JP2011125154A - Demagnetization determining system of rotating electric machine - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a technique which accurately determines demagnetization of a permanent magnet, regardless of the operation state of a rotating electric machine. <P>SOLUTION: The demagnetization determining system 10 includes an actual q-axis voltage acquisition section 1 for acquiring a q-axis target voltage that performs feedback control as an actual q-axis voltage vq, wherein a current flowing into a polyphase coil of a permanent magnet-type rotating electric machine is subjected to coordinate transformation, into a vector component of the d-axis which is the direction of a magnetic field generated by a permanent magnet and the q-axis orthogonal to the d-axis; a reference q-axis voltage acquiring section 2 for acquiring the q-axis target voltage, when the permanent magnet is not demagnetized as a reference q-axis voltage vqr, and a demagnetization determining section 7 for determining the demagnetization state of the permanent magnet, based on an operation state division indicating at least a power operation state or a regenerative operation state, the actual q-axis voltage vq and the reference q-axis voltage vqr. <P>COPYRIGHT: (C)2011,JPO&INPIT

Description

  The present invention relates to a demagnetization determination system for a rotating electrical machine that determines demagnetization of a permanent magnet of a permanent magnet type rotating electrical machine.

  In recent years, attempts to reduce the environmental load due to the consumption of fossil fuels have been widely carried out, and in the industry, vehicles having a smaller environmental load than vehicles driven by an internal combustion engine have been proposed. An electric vehicle driven by a rotating electrical machine and a hybrid vehicle driven by an internal combustion engine and a rotating electrical machine are examples. Electric vehicles and hybrid vehicles are equipped with rotating electric machines equipped with permanent magnets such as interior permanent magnet synchronous motors / generators (IPMSM). However, a phenomenon of demagnetization in which the magnetic flux decreases due to a temperature rise or the like occurs in the permanent magnet. Since the rotating electrical machine that functions as a motor generates a force (torque) by a magnetic field and an electric current, the torque required by demagnetization may not be obtained.

  The magnetic flux of the permanent magnet can be known from the counter electromotive force generated in proportion to the rotational speed of the rotor. In vector control known as a method of controlling a rotating electrical machine, the current flowing in each of the three phases of the rotating electrical machine is converted into vector components of the d axis that is the direction of the magnetic field generated by the permanent magnet and the q axis that is orthogonal to the d axis. Perform coordinate control and feedback control. In this vector control, the formula for deriving the q-axis voltage command includes a term of back electromotive force. Therefore, it is possible to determine whether or not demagnetization has occurred by monitoring the q-axis voltage command. Japanese Patent Laid-Open No. 2005-51892 (Patent Document 1) stores a normal q-axis voltage command in which no demagnetization occurs, an actual q-axis voltage command while the rotating electrical machine is driving, and a normal time A technique for determining demagnetization by comparing with a q-axis voltage command is disclosed. In Patent Document 1, the normal q-axis voltage command value at which no demagnetization occurs is corrected based on the increase or decrease in inverter input voltage, dead time, inverter switching frequency, and current vector angle.

JP 2005-51892 A (paragraphs 10 to 17)

  However, an experiment by the inventors found that the q-axis voltage command changes depending on whether the operating state of the rotating electrical machine is power running or regenerative. Therefore, in the method of Patent Document 1, the estimation accuracy of the magnitude of demagnetization becomes low, and there is a possibility that demagnetization cannot be accurately determined. For this reason, a technique for accurately determining the demagnetization of the permanent magnet is required regardless of the operating state of the rotating electrical machine.

The characteristic configuration of the demagnetization determination system for a rotating electrical machine according to the present invention in view of the above problems is as follows.
When the feedback control is performed by converting the current flowing in the multiphase coil of the permanent magnet type rotating electrical machine into a vector component of the d-axis that is the direction of the magnetic field generated by the permanent magnet and the q-axis orthogonal to the d-axis. an actual q-axis voltage acquisition unit that acquires a q-axis target voltage as an actual q-axis voltage;
A reference q-axis voltage acquisition unit configured to acquire, as a reference q-axis voltage, a q-axis target voltage that is set in advance according to the rotation speed and target torque of the rotating electrical machine and in which the demagnetization is not generated in the permanent magnet;
Determining the demagnetization state of the permanent magnet on the basis of an operation state classification indicating whether the operation state of the rotating electrical machine is at least a power running operation state or a regenerative operation state, and the actual q-axis voltage and the reference q-axis voltage And a demagnetization determination unit.

  According to this characteristic configuration, when the demagnetization of the permanent magnet is determined using the actual q-axis voltage and the reference q-axis voltage, whether the operation state of the rotating electrical machine is a power running operation state or a regenerative operation state. Since it determines according to the driving | running state division showing a division, it becomes possible to determine a demagnetization more correctly. Therefore, according to this characteristic configuration, it is possible to provide a demagnetization determination system for a rotating electrical machine that can accurately determine the demagnetization of a permanent magnet regardless of the operating state of the rotating electrical machine.

  Here, the demagnetization determination unit of the demagnetization determination system for a rotating electrical machine according to the present invention corrects one of the actual q-axis voltage and the reference q-axis voltage according to the operating state classification, It is preferable to determine the demagnetization state of the permanent magnet in comparison with either the actual q-axis voltage or the reference q-axis voltage. According to this, since demagnetization can be determined by correcting the actual q-axis voltage or the reference q-axis voltage, it is not necessary to set the reference q-axis voltage separately for power running and regeneration. Therefore, it is possible to accurately determine the demagnetization of the permanent magnet regardless of the operating state of the rotating electrical machine, and to reduce the memory usage of the control device.

  Here, the demagnetization determination unit of the demagnetization determination system for a rotating electrical machine according to the present invention has a value that differs depending on the operating state classification for either the actual q-axis voltage or the reference q-axis voltage. It is preferable to determine the demagnetization state by calculating the amount of demagnetization based on the difference between the correction q-axis voltage obtained by multiplying the correction parameter and either the actual q-axis voltage or the reference q-axis voltage. It is. According to this, by the simple calculation of multiplying the actual q-axis voltage or the reference q-axis voltage by the correction parameter, the correction according to the operation state classification can be performed and the demagnetization amount of the permanent magnet can be calculated. Further, by performing the correction, the estimation accuracy of the magnitude of demagnetization is increased, and a more accurate demagnetization determination is possible. Note that the amount of demagnetization is not limited to the amount of decrease in magnetic flux, but also includes the rate at which the magnetic flux decreases, that is, the demagnetization factor.

  In the demagnetization determination system for a rotating electrical machine according to the present invention, the operating state classification includes a zero torque operating state in which the output torque of the rotating electrical machine is controlled to be zero in addition to the power running operation state and the regenerative operation state. In addition, it is preferable that the correction parameter has a different value depending on each operation state classification. Experiments by the inventors have confirmed that the q-axis voltage is different in each of the power running state, the regenerative operation state, and the zero torque operation state. Therefore, it is possible to estimate the demagnetization amount of the permanent magnet with higher accuracy by providing the operation state classification according to the actual situation and using the correction parameters having different values depending on each operation state classification.

  Further, in the demagnetization determination system for a rotating electrical machine according to the present invention, the correction parameter represents a relationship between one of the actual q-axis voltage and the reference q-axis voltage and a DC power supply voltage for each of the operating state categories. It is preferable that the demagnetization determination unit determines a correction parameter based on the operating state classification and the DC power supply voltage. The d-axis and q-axis in vector control are computational concepts, and an actual rotating electrical machine is driven by applying a multiphase alternating current to each winding of, for example, a three-phase multilayer winding. In an automobile or the like, this multiphase AC is generated by AC conversion of a DC power source such as a secondary battery via an inverter. In many cases, this inverter is constituted by a bridge circuit using a power semiconductor element such as an IGBT (insulated gate bipolar transistor). In this bridge circuit, an arm corresponding to each phase is composed of a high-side element and a low-side element that are complementarily connected. When the high-side element and low-side element of the same arm are turned on at the same time, the circuit is short-circuited without passing through the windings of the rotating electrical machine, so there is a dead time when both elements are turned off simultaneously. Provided. Here, the semiconductor elements constituting the inverter are often switched by PWM (pulse width modulation) control, and the dead time is set by reducing the pulse width of PWM control. When the DC voltage applied to the inverter changes, even if the dead time is the same time, the cut power is different. Therefore, the q-axis voltage used for calculating the amount of demagnetization is preferably a value corrected according to the DC power supply voltage. The demagnetization determination system for a rotating electrical machine according to the present invention further includes a correction parameter that represents the relationship between one of the actual q-axis voltage and the reference q-axis voltage and the DC power supply voltage for each operating state section. Correction according to the voltage is effectively performed.

A block diagram schematically showing a system configuration example of a vehicle including a demagnetization determination system Block diagram schematically showing a configuration example of a rotating electrical machine control system Block diagram schematically showing a configuration example of a converter and an inverter Explanatory diagram showing the concept of q-axis voltage map Explanatory drawing which shows the concept of the correction parameter at the time of power running Explanatory drawing which shows the concept of the correction parameter at the time of regeneration Explanatory drawing showing the concept of correction parameters during zero torque operation Flow chart showing an example of a demagnetization determination procedure Block diagram showing another configuration example of demagnetization determination system

  Hereinafter, an embodiment of a demagnetization determination system according to the present invention will be described with reference to the drawings, taking as an example a control system that drives and controls a rotating electrical machine serving as a drive source for a vehicle such as a hybrid vehicle or an electric vehicle. This rotating electric machine is a permanent magnet embedded synchronous machine, and functions as an electric motor or a generator depending on the situation. Hereinafter, the rotary electric machine will be described as appropriately referred to as a motor, and this refers to a rotary electric machine that functions as an electric motor and a generator. 1 to 3 show an example of a system configuration of a vehicle including such a motor M control system (rotary electrical machine control system 20), an example of a configuration of a rotary electrical machine control system 20 including a demagnetization determination system 10, and a rotary electrical machine control. The structural example of the converter 50 and the inverter 60 of the system 20 is shown typically.

  The motor M is electrically connected to the converter 50 that boosts the output voltage of the battery B via the inverter 60. When functioning as an electric motor, the motor M is supplied with electric power via the converter 50 and the inverter 60 and is driven. (Powering operation). On the other hand, when the motor M functions as a generator, the generated AC power is converted into DC by the inverter 60, and is stepped down in the converter 50 as needed and regenerated to the battery B (regenerative operation). The converter 50 and the inverter 60 are configured to have a plurality of switching elements. It is preferable to apply an insulated gate bipolar transistor (IGBT) or a metal oxide semiconductor field effect transistor (MOSFET) to the switching element. As shown in FIG. 3, in this embodiment, an IGBT is used as a switching element.

  As is well known, the converter 50 includes a reactor, a filter capacitor, a pair of upper and lower IGBTs connected in series, a discharge resistor, a smoothing capacitor, and the like. The emitter of the upper IGBT of converter 50 is connected to the collector of the lower IGBT, and is connected to the positive side of battery B through the reactor. The collector of the upper IGBT is connected to the input plus side of the inverter 60. The emitter of the lower IGBT is connected to the negative side (ground) of battery B and to the input negative side of inverter 60. The filter capacitor is connected between the plus and minus of the output of the battery B, and the discharging resistor and the smoothing capacitor are between the plus and minus of the output of the converter 50, that is, between the collector of the upper IGBT and the emitter of the lower IGBT. Connected in parallel.

  When the upper IGBT is turned off and the lower IGBT is switched by, for example, PWM (pulse width modulation) control, the output voltage of the battery B is boosted and supplied to the inverter 60. When the lower IGBT is turned off and the upper IGBT is turned on, the electric power generated by the motor M is regenerated to the battery B. When the electric power generated by the motor M is stepped down and regenerated in the battery B, the upper IGBT may be PWM-controlled.

  The inverter 60 that converts direct current into three-phase alternating current is constituted by a three-arm bridge circuit as is well known. Two IGBTs are connected in series between the input plus side and the input minus side of the inverter 60, and this series circuit is connected in parallel in three lines (three arms). That is, a bridge circuit in which a set of series circuits (arms) corresponds to each of the stator coils Mu, Mv, and Mw corresponding to the u phase, v phase, and w phase of the motor M is configured. The collector of the upper stage IGBT of each phase is connected to the input plus side of the inverter 60, and the emitter is connected to the collector of the lower stage IGBT of each phase. Further, the emitter of the IGBT on the lower side of each phase is connected to the input minus side (for example, ground) of the inverter 60. The intermediate point of the series circuit of IGBTs of each phase that is a pair, that is, the connection point of the IGBT, is connected to the stator coils Mu, Mv, and Mw of the motor M, respectively.

  Note that a flywheel diode (regenerative diode) is connected in parallel to each IGBT. The flywheel diode is connected in parallel to the IGBT such that the cathode terminal is connected to the collector terminal of the IGBT and the anode terminal is connected to the emitter terminal of the IGBT. The gate of each IGBT is connected to an ECU (electronic control unit) 30 via a driver circuit 33 and is individually controlled to be switched. In the present embodiment, the ECU 30 is referred to as a TCU (trans-axle control unit) 30 in order to distinguish it from other ECUs.

  The TCU 30 is configured with a logic circuit such as a microcomputer as a core. In the present embodiment, the TCU 30 includes a CPU (central processing unit) 40 that is a microcomputer, an interface circuit 31, and other peripheral circuits. The interface circuit 31 includes EMI (electro-magnetic interference) countermeasure parts, a buffer circuit, and the like. The CPU 40 implements the demagnetization determination system 10 and the rotating electrical machine control system 20 of the present invention by cooperation of hardware and software. The potential difference between the high level and the low level of the pulsed drive signal input to the gate of the IGBT or MOSFET that switches high voltage is higher than the drive voltage of a general electronic circuit such as a microcomputer. Therefore, the drive signal is boosted through the driver circuit 33 and then input to the converter 50 and the inverter 60.

  The CPU 40 includes at least a CPU core 41, a program memory 42, a parameter memory 43, and a work memory 44. In addition, although it has a communication control part, an A / D converter, a timer, a port, etc., it is a general matter and illustration is omitted for easy explanation. The CPU core 41 is the core of the CPU 40, and includes an instruction register, an instruction decoder, an ALU (arithmetic logic unit), a flag register, a general-purpose register, an interrupt controller, and the like that perform various operations. The program memory 42 is a nonvolatile memory that stores a rotating electrical machine control program, a demagnetization determination program, and the like. The parameter memory 43 is a non-volatile memory that stores various parameters that are referred to when the program is executed. A q-axis voltage map and correction parameters to be described later are stored in the parameter memory 43. The parameter memory 43 may be included in the program memory 42 without being distinguished from the program memory 42. The program memory 42 and the parameter memory 43 are preferably constituted by a flash memory, for example. The work memory 44 is a memory that temporarily stores temporary data during program execution. The work memory 44 is composed of DRAM (dynamic RAM) or SRAM (static RAM) that is volatile and can read and write data at high speed.

  The communication control unit controls communication with other systems in the vehicle. In the present embodiment, communication with the travel control system 70, other systems, sensors, and the like is controlled via a CAN (controller area network) 80 that is a network in the vehicle. Similar to the TCU 30, the traveling control system 70 is configured with an electronic circuit such as a CPU as a core, and is configured as an ECU together with peripheral circuits. The A / D converter converts an analog electric signal into digital data. For example, when the detection results of the motor currents Iu, Iv, and Iw flowing through the stator coils Mu, Mv, and Mw of the motor M are output as analog signals from the current sensor 35, the analog signals are converted into digital values. Note that the three phases u phase, v phase, and w phase are balanced and the instantaneous value is zero, so that only the current for two phases may be detected, and the remaining one phase may be obtained by calculation in the CPU 40. . In this embodiment, the case where all three phases are detected is illustrated.

  The timer measures time with the clock period of the CPU 40 as the minimum resolution. For example, the timer monitors the program execution cycle and notifies the interrupt controller of the CPU core 41. The timer measures the valid time of the gate drive signals (pu, nu, pv, nv, pw, nw) for driving the IGBT, and generates the gate drive signal. The port is a terminal control unit that outputs an IGBT gate drive signal or the like via a terminal of the CPU 40 and receives a rotation detection signal R from the rotation detection sensor 37 input to the CPU 40. The rotation detection sensor 37 is a sensor that is installed in the vicinity of the motor M and detects the rotation position and rotation speed of the rotor of the motor M, and is configured using, for example, a resolver.

  In the present embodiment, the demagnetization determination system 10 and the rotating electrical machine control system 20 are configured with the CPU 40 as a core. That is, the demagnetization determination system 10 and the rotating electrical machine control system 20 include hardware including mainly the CPU core 41 and the work memory 44, and software such as programs and parameters stored in the program memory 42 and the parameter memory 43. Realized by collaboration. However, the embodiment of the demagnetization determination system 10 and the rotating electrical machine control system 20 is not limited to such cooperation between hardware and software, and hardware using ASIC (Application Specific Integrated Circuit) or the like. It may be composed only of wear.

  By the way, as a method for controlling an AC motor, a control method called vector control is known. In the vector control, the coil current flowing through the three-phase (multi-phase) stator coil of the AC motor is converted into a d-axis that is a direction of a magnetic field generated by a permanent magnet disposed on the rotor, and q orthogonal to the d-axis. The feedback control is performed by converting the coordinates to the vector component with the axis. The rotating electrical machine control system 20 of this embodiment also employs this vector control. As shown in FIG. 2, the rotating electrical machine control system 20 includes a torque control unit 11, a current control unit 12, an inverter control unit 13, a rotation state calculation unit 14, and a converter control unit 15. The Moreover, since the demagnetization determination system 10 of this embodiment performs demagnetization determination using the command value in vector control, in this embodiment, it has illustrated as one function part of the rotary electric machine control system 20. .

  For coordinate conversion in vector control, it is necessary to know the rotation state of the motor M in real time. Therefore, as shown in FIGS. 1 and 3, a rotation detection sensor 37 such as a resolver is provided in the vicinity of the motor M. The detection result is transmitted to the register in the CPU core 41 and the work memory 44 via the interface circuit 31 as described above. The rotation state calculation unit 14 obtains the rotor position (electrical angle θ) and the rotation speed (angular velocity ω) based on the detection result R of the rotation detection sensor 37. The obtained electrical angle θ and angular velocity ω are used in each part of the rotating electrical machine control system 20 including the demagnetization determination system 10. When the rotation detection sensor 37 such as a resolver provides information on the rotor position and the rotation speed in a form that the demagnetization determination system 10 and the rotating electrical machine control system 20 can use for calculation, the rotation state calculation unit 14 is provided. It does not have to be.

  The torque controller 11 includes a target torque calculator 11a and a two-phase converter 11b. The target torque calculation unit 11a is a functional unit that calculates a target torque (torque command) Ti in accordance with the required torque Tr transmitted from the travel control system 70. The required torque Tr and the target torque Ti may be the same, but when the required torque Tr is too large, the target torque calculation unit 11a suppresses it within an allowable value. Further, the required torque Tr may be corrected in order to suppress torque ripple. The two-phase conversion unit 11b is a functional unit that sets target currents (current commands) id and iq for current feedback control according to the target torque Ti. The target currents id and iq are calculated corresponding to the d-axis and q-axis described above. Therefore, the torque control unit 11 converts the target torque Ti into the d-axis target current id and the q-axis target current iq based on the electrical angle θ obtained by the rotation state calculation unit 14 in the two-phase conversion unit 11b. To do.

  The current control unit 12 performs, for example, proportional-integral control (PI control) or proportional-calculus control (PID control) based on the deviation between the target currents id, iq and the fed back motor current, and the target voltage (voltage command). ) A functional unit for setting vd and vq. Here, a case where proportional calculus control (PID control) is performed is illustrated, and the current control unit 12 includes a PID control unit 12b and a two-phase conversion unit 12a. Since the current values detected by the current sensor 37 are the three-phase currents Iu, Iv, and Iw, the two-phase conversion unit 12a uses the two-phase current Id based on the electrical angle θ obtained by the rotation state calculation unit 14. , Iq. The PID control unit 12b performs PID control based on the deviation between the target currents id, iq and the two-phase motor currents Id, Iq and the angular velocity ω obtained by the rotational state calculation unit 14, and performs the target voltage vd, vq is calculated. Here, the target voltage vd is a d-axis target voltage, and the target voltage vq is a q-axis target voltage. The q-axis target voltage vq corresponds to the actual q-axis voltage of the present invention.

  The inverter control unit 13 generates gate drive signals pu, nu, pv, nv, pw, nw for driving the three-phase IGBT of the inverter 60 based on the target voltages vd, vq. The inverter control unit 13 includes a three-phase conversion unit 13a and a PWM control unit 13b. Since the inverter 60 is provided for three phases corresponding to the three-phase stator coils Mu, Mv, and Mw of the motor M, the two-phase target voltages vd and vq are obtained by the rotation state calculation unit 14. Based on the electrical angle θ, the coordinates are converted into three-phase target voltages (vu, vv, vw). The PWM control unit 13b includes six gate drive signals pu that individually control the gates of the IGBTs in each stage, each phase being constituted by two stages of IGBTs on the upper stage side and the lower stage side, and a total of six stages in three phases. nu, pv, nv, pw, nw are generated. The gate drive signals pu, nu, pv, nv, pw, and nw correspond to the gate drive signals of the IGBTs in the u-phase upper stage, u-phase lower stage, v-phase upper stage, v-phase lower stage, w-phase upper stage, and w-phase lower stage in order. . The valid time of each gate drive signal is measured by the timer of the CPU 40, and pulsed gate drive signals pu, nu, pv, nv, pw, nw are output through the ports.

  Converter control unit 15 generates IGBT gate drive signals pc, nc of converter 50 based on a boost command set in accordance with required torque Tr (or target torque Ti in some cases). The voltage VB of the battery B and the output voltage VC of the converter 50 (input of the inverter 60) are transmitted to the converter control unit 15, and a boost command is calculated based on these voltage values. The gate drive signal pc is a drive signal for the upper-stage IGBT, and the gate drive signal nc is a drive signal for the lower-stage IGBT.

  The demagnetization determination system 10 is a functional unit that calculates the demagnetization amount of the permanent magnet mounted on the motor M and determines the degree of demagnetization. The magnetic flux of the permanent magnet can be known from the back electromotive force generated in proportion to the rotational speed (angular velocity ω) of the rotor of the motor M. The voltage equation in vector control is: φ: armature linkage flux, ω: angular velocity, id: d-axis current, iq: q-axis current, Ld: d-axis inductance, Lq: q-axis inductance, Ra: armature resistance, p: As a differential operator, it is expressed as the following formula (1).

  From equation (1), the q-axis target voltage vq is represented by the following equation (2) and includes a term consisting of the product (ωφ) of the angular velocity ω and the linkage flux φ of the armature.

  In the states shown in equations (1) and (2), it is assumed that the magnetic flux of the permanent magnet of the armature is ideal, and here, the case where the magnetic flux of the permanent magnet is reduced is considered. If K is a coefficient of 1 or less representing the residual rate of magnetic flux, the actual flux linkage φ of the armature is “K · φ”. Therefore, the q-axis target voltage vq calculated by the equation (2) is actually as shown in the following equation (3).

  Here, assuming that the ideal q-axis target voltage obtained by Equation (2) is the reference q-axis voltage vqr, and the q-axis target voltage actually calculated in the rotating electrical machine control system 20 is the actual q-axis voltage vq, Based on the equations (4) and (5), the demagnetizing factor J of the permanent magnet can be obtained from the difference between the actual q-axis voltage vq and the reference q-axis voltage vqr. Further, since the initial magnetic flux before demagnetization, for example, the magnetic flux φ at the time of magnetization, is known, the amount of demagnetization can also be obtained by multiplying the magnetic flux φ by the demagnetization factor.

  As described above, the calculation formula for deriving the q-axis target voltage in the vector control includes the term of the back electromotive force. Therefore, by monitoring the actual q-axis voltage vq, it is determined whether or not demagnetization has occurred. It can be performed. The normal q-axis target voltage (reference q-axis voltage vqr) at which no demagnetization occurs is stored as the q-axis voltage map 3 in the program memory 42 or the parameter memory 43. FIG. 4 is an explanatory diagram showing the concept of the q-axis voltage map 3. As is clear from the configuration of the current control unit 12 shown in FIG. 2 and Expression (1), the q-axis target voltage is a function of the target currents id and iq. Accordingly, the reference q-axis voltage vqr is a point on the q-axis voltage map 3 that can be read with the target currents id and iq as arguments as shown in FIG.

  Similarly, as is clear from FIG. 2 and equation (1), the reference q-axis voltage vqr also depends on the angular velocity ω. Therefore, a plurality of typical rotation numbers, for example, the q-axis voltage map 3 at the time of low rotation, medium rotation, and high rotation are stored in the program memory 42 and the parameter memory 43. Of course, the q-axis voltage map 3 may be stored corresponding to four or more rotation speeds. Since the q-axis target voltage also depends on the DC voltage (VC) of the inverter 60 as will be described later, the q-axis voltage map 3 is set corresponding to the value of a typical DC voltage. Note that a curve a in FIG. 4 shows a locus of the q-axis target voltage vd during the maximum torque control. Further, the origin side with the curve b in FIG. 4 as a boundary indicates a variable range of the q-axis target voltage vd.

  The demagnetization determination system 10 is a functional unit that determines demagnetization of a permanent magnet based on the principle described above. In the present embodiment, the demagnetization determination system 10 includes an actual q-axis voltage acquisition unit 1, a reference q-axis voltage acquisition unit 2, and a demagnetization determination unit 7, as shown in FIG. .

  The actual q-axis voltage acquisition unit 1 is a functional unit that acquires the q-axis target voltage vq calculated by the current control unit 12 as an actual q-axis voltage. In the present embodiment, the rotating electrical machine control system 20 including the demagnetization determination system 10 is configured with a microcomputer as a core. Therefore, the real q-axis voltage acquisition unit 1 realizes, for example, the q-axis target voltage vq by cooperation of hardware such as a general-purpose register of the CPU core 41 and software that causes the hardware to latch the q-axis target voltage vq. Is done. In general, the calculation cycle in which the q-axis target voltage vq is calculated is shorter than the calculation cycle in which the demagnetization determination calculation is performed. For example, if the calculation cycle in which the q-axis target voltage vq is calculated is T [ms], the calculation cycle in which the demagnetization determination calculation is executed is 10 T [ms]. Further, in order to suppress torque ripple, the q-axis target voltage vq may vary finely. Therefore, the real q-axis voltage acquisition unit 1 acquires one q-axis target voltage vq by thinning out or smoothing the q-axis target voltage vq that can be acquired 10 times within a calculation period of 10T [ms]. The filter 1a for this is comprised. This filter 1a can also be configured by software.

  The reference q-axis voltage acquisition unit 2 is a functional unit that acquires a q-axis target voltage in a state where no demagnetization occurs in the permanent magnet as a reference q-axis voltage vqr. The reference q-axis voltage vqr is a value set in advance according to the rotation speed of the motor M and the target torque. As described above, the reference q-axis voltage vqr is stored as the q-axis voltage map 3 with the target currents id and iq and the angular velocity ω as arguments. As shown in FIG. 2, the target currents id and iq are calculated from the target torque (Ti) in the torque control unit 11, and as described above, the q-axis voltage map 3 has the rotational speed (angular velocity ω). Multiple sheets are set accordingly. Therefore, it can be said that the reference q-axis voltage vqr is a value set in advance according to the rotation speed of the motor M and the target torque.

  In the present embodiment, the reference q-axis voltage acquisition unit 2 is similar to the real q-axis voltage acquisition unit 1, for example, hardware such as a general-purpose register of the CPU core 41 and the parameter memory 43, and the reference q-axis to the hardware. This is realized in cooperation with software for latching the voltage vqr. In the present embodiment, the reference q-axis voltage acquisition unit 2 is exemplified in FIG. 2 as including the q-axis voltage map 3, but the q-axis voltage map 3 is the reference q-axis voltage acquisition unit 2. And may be different. It is clear that both the reference q-axis voltage acquisition unit 2 and the q-axis voltage map 3 are functional units realized by the cooperation of hardware and software, and their configurations are not limited to the scope disclosed in the drawings. is there.

  Similar to the cycle in which the target voltages vd and vq are calculated, the calculation cycle in which the target currents id and iq are calculated is also shorter than the calculation cycle in which the demagnetization determination calculation is executed. For example, if the calculation cycle in which the target currents id and iq are calculated is T [ms], the calculation cycle in which the demagnetization determination calculation is performed is 10 T [ms]. Moreover, in order to suppress torque ripple, the target currents id and iq may also fluctuate finely. Therefore, the reference q-axis voltage acquisition unit 2 also thins out or smoothes the target currents id and iq that can be acquired 10 times during the calculation period of 10T [ms], similarly to the real q-axis voltage acquisition unit 1. And a filter 2a for obtaining a set of target currents id and iq. Naturally, the filter 2a can also be configured by software.

  The demagnetization determination unit 7 is a functional unit that determines the demagnetization state by calculating the demagnetization amount of the permanent magnet using the actual q-axis voltage vq and the reference q-axis voltage vqr. As is clear from the above description using the equations (4) and (5), the amount of demagnetization (the amount of magnetic flux [wb]) is obtained by multiplying the initial magnetic flux φ [wb] by the demagnetizing factor J. It is possible to calculate by. Therefore, obtaining the demagnetization factor J is substantially the same as obtaining the demagnetization amount, and the demagnetization determination unit 7 calculates the demagnetization factor of the permanent magnet and determines the demagnetization state. Good.

  As shown in FIG. 2, the demagnetization determination unit 7 includes a demagnetization amount calculation unit 5 that calculates the demagnetization amount of the permanent magnet, and a determination unit 6 that determines the degree of demagnetization based on the calculated demagnetization amount. With. Of course, the demagnetization amount calculation unit 5 may calculate the demagnetization rate instead of the demagnetization amount, and the determination unit 6 also determines the degree of demagnetization based on the demagnetization rate instead of the demagnetization amount. Also good. The demagnetization determination unit 7 of the present embodiment further corrects the q-axis voltage according to an operation state classification that represents at least a classification of whether the operation state of the motor M (rotary electric machine) is a power running operation state or a regenerative operation state. A shaft voltage correction unit 4 is provided. In the embodiment shown in FIG. 2, the actual q-axis voltage vq is corrected according to the operating state classification. The driving state classification is transmitted as the driving state classification signal DS from the traveling control system 70 or the like to the demagnetization determination unit 7.

  Further, the voltage value VC of the DC power supply voltage of the inverter 60 is also transmitted to the demagnetization determination unit 7. As described above, the d-axis and q-axis in vector control are computational concepts, and an actual rotating electrical machine is applied with multi-phase AC to each winding of the multi-phase winding during powering operation. In the regenerative operation, multiphase alternating current corresponding to each winding is generated. In the present embodiment, a multiphase alternating current (three-phase alternating current) is generated by converting the direct current of the voltage VC through the inverter 60 during power running operation, and the generated multiphase alternating current is generated during regenerative operation. Is converted to a direct current of the voltage VC through the inverter 60.

  The inverter 60 is configured by a bridge circuit using a power semiconductor element such as IGBT as shown in FIG. In this bridge circuit, if the high-side IGBT and the low-side IGBT of the same arm are turned on at the same time, the circuit is short-circuited without passing through the winding of the motor M. A dead time is provided. The inverter 60 is switched by PWM control, and the dead time is set by cutting the pulse width of the pulsed switching signal. The voltage VC varies due to voltage variation of the battery B, boosting by the converter 50, and the like. For example, during powering operation, if the DC voltage VC applied to the inverter 60 changes, even if the dead time is the same time, a difference occurs in the reduced power. Therefore, the value VC of the DC power supply voltage to the inverter 60 is also transmitted to the demagnetization determination unit 7, and the q-axis voltage is corrected according to the voltage VC. In addition, when the motor M is in the power running operation state and in the regenerative operation state, even if the DC power supply voltage value VC is the same, the amount of power affected by the dead time is different. Experiments by the inventor have confirmed that the q-axis voltage value is different in each of the power running operation state and the regenerative operation state, and that the tendency of the change according to the DC power supply voltage value VC is different.

  FIG. 5 is a graph showing the relationship between the DC power supply voltage value VC and the q-axis voltage vq when the motor M is in the power running state. Although a plurality of lines are shown in FIG. 5, one line represents the q-axis voltage vq defined by the target currents id and iq (q-axis voltage vq corresponding to one point on the map of FIG. 4). The change with respect to value VC of DC power supply voltage is shown. The absolute value of the target currents id and iq is larger on the lower line on the graph. Even if the q-axis voltage vq is a different value, the change tendency of the q-axis voltage vq according to the change of the value VC of the DC power supply voltage, that is, the slope of the graph is almost the same.

  FIG. 6 is a graph showing the relationship between the DC power supply voltage VC and the q-axis voltage vq when the motor M is in the regenerative operation state. Although a plurality of lines are shown, the meaning is the same as in FIG. 5, and the absolute value of the target currents id and iq is larger in the lower line on the graph. Even in the regenerative operation state, as in the power running operation state, even if the q-axis voltage vq is a different value, the tendency of the change in the q-axis voltage vq according to the change in the DC power supply voltage value VC, that is, the graph The slopes of are almost the same. However, the slope of the graph (absolute value of the slope) is different between the powering operation state and the regenerative operation state.

  Therefore, the q-axis voltage correction unit 4 corrects the q-axis voltage based on the operation state classification signal DS indicating whether the motor M is in the power running operation state or the regenerative operation state and the value VC of the DC power supply voltage. Is done. In the present embodiment, the q-axis voltage correction unit 4 corrects the actual q-axis voltage vq to become a corrected q-axis voltage vqm. For example, the q-axis voltage correction unit 4 calculates a correction q-axis voltage vqm by multiplying a correction parameter that is a different value depending on the operating state classification. This correction parameter represents the relationship between the DC power supply voltage and the q-axis voltage for each operating state category. Specifically, the correction parameter is a value corresponding to the slope of the graph for each operating state category shown in FIGS.

  Note that the operation state classification can include a zero torque operation state in which the output torque of the rotating electrical machine is controlled to be zero in addition to the power running operation state and the regenerative operation state. The correction parameter is set as a different value depending on each operation state classification including the zero torque operation state. FIG. 7 is a graph showing the relationship between the DC power supply voltage value VC and the q-axis voltage vq when the rotating electrical machine is in the zero torque operation state. Since the q-axis voltage vq is single because of the zero torque operation state, the q-axis voltage vq changes according to the change of the DC power supply voltage, and the tendency of the change is the power running operation state and the regenerative operation state. Is different.

  Hereinafter, the procedure of the demagnetization determination process by the demagnetization determination system 10 will be described using the flowchart shown in FIG. First, it is determined whether or not a precondition for determining demagnetization is satisfied (# 1). Here, the preconditions include whether or not the temperature of the rotating electrical machine M is within the determination target temperature range, whether or not the rotating electrical machine M is being driven, and the like. Also included is whether to continue using the calculation result of the demagnetization determination performed so far, or whether to start the calculation again from the beginning. If the precondition is not satisfied, the time counter and the detection counter are cleared to the initial state (= 0) (# 16, # 17). The functions of the time counter and the detection counter will be described later.

  When it is determined in step # 1 that the precondition is satisfied, a control variable used in the demagnetization determination system 10 is acquired (# 2). The control variables include target current id, iq, angular velocity ω, actual q-axis voltage vq, DC power supply voltage value VC, operation state classification, correction parameter, and the like. As described above, the control variable is filtered as necessary and latched in a register or the like. Then, the actual q-axis voltage vq is corrected in the q-axis voltage correction unit 4, and the corrected q-axis voltage vqm is calculated (# 3). Next, the map index calculation is executed using the target currents id, iq and the angular velocity ω as arguments, and the reference q-axis voltage vqr is acquired as a map value from the q-axis voltage map 3 (# 4, # 5). Steps # 4 to # 5 may be executed prior to step # 3.

  Next, a demagnetization amount (or demagnetization factor) is calculated in the demagnetization amount calculation unit 5 (# 6). Subsequently, the determination threshold value and the demagnetization amount (or demagnetization factor) are compared (# 7). Note that the degree of demagnetization such as weak demagnetization state or strong demagnetization state can also be determined from the magnitude of the demagnetization amount (or demagnetization factor). For example, it can be determined that a weak demagnetization state is obtained when the demagnetization factor is 30% or more, and a strong demagnetization state when the demagnetization factor is 50% or more. For ease of explanation, it is assumed here that one type of determination threshold is compared with the amount of demagnetization (or demagnetization rate), and it is simply determined whether or not the state is demagnetized.

  When the demagnetization amount (or demagnetization factor) exceeds the determination threshold value and is determined to be NG in the demagnetization determination in step # 7, the detection counter is incremented (# 8). If it is not determined as NG in the demagnetization determination in step # 7, step # 8 is skipped and the process proceeds to step # 9. In step # 9, it is determined whether or not the value of the detection counter is greater than zero. As described above, since the initial value of the detection counter is zero, it is determined in step # 9 whether or not at least one NG determination has been made in the past. If it is determined in step # 9 that the value of the detection counter is 1 or more, the time counter is incremented (# 10). If it is determined in step # 9 that the value of the detection counter is the initial value, step # 10 is skipped and the process proceeds to step # 11. Detailed functions of the detection counter and the time counter in steps # 8 to # 10 will be described later.

  In step # 11, it is determined whether or not the value of the time counter is 200 or less. If it is 200 or less, it is next determined whether or not the value of the detection counter is 160 or more (# 12). If it is determined that the value of the detection counter is 160 or more, it is determined that the demagnetization state is deterministic, and a definite flag F (see FIG. 2) is output (# 13). The confirmation flag F is transmitted to the rotating electrical machine control system 20, the travel control system 70, and other in-vehicle systems. Each system that has received the confirmation flag F emits a warning lamp in the vehicle, issues an alarm message, or stops the rotating electrical machine. That is, the detection counter has a function for improving the accuracy of the demagnetization determination by accumulating the determination results for the demagnetization amount (demagnetization factor) a plurality of times. Although it is not preferable to make the user uneasy or prevent smooth running of the vehicle by erroneous detection, it is possible to suppress erroneous detection satisfactorily by providing a detection counter.

  On the other hand, if the value of the detection counter is less than 160 in step # 12, step # 13 is skipped, and the process is executed again sequentially from step # 1. A series of processes from step # 1 to step # 13 is one calculation process for demagnetization determination, and is executed, for example, every calculation cycle 10T [ms] as described above. When this calculation cycle is repeated a plurality of times, that is, at least 160 times or more, and the value of the detection counter reaches 160 (a predetermined number of times), the determination result that the demagnetization state is deterministic A determination flag F is output.

  The time counter is a counter that monitors whether or not the detection counter reaches 160 within a 200 calculation cycle after activation (during a predetermined number of repetitions). The time counter is started after it is determined that the demagnetization state is first in the repeated processing and the detection counter becomes 1 or more. Therefore, the time counter has a function of monitoring a timeout after it is first determined that the demagnetization state is present. If the detection counter does not reach 160 within 200 calculation cycles after the time counter is activated, the determination result of the demagnetization state is not stable and the result of the demagnetization determination in each calculation cycle is noise. There is a high possibility that it is disturbed by factors such as. Therefore, in step # 11, it is determined whether or not a timeout has occurred. If it is determined that a time-out has occurred, the time counter and the detection counter are cleared to the initial state (# 14, # 15), and the demagnetization determination is executed from the beginning. If only the detection counter is provided and the determination results of step # 7 are integrated without providing a time counter, the determination results due to erroneous detection are also accumulated. By providing the time counter, accumulation of such erroneous detections is suppressed.

  In the embodiment illustrated in FIG. 8, it has been described that in step # 7, one type of determination threshold value and the demagnetization amount (or demagnetization factor) are compared to determine whether or not a demagnetization state. However, two or more determination threshold values and two or more detection counters may be provided to determine a plurality of demagnetization states. As an example, it can be determined that the demagnetization rate is 30% and the weak demagnetization state, and the demagnetization rate is 50% and the strong demagnetization state. As for the confirmation flag F, two types of flags may be output, or information indicating a plurality of demagnetization states may be placed on the signal line. Each system that has received the confirmation flag F executes different processing depending on the degree of the demagnetization state such as the weak demagnetization state or the strong demagnetization state. For example, when it is determined that the state is weakly demagnetized, the corresponding system emits a warning lamp in the vehicle or issues an alarm message. Further, when it is determined that the state is a strong demagnetization state, the traveling control system 70 and the rotating electrical machine control system 20 stop the rotating electrical machine M.

  In the above description, the actual q-axis voltage vq is corrected to obtain the corrected q-axis voltage vqm, and the demagnetization amount is calculated based on the difference between the reference q-axis voltage vqr and the corrected q-axis voltage vqm. did. However, as in the demagnetization determination system 10A shown in FIG. 9, the reference q-axis voltage vqr is corrected to obtain the corrected q-axis voltage vqm, and is reduced based on the difference between the actual q-axis voltage vq and the corrected q-axis voltage vqm. The magnetic quantity may be calculated. That is, the demagnetization determination unit 7 calculates a corrected q-axis voltage vqm obtained by multiplying one of the actual q-axis voltage vq and the reference q-axis voltage vqr by a correction parameter that is a different value depending on the operating state classification. The demagnetization state can be determined by calculating the amount of demagnetization based on the difference between the actual q-axis voltage vq and the reference q-axis voltage vqr. Similarly, the correction parameter indicating the relationship between the DC power supply voltage VC and the q-axis voltage for each operation state category may represent the relationship between the DC power supply voltage VC and the actual q-axis voltage vq, or the DC power supply It may represent the relationship between the voltage VC and the reference q-axis voltage vqr.

  Further, based on the difference between the corrected q-axis voltage vqm obtained by correcting the actual q-axis voltage vq and the reference q-axis voltage vqr, or the corrected q-axis voltage vqm and the actual q-axis voltage vq obtained by correcting the reference q-axis voltage vqr Although it has been described that the amount of demagnetization is calculated based on the difference, the present invention is not limited to these. For example, the reference q-axis voltage vqr corresponding to the operating state classification may be stored in the q-axis voltage map 3 in advance as the one corresponding to the corrected q-axis voltage vqm. In determining the demagnetization, the demagnetization state may be determined by comparing the q-axis voltages without necessarily obtaining the demagnetization amount and the demagnetization factor. For example, since the initial magnetic flux φ is already known, if a threshold value corresponding to the difference in q-axis voltage is set according to the rotational speed ω, the demagnetization determining unit 7 is based on the difference in q-axis voltage. It is possible to determine the state of demagnetization.

  According to the present invention, the demagnetization amount of the permanent magnet can be estimated with high accuracy regardless of the operating state of the permanent magnet embedded type rotating electrical machine. The present invention can be applied to an electric vehicle driven by a permanent magnet embedded rotary electric machine and a hybrid vehicle driven by an internal combustion engine and a rotary electric machine.

1: real q-axis voltage acquisition unit 2: reference q-axis voltage acquisition unit 7: demagnetization determination unit 10, 10A: demagnetization determination system M: motor (rotary electric machine)
Mu, Mv, Mw: multi-phase coil vq: q-axis target voltage, real q-axis voltage vqm: corrected q-axis voltage vqr: reference q-axis voltage

Claims (5)

When the feedback control is performed by converting the current flowing in the multiphase coil of the permanent magnet type rotating electrical machine into a vector component of the d-axis that is the direction of the magnetic field generated by the permanent magnet and the q-axis orthogonal to the d-axis. an actual q-axis voltage acquisition unit that acquires a q-axis target voltage as an actual q-axis voltage;
A reference q-axis voltage acquisition unit configured to acquire, as a reference q-axis voltage, a q-axis target voltage that is set in advance according to the rotation speed and target torque of the rotating electrical machine and in which the demagnetization is not generated in the permanent magnet;
Determining the demagnetization state of the permanent magnet on the basis of an operation state classification indicating whether the operation state of the rotating electrical machine is at least a power running operation state or a regenerative operation state, and the actual q-axis voltage and the reference q-axis voltage A demagnetization determination unit for a rotating electrical machine, comprising:   The demagnetization determination unit corrects either the actual q-axis voltage or the reference q-axis voltage according to the operating state classification, and determines whether the actual q-axis voltage or the reference q-axis voltage is the other. The demagnetization determination system for a rotating electrical machine according to claim 1, wherein the demagnetization state of the permanent magnet is determined by comparison.   The demagnetization determination unit includes a correction q-axis voltage obtained by multiplying one of the actual q-axis voltage and the reference q-axis voltage by a correction parameter that is a different value depending on the operating state classification, The demagnetization determination system for a rotating electrical machine according to claim 2, wherein a demagnetization state is determined by calculating a demagnetization amount based on a difference between an actual q-axis voltage and the reference q-axis voltage.   The operation state classification includes a zero torque operation state in which the output torque of the rotating electrical machine is controlled to be zero in addition to the power running operation state and the regenerative operation state, and the correction parameter varies depending on each operation state classification. The demagnetization determination system for a rotating electrical machine according to claim 3, which is a value.   The correction parameter represents a relationship between any one of the actual q-axis voltage and the reference q-axis voltage and a DC power supply voltage for each of the operation state categories, and the demagnetization determination unit includes the operation state category and the DC The demagnetization determination system for a rotating electrical machine according to claim 3 or 4, wherein the correction parameter is determined based on the power supply voltage.

Images