Motor magnet temperature estimation device and hybrid vehicle equipped with the same

Abstract

To achieve balance between the estimation accuracy of magnet temperature and drive force control of a motor.SOLUTION: A magnet temperature estimation device for a motor 2 comprises an inverter 10 equipped with a plurality of power devices 40, and a magnet temperature estimation unit 34 that estimates the magnet temperature of the motor 2 based on a voltage value corresponding to current flowing through the inverter 10. When estimating the magnet temperature, the magnet temperature estimation unit 34 turns off switching elements 41 of all of the plurality of power devices 40 so as to cut off the power supply to the motor 2, estimates the induced voltage caused by the rotation of the motor 2 based on the voltage value within a period from when the switching elements 41 are turned off until a predetermined waiting time has elapsed, and estimates the magnet temperature based on the correlation between the induced voltage and magnet temperature. The magnet temperature estimation unit 34 also sets the waiting time to increase or decrease depending on a load on the motor 2.SELECTED DRAWING: Figure 15

Description

Technical Field

[0001] The present disclosure relates to a magnet temperature estimation device for a motor and a hybrid vehicle including the same.

Background Art

[0002] For example, Patent Document 1 discloses an example of a magnet temperature estimation device for a motor. Specifically, the device disclosed in Patent Document 1 is configured to estimate an induced voltage generated as the motor rotates and estimate the magnet temperature of the motor based on the magnetic flux density corresponding to the induced voltage.

[0003] Furthermore, the device disclosed in Patent Document 1 includes a voltage sensor that detects the voltage between the neutral point of the motor and the connection cable for energizing the coil, and regards the detection value of this voltage sensor as the induced voltage. This device calculates the magnetic flux density of the magnet based on the frequency spectrum of the detection value regarded as the induced voltage.

Prior Art Documents

Patent Documents

[0004]

Patent Document 1

Summary of the Invention

Problems to be Solved by the Invention

[0005] By the way, when using a voltage sensor as described in Patent Document 1, the detection value includes not only the induced voltage but also the voltage value corresponding to the drive current of the motor. There is room for improving the estimation accuracy of the magnet temperature in the method described in Patent Document 1.

[0006] In addition, in order to perform temperature estimation during the driving of the motor, not only is it necessary to simply improve the estimation accuracy of the magnet temperature, but also a mechanism that minimizes the influence on the driving force control of the motor is required.

[0007] The present disclosure has been made in view of such points, and an object thereof is to achieve a balance between the estimation accuracy of the magnet temperature and the driving force control of the motor.

Means for Solving the Problems

[0008] A first aspect of the present disclosure relates to a magnet temperature estimation device for a motor. This magnet temperature estimation device includes an inverter in which a plurality of power devices including a switching element and a freewheeling diode connected in anti-parallel with the switching element are mounted, a motor electrically connected to the inverter, a battery that supplies power to the motor via the inverter, and a magnet temperature estimation unit that is electrically connected to the inverter and estimates the magnet temperature of the motor based on a voltage value corresponding to the current flowing through the inverter. When estimating the magnet temperature, the magnet temperature estimation unit turns off all the switching elements of the plurality of power devices so as to cut off the power supply to the motor, estimates the induced voltage associated with the rotation of the motor based on the voltage value within a period until a predetermined standby time elapses after turning off the switching element, and estimates the magnet temperature based on the correlation between the induced voltage and the magnet temperature. The magnet temperature estimation unit also sets the standby time so as to increase or decrease according to the load of the motor.

[0009] When the switching element is turned off, the voltage value changes to an induced voltage suitable for estimating the magnet temperature. The changing voltage value includes not only the induced voltage but also those caused by the back electromotive force. On the other hand, the voltage value after the change can be regarded as the induced voltage itself.

[0010] Using the former voltage value can minimize the influence on the drive control of the motor, but since it is not the induced voltage itself, it is disadvantageous for improving the estimation accuracy of the magnet temperature.

[0011] While using the latter voltage value is advantageous for improving the estimation accuracy of the magnet temperature, it is disadvantageous for suppressing the influence on the drive control of the motor.

[0012] However, generally, depending on the motor load based on the driving state of a hybrid vehicle or the like, whether to prioritize the estimation accuracy of the magnet temperature or to prioritize the suppression of the influence on the drive control of the motor can vary. For example, when the motor load is relatively high, it is considered that the suppression of the influence on the drive control of the motor should be prioritized. On the other hand, when the motor load is relatively low (e.g., at no load), the suppression of the influence on the drive control of the motor is not so much of a problem, and improving the estimation accuracy of the magnet temperature can be prioritized.

[0013] Therefore, as in the first aspect, by increasing or decreasing the waiting time for determining the acquisition timing of the voltage value, the balance between the estimation accuracy of the magnet temperature and the drive force control of the motor can be made appropriate according to the motor load at that time.

[0014] Further, according to the second aspect of the present disclosure, after the voltage value is acquired, the magnet temperature estimation unit may turn on the plurality of switching elements so as to resume power supply to the motor.

[0015] According to the second aspect, power supply to the motor can be resumed immediately after the voltage value is acquired. Thereby, the influence on the drive force control of the motor can be suppressed as much as possible.

[0016] Further, according to the third aspect of the present disclosure, when the load of the motor is high, the magnet temperature estimation unit may shorten the waiting time compared to when the load is low.

[0017] As described above, when the motor load is relatively low, it is considered that the suppression of the influence on the drive control of the motor should be prioritized. On the other hand, when the motor load is relatively low (e.g., at no load), the suppression of the influence on the drive control of the motor is not so much of a problem, and improving the estimation accuracy of the magnet temperature can be prioritized.

[0018] By configuring as in the third aspect described above, an appropriate balance according to the load of the motor can be realized.

[0019] Further, according to a fourth aspect of the present disclosure, the magnet temperature estimation unit may estimate the induced voltage by selectively considering or not considering the back electromotive force in the voltage value according to the level of the load of the motor.

[0020] According to the fourth aspect, considering the back electromotive force in the voltage value corresponds to detecting the voltage value in the transient state before reaching the induced voltage. In this case, although it is advantageous in suppressing the influence on the driving force control of the motor, it is disadvantageous in improving the estimation accuracy of the magnet temperature.

[0021] Therefore, by selectively considering or not considering the back electromotive force according to the load of the motor, the balance between the estimation accuracy of the magnet temperature and the driving force control of the motor can be appropriately adjusted according to the motor load at that time.

[0022] Further, according to a fifth aspect of the present disclosure, when the load of the motor is higher than no load, the magnet temperature estimation unit uses, as the voltage value, the voltage value in the transient state after turning off the switching element, and converts the voltage value into the induced voltage based on a model depending on the inductance of the motor. When the load of the motor is no load, the magnet temperature estimation unit uses, as the voltage value, the voltage value after transitioning from the transient state to the steady state, and regards the voltage value as the induced voltage.

[0023] Using the voltage value in the transient state is advantageous in suppressing the influence on the driving force control of the motor as described above, but may be disadvantageous in improving the estimation accuracy of the magnet temperature in that a model depending on the inductance is used.

[0024] However, when the motor is unloaded, it is considered that there is no need to consider its influence on the driving force control anymore. Therefore, as in the fifth aspect, by properly using voltage values according to whether the motor is unloaded or not, it is possible to appropriately balance the estimation accuracy of the magnet temperature and the driving force control of the motor.

[0025] Also, according to a sixth aspect of the present disclosure, when an allowable current value is set as a current value that serves as an index for magnetic saturation of the coil of the motor, which is set corresponding to the specifications of the motor, the magnet temperature estimation unit, when the load of the motor is higher than no load, when the load of the motor is relatively high, estimates the induced voltage based on the voltage value acquired before the transient current accompanying the turn-off of the switching element drops to the allowable current value and the model, and when the load of the motor is relatively low, estimates the induced voltage based on the voltage value acquired after the transient current drops to the allowable current value and the model. This may be the case.

[0026] Using the voltage value detected before it drops to the allowable current value is particularly advantageous in suppressing the influence on the driving force control of the motor, but it is disadvantageous in enhancing the estimation accuracy of the magnet temperature due to the influence of magnetic saturation on the value of inductance.

[0027] For example, when the load of the motor is relatively high, it is considered that it is required to further suppress the influence on the driving force control of the motor compared to when it is relatively low.

[0028] Therefore, by properly using the detection timing of the voltage value as in the sixth aspect, it is possible to make the balance between the estimation accuracy of the magnet temperature and the driving force control of the motor more appropriate.

[0029] Further, according to a seventh aspect of the present disclosure, the motor is mounted on a hybrid vehicle including an engine that cooperates with the motor, and when the driving torque of the hybrid vehicle is output only from the motor, the magnet temperature estimation unit determines that the load on the motor is relatively high, and when the driving torque is output from both the motor and the engine, the magnet temperature estimation unit determines that the load on the motor is relatively low, which may be acceptable.

[0030] When the driving torque of a hybrid vehicle is output only from the motor, it is considered that the influence on the driving force control of the motor should be suppressed as much as possible. On the other hand, when the driving torque of a hybrid vehicle is output from both the engine and the motor, it is considered that the suppression of the influence on the driving force control of the motor is not required as much as in the former case.

[0031] Therefore, by causing the magnet temperature estimation unit to make a determination as in the seventh aspect, the balance between the estimation accuracy of the magnet temperature and the driving force control of the motor can be made more appropriate.

[0032] Further, according to an eighth aspect of the present disclosure, when the load on the motor is no load, the magnet temperature estimation unit corrects the model so that the induced voltage based on the model approaches the induced voltage based on the voltage value obtained after the transition to the steady state.

[0033] According to the eighth aspect, even when the influence of the back electromotive force cannot be ignored, the estimation accuracy of the magnet temperature can be improved by appropriately correcting the model.

[0034] Further, according to a ninth aspect of the present disclosure, the magnet temperature estimation unit further has a dynamic thermal model for calculating the magnet temperature, calculates an error and a Kalman gain between an estimated magnet temperature, which is the magnet temperature estimated by the magnet temperature estimation unit, and a calculated magnet temperature, which is the magnet temperature calculated by the dynamic thermal model, corrects the dynamic thermal model by introducing the error and the Kalman gain, and adopts a calculated value by the corrected dynamic thermal model as the magnet temperature, which may be acceptable.

[0035] The magnet temperature (estimated magnet temperature) estimated by the magnet temperature estimation unit includes switching noise. On the other hand, since the dynamic thermal model is not affected by switching, the calculated magnet temperature does not include switching noise. Therefore, using the method of a known Kalman filter, the dynamic thermal model is corrected based on the error between these estimated magnet temperatures and calculated magnet temperatures and the Kalman gain, and the calculated value by the corrected dynamic thermal model is adopted as the temperature of the power device.

[0036] In this way, the dynamic thermal model is sequentially corrected, and the calculated magnet temperature converges to the true value. Therefore, the magnet temperature estimation unit can estimate the magnet temperature with reduced switching noise. Accordingly, the estimation accuracy of the magnet temperature is improved.

[0037] Also, according to the tenth aspect of the present disclosure, the magnet temperature estimation unit further has a dynamic thermal model for calculating the magnet temperature, gain setting data related to a specific output related to the noise of the inverter, and a gain corresponding to the dynamic thermal model. The error between the estimated magnet temperature, which is the magnet temperature estimated by the magnet temperature estimation unit, and the calculated magnet temperature, which is the magnet temperature calculated by the dynamic thermal model, is calculated, the gain is obtained from the gain setting data, and the dynamic thermal model is corrected by introducing the error and the gain, and the calculated value by the corrected dynamic thermal model is adopted as the magnet temperature.

[0038] Calculating the Kalman gain K has a large computational load. Therefore, there is a possibility of a delay in responsiveness, and the hardware becomes large. On the other hand, if it is configured to predict the change in switching noise in advance and thereby change the gain (so-called feedforward control), the calculation of the Kalman gain becomes unnecessary. Therefore, together with the improvement of the estimation accuracy of the magnet temperature, the computational load can be reduced and high responsiveness can be realized.

[0039] Further, according to the 11th and 12th aspects of the present disclosure, the specific output may be the current value output by the inverter, and the gain setting data may be set such that the gain decreases as the current value increases. Alternatively, the specific output may be the carrier frequency of the inverter, and the gain setting data may be set such that the gain decreases as the carrier frequency decreases.

[0040] That is, the switching noise increases as the current output by the inverter increases, and the switching noise decreases as the current output by the inverter decreases. Also, the switching noise increases as the carrier frequency of the inverter decreases, and the switching noise decreases as the carrier frequency of the inverter increases. That is, the current value output by the inverter and the carrier frequency of the inverter are correlated with the change in the switching noise.

[0041] Therefore, by changing the gain in accordance with these changes, an appropriate gain can be obtained predictively.

[0042] Further, the 13th aspect of the present disclosure relates to a hybrid vehicle including the magnet temperature estimation device, the motor, and an engine that cooperates with the motor.

Advantages of the Invention

[0043] As described above, according to the present disclosure, it is possible to achieve a balance between the estimation accuracy of the magnet temperature and the driving force control of the motor.

Brief Description of the Drawings

[0044]

Figure 1

Figure 2

Figure 3

Figure 4

Figure 5

Figure 6

Figure 7

Figure 8

Figure 9

Figure 10

Figure 11

Figure 12

Figure 13

Figure 14

Figure 15

Figure 16

Figure 17

MODE FOR CARRYING OUT THE INVENTION

[0045] Hereinafter, embodiments of the present disclosure will be described with reference to the drawings. Note that the following description is an example.

[0046] (1) Moving body (1-1) Overall configuration FIG. 1 is a schematic diagram illustrating the configuration of an automobile 1 as a moving body. The automobile 1 shown in FIG. 1 is a four-wheel hybrid vehicle. The automobile 1 as a hybrid vehicle includes, as a drive source, a motor 2 and an engine 3 that cooperates with the motor 2. The motor 2 and the engine 3 cooperate with each other to rotationally drive drive wheels 9, 9 located, for example, on the rear side of the vehicle body among the four wheels. By this rotational drive, the automobile 1 moves (travels).

[0047] The motor 2 is electrically connected to an inverter 10. The motor 2 is a motor that functions as a drive source of the automobile 1 together with the engine 3. The motor 2 is a permanent magnet synchronous motor.

[0048] An example of the motor 2 is shown in FIG. 2. The motor 2 generally includes a motor case 21, a shaft 22, a rotor 23, and a stator 24.

[0049] The rotor 23 has a plurality of magnets 25 and is configured to output rotational power. The plurality of magnets 25 includes a first magnet 25a that directs an S pole toward the stator 24 and a second magnet 25b that is adjacent to the first magnet 25a in the rotational direction and directs an N pole toward the inner peripheral surface of the stator 24. In the following description, the first magnet 25a is referred to as the S magnet 25a, and the second magnet 25b is referred to as the N magnet 25b.

[0050] The stator 24 has a plurality of coils 26 arranged to face the rotor 23 with a gap in the radial direction. Specifically, the stator 24 according to the present embodiment has a stator core 24a formed by laminating a plurality of metal plates and a plurality of coils 26 formed by winding electric wires around the stator core 24a.

[0051] The plurality of coils 26 constitute a three-phase coil group consisting of a U-phase, a V-phase, and a W-phase with different phases of the flowing current. Each coil group is arranged in order along the rotation direction. For example, in the configuration shown in FIG. 2, the plurality of coils 26 consist of a total of 12 coils 26. The 12 coils 26 are divided into a U-phase coil group consisting of 4 coils 26, a V-phase coil group consisting of 4 coils 26, and a W-phase coil group consisting of 4 coils 26.

[0052] Hereinafter, the coil 26 belonging to the U-phase coil group may be referred to as the U-phase coil 26, the coil 26 belonging to the V-phase coil group may be referred to as the V-phase coil 26, and the coil 26 belonging to the W-phase coil group may be referred to as the W-phase coil 26.

[0053] Note that although FIG. 2 illustrates the motor 2 with 8 poles and 12 slots, the configuration of the motor 2 is not limited to this example. The motor 2 may be configured to have more pole numbers and slot numbers. For example, taking N as an integer, the motor 2 can be configured with 2×N magnets 25 and 3×N slots.

[0054] To energize the coils 26, three output wirings (connection cables) 14, 14, 14 are connected to these coils 26. The three connection cables 14, 14, 14 include a connection cable 14 connected to the U-phase coil group, a connection cable 14 connected to the V-phase coil group, and a connection cable 14 connected to the W-phase coil group. These connection cables 14, 14, 14 are led out to the outside of the motor case 21 and connected to the battery 4 via an inverter 10 described later. When the vehicle 1 is a hybrid vehicle, the battery 4 may be a battery with a rated voltage of 50 V or less, specifically, a 48 V DC battery (low-voltage battery).

[0055] When an alternating current is applied to the coil 26, the coils 26 forming each coil group generate a magnetic field. Then, the coils 26 that have generated the magnetic field and the magnets 25 exert an attractive or repulsive force on each other, causing the rotor 23 to be rotationally driven. The rotationally driven rotor 23 outputs rotational power via the shaft 22. This rotational power is input to the transmission 8 via the shaft 22 or the like.

[0056] In addition, a motor cooling system 2a shown in FIG. 1 is attached to the motor 2. The motor cooling system 2a cools the motor 2 by circulating and supplying a refrigerant.

[0057] The inverter 10 is connected to the battery 4 via the high-voltage wiring 5. The inverter 10 is also connected to the controller 7 via the low-voltage wiring 6 for control. As will be described later, a power device 40 including a switching element 41 and a freewheeling diode 42 connected in anti-parallel with the switching element 41 is mounted on the inverter 10.

[0058] The battery 4 supplies power to the motor 2 via the inverter 10. Specifically, the battery 4 is composed of, for example, a plurality of lithium-ion batteries. The rated voltage of the battery 4 is as described above. The battery 4 supplies high-voltage DC power to the inverter 10. The inverter 10 is PWM-controlled by the controller 7, converts the DC power supplied from the battery 4 into a three-phase alternating current, and supplies it to the motor 2. Thereby, the motor 2 rotates.

[0059] A transmission 8 is connected to the rear of the motor 2. When the vehicle 1 is running, the rotational power output by the motor 2 is transmitted to the drive wheels 9 via the propeller shaft, differential gear, and a pair of drive shafts after being shifted by the transmission 8.

[0060] The controller 7 constitutes the control system of the vehicle 1. This controller 7 is composed of one or more of an engine control unit (ECU), a motor control unit (MCU), a transmission control unit (TCU), a brake control (BCU), and a general control unit (GCU).

[0061] Various sensors are electrically connected to the control system of the vehicle 1. The controller 7 determines the loads on the motor 2 and the engine 3 based on, for example, the detection signal of the accelerator opening sensor 7a. The accelerator opening sensor 7a is shown only in FIG. 1.

[0062] (1-2) Inverter 10 As shown in a simplified manner in FIG. 1, the inverter 10 has a power module 20 including a substrate 20a and a radiator 20b mounted thereon. Further connected to the power module 20 is a magnet temperature estimation device 30 for the motor (this will be described separately later). Six power devices 40 are installed on the surface of the substrate 20a. As shown in FIG. 3, each of these power devices 40 includes a switching element 41 and a freewheeling diode 42 connected in anti-parallel with this switching element 41.

[0063] A known inverter circuit is provided on the substrate 20a. The inverter circuit converts the DC power of the battery 4 into AC power of three phases (U phase, V phase, W phase) and outputs it to the motor 2 rotating with three-phase AC. The inverter circuit is provided with a positive-side wiring 11 connected to the positive side of the battery 4 and a negative-side wiring 12 connected to the negative side of the battery 4. And between these positive-side wiring 11 and negative-side wiring 12, three circuits (half-bridge circuits 13) corresponding to each phase are connected in parallel.

[0064] Two power devices 40, 40 are connected in series to each of these half-bridge circuits 13, 13, 13. And an output wiring 14 for outputting AC power to the motor 2 is connected between these two power devices 40, 40. Each output wiring 14 is connected to the coil 26 of the motor 2. A smoothing capacitor 15 is connected in parallel with these half-bridge circuits 13, 13, 13 between the positive electrode side wiring 11 and the negative electrode side wiring 12 on the input side rather than these half-bridge circuits 13, 13, 13.

[0065] The freewheeling diode 42 is provided to prevent the switching element 41 from being damaged by the instantaneous back electromotive force generated when the corresponding switching element 41 is turned off. Therefore, when the corresponding switching element 41 is turned off, a forward voltage is applied to the freewheeling diode 42 and a forward current flows.

[0066] The switching element 41 that turns on and off at high speed and intermittently cuts off a large current generates heat. Therefore, in order to cool the power device 40, as shown in FIG. 1, a heat sink 20b is attached to the back surface of the substrate 20a. The heat sink 20b is configured to be cooled by a refrigerant circulated and supplied from a cooling device (not shown).

[0067] The exemplary power device 40 shows a MOSFET (Metal-Oxide-Semiconductor Field-Effect Transistor) or an RC-IGBT (Reverse Conducting IGBT) in which the switching element 41 and the freewheeling diode 42 are integrally formed.

[0068] (1-3) Magnet temperature estimation device 30 As shown in FIG. 1, the magnet temperature estimation device 30 (also simply referred to as "temperature estimation device 30") of the present embodiment is provided in the power module 20 in a form independent of the substrate 20a. Specifically, the temperature estimation device 30 is attached to the power module 20 in a state of being externally attached to the substrate 20a.

[0069] Note that a part of the temperature estimation device 30 may be incorporated into the substrate 20a and configured integrally with the substrate 20a. Also, it is not essential to provide all of the temperature estimation device 30 in the power module 20. A part of the temperature estimation device 30, for example, the magnet temperature estimation unit 34 described later, may be implemented as an element (functional block) of the controller 7 such as the aforementioned MCU.

[0070] As shown in FIG. 3, the temperature estimation device 30 includes a diode current detection unit 31, a voltage division unit 32, a diode voltage detection unit 33, an AD converter 35, a magnet temperature estimation unit 34, and a thermal model estimation unit 38. The voltage division unit 32 includes a voltage division diode 32a (first element) and a voltage division resistor 32b (second element). The diode voltage detection unit 33 includes a protection diode 33a and a transformer unit 33b.

[0071] This temperature estimation device 30 estimates the temperature of the magnet 25 of the motor 2 based on the induced electromotive force related to the U-phase coil 26. Specifically, the temperature estimation device 30 estimates the temperature of the magnet 25 of the motor 2 based on the diode current and diode voltage related to the U-phase power device 40. Note that similar temperature estimation may be performed for the power devices 40 of the V-phase and W-phase.

[0072] (Diode Current Detection Unit 31) The diode current detection unit 31 detects the forward current flowing through the freewheeling diode 42 of the power device 40 to be detected. The diode current detection unit 31 may be any device that can measure a current value, such as a current detector or a current sensor.

[0073] The diode current detection unit 31 is installed, for example, at a predetermined location on the wiring on the upstream side or downstream side of the power device 40 to be detected as shown in FIG. 3. Then, during the period when the switching element 41 is off, the current (forward current) flowing through the freewheeling diode 42 is detected.

[0074] (Voltage Division Unit 32) The voltage dividing section 32 has a function of dividing the high forward voltage of the reflux diode 42 into two at a predetermined ratio. The voltage dividing diode 32a (first element) and the voltage dividing resistor 32b (second element) are connected in series with each other and are connected in parallel with the reflux diode 42 to be estimated. The voltage dividing diode 32a is arranged in a forward parallel configuration in the same direction as the reflux diode 42. That is, it is arranged in a direction that is reverse parallel to the switching element 41.

[0075] The voltage dividing diode 32a has a slightly lower forward voltage than the reflux diode 42 when compared with the same forward current. The magnitude of the voltage dividing resistor 32b is selected corresponding to the forward voltage of the voltage dividing diode 32a.

[0076] These voltage dividing diode 32a and voltage dividing resistor 32b constitute a voltage dividing circuit, and the forward voltage of the reflux diode 42 is detected using this voltage dividing circuit. That is, the diode voltage detection section 33 is connected in parallel with the voltage dividing resistor 32b, and detects the forward voltage of the reflux diode 42 based on the voltage division of the voltage dividing resistor 32b.

[0077] (Diode voltage detection section 33) The diode voltage detection section 33 is composed of a protection diode 33a, a transformer section 33b, etc. as described above. The diode voltage detection section 33 has a function of outputting the voltage division of the voltage dividing resistor 32b as an analog signal of a predetermined magnitude in a state insulated from the circuit of the inverter 10.

[0078] The protection diode 33a is arranged in a direction opposite to that of the reflux diode 42 and the voltage dividing diode 32a and is connected in parallel with the voltage dividing resistor 32b. Since the forward voltage of the reflux diode 42 is high, if the voltage dividing diode 32a fails and short - circuits, the high voltage will act on the voltage dividing resistor 32b. In such a case, there is a risk that the temperature estimation device 30 will be damaged.

[0079] On the other hand, by providing such a protection diode 33a, even if the voltage-dividing diode 32a is short-circuited, current can flow from the positive electrode side to the negative electrode side through the protection diode 33a. Therefore, it is possible to prevent an abnormal high voltage from being applied to the voltage-dividing resistor 32b. And it is possible to prevent an abnormal large current from flowing through the transformer section 33b. Therefore, the temperature estimation device 30 can be protected.

[0080] The AD converter 35 is connected to the voltage-dividing section 32 via the transformer section 33b. The AD converter 35 converts an analog signal of voltage or current input from the transformer section 33b into a digital signal. A load resistor 36 is connected in parallel between the AD converter 35 and the transformer section 33b.

[0081] The transformer section 33b consists of a known pulse transformer connected in parallel with the voltage-dividing resistor 32b, and the AD converter 35 is separated from the voltage-dividing section 32 by the transformer section 33b. Therefore, the AD converter 35 can be made independent (insulated) from a circuit of a power system with a high voltage of several hundred volts. Even if the power of the voltage-dividing section 32 becomes abnormal, it will not directly affect the AD converter 35.

[0082] Moreover, the protection diode 33a prevents a large current from flowing through the transformer section 33b. Therefore, since the transformer section 33b does not need to be prepared for a large current, it can be miniaturized. And since only a relatively small current corresponding to the voltage division of the voltage-dividing resistor 32b flows through the transformer section 33b, the transformer section 33b can increase to an appropriate voltage for the AD converter 35 with a relatively small turns ratio.

[0083] Specifically, the voltage that the AD converter 35 can tolerate is generally about ±5V. Therefore, the voltage Vad on the secondary side of the transformer unit 33b needs to be within that range. In contrast, the transformer unit 33b converts the voltage on its primary side (i.e., the divided voltage Vrf of the voltage divider resistor 32b) according to the ratio of the number of turns on the primary side (T1) of the transformer unit 33b connected in parallel with the voltage divider resistor 32b to the number of turns on the secondary side (T2) of the transformer unit 33b connected to the AD converter 35 (turns ratio Tvf: T2 / T1), and outputs it to the secondary side (Vad = Tvf·Vrf).

[0084] Since the current flowing through the primary side of the transformer unit 33b is relatively small, by setting the turns ratio Tvf to about several times (for example, 4 times), the voltage on the primary side can be increased and output to the secondary side. Thus, the voltage on the secondary side of the transformer unit 33b can be within the range of about ±5V. Therefore, the divided voltage Vrf of the voltage divider resistor 32b, and thus the forward voltage Vf of the freewheeling diode 42, can be stably and accurately detected.

[0085] (Magnet temperature estimation unit 34) The magnet temperature estimation unit 34 is composed of an IC chip or the like onboard the AD converter 35. That is, the magnet temperature estimation unit 34 is provided in the AD converter 35 as a functional configuration.

[0086] Specifically, the magnet temperature estimation unit 34 is composed of known hardware such as a processor and a memory provided in the AD converter 35, and software such as programs and data implemented in the memory. The magnet temperature estimation unit 34 estimates the magnet temperature of the motor 2 using the detected forward current if and forward voltage Vf through the cooperation of these hardware and software.

[0087] Hereinafter, the method for estimating the magnet temperature by the magnet temperature estimation unit 34 will be described in detail.

[0088] (2) Method for estimating magnet temperature (2-1) Estimation principle For example, during the driving of the motor 2, for any one of the U-phase, V-phase, and W-phase, the switching element 41 connected to the positive electrode side of the battery 4 is turned on, and the switching element 41 connected to the negative electrode side of the battery is turned off. At the same time, for the remaining two of the U-phase, V-phase, and W-phase, the switching element 41 connected to the positive electrode side of the battery is turned off, and the switching element 41 connected to the negative electrode side of the battery is turned on.

[0089] Suppose that the former is the U-phase and the latter are the V-phase and W-phase. For the U-phase, current will flow from the battery 4 towards the coil 26 of the motor 2 through the positive electrode side wiring 11 and the positive electrode side switching element 41. On the other hand, for the V-phase and W-phase, current will return from the coil 26 towards the battery 4 through the negative electrode side wiring 12 and the negative electrode side switching element 41.

[0090] Here, during the driving of the motor 2, assume that all the switching elements 41 on both the positive electrode side and the negative electrode side for the U-phase, V-phase, and W-phase are turned off. This is equivalent to turning off the switching element 41 so as to cut off the power supply to the motor 2.

[0091] In this case, regarding the current supplied to the motor 2, it is assumed that it is temporarily maintained due to the action of inductance like an elastic body in a mechanical system. At this time, since the current supplied to the motor 2 cannot flow through the positive electrode side freewheeling diode 42, it flows towards the coil 26 through the negative electrode side wiring 12 and the negative electrode side freewheeling diode 42 (refer to the dashed arrow in Figure 3). On the other hand, since the current returning from the motor 2 cannot flow through the negative electrode side freewheeling diode 42, it returns to the battery 4 through the positive electrode side wiring 11 and the positive electrode side freewheeling diode 42 as the forward current described above (refer to the dashed arrow in Figure 3).

[0092] Also, in this case, the forward current decreases exponentially according to the time constant τ (= L / R) obtained by dividing the inductance L of coil 26 by the winding resistance R. During the decrease of the forward current, the corresponding forward voltage is the sum of the back electromotive force associated with the change in current and the induced voltage generated by the magnetic flux emitted from magnet 25 rotating with rotor 23 passing through coil 26.

[0093] Note that the period during which all switching elements 41 are turned off is actually a short period. Therefore, even if the forward current decreases to zero, rotor 23 itself will continue to rotate. Also, when the decrease of the forward current stops and reaches a certain value, the back electromotive force generated due to the current change becomes zero. Therefore, the forward voltage detected after the forward current becomes zero is the induced voltage itself without including the back electromotive force. Repeating, the forward voltage after the forward current becomes zero can be regarded as the induced voltage.

[0094] On the other hand, as described in Japanese Patent Application Laid-Open No. 2021-118653 and the like, the applicant of the present application has already publicly disclosed a method for estimating the magnet temperature based on the induced voltage. Using the forward voltage detected at the timing after the forward current becomes zero corresponds to using a highly accurate induced voltage without including the back electromotive force. Using a highly accurate induced voltage contributes to highly accurate estimation of the magnet temperature.

[0095] However, turning off all switching elements 41 may affect the drive control of motor 2 even if it is a short period. For example, cutting off the power supply when motor 2 is operating at a high load is inconvenient because it may affect the riding comfort of automobile 1.

[0096] On the one hand, if the forward voltage detected at the timing before the forward current becomes zero is regarded as the induced voltage, the influence on the drive control of the motor 2 is suppressed because the off period of the switching element 41 becomes relatively short. However, in this case, since the forward voltage also includes the influence of the back electromotive force, it is inconvenient for using a highly accurate induced voltage and thus for highly accurately estimating the magnet temperature.

[0097] Therefore, in the present embodiment, it is decided to selectively use the detection timing of the forward voltage according to the load of the motor 2. Further, even when using a forward voltage in which the influence of the back electromotive force cannot be ignored, by modeling the influence of the back electromotive force, it has been possible to achieve high accuracy of the induced voltage.

[0098] First, let the value of the induced voltage be V emf and the value of the forward voltage be V p Then, when the influence of the back electromotive force can be ignored, the induced voltage V emf is the forward voltage V p itself as described above. This can be expressed as the following formula (1).

[0099] V emf =V p …(1) On the other hand, when the influence of the back electromotive force cannot be ignored, the forward voltage V p is expressed by adding the value of the back electromotive force to the induced voltage V emf . The value of the back electromotive force is described by Faraday's law, and its magnitude is proportional to the time derivative of the forward current. This forward current, for example, when the current value (initial current) at the time when the switching element 41 is turned off is I0, decreases exponentially from that current value I0 according to the time constant τ based on the inductance L and the winding resistance R. In this case, by transforming the formula representing the forward voltage V p into a formula representing the induced voltage V emf , it can be modeled as shown in the following formula (2).

[0100] V emf =Ka ·V p +K b ·I0 + K c …(2) In the above formula (2), the parameter K a , K b , K c are constants based on the inductance L and the winding resistance R, respectively.

[0101] FIG. 4 is a diagram illustrating the forward current (upper part) and the forward voltage (lower part) in the transient state after the switching element 41 is turned off. As illustrated in FIG. 4, in the first region R1 described later, the forward voltage V p will only include the induced voltage in the steady state. On the other hand, in the second region R2 and the third region R3 of the figure, the forward voltage V p will include both the induced voltage and the back electromotive force..

[0102] Therefore, in the present embodiment, the detection timing of the forward voltage is used properly according to the load of the motor 2, and equations (1) and (2) are used properly according to the detection timing.

[0103] In other words, the magnet temperature estimation unit 34 according to the present embodiment can estimate the induced voltage by properly determining whether to consider the back electromotive force in the value (voltage value) of the forward voltage according to the level of the load of the motor 2.

[0104] Note that, to calculate the value of the back electromotive force, the time change of the magnetic flux linkage of the coil 26 is used. This time change is the origin of the time differentiation of the forward current described above. Here, the magnetic flux linkage is usually proportional to the product of the inductance L of the coil 26 and the current flowing through the coil 26 (forward current). That is, the magnetic flux linkage increases linearly as the forward current increases.

[0105] However, as shown in FIG. 5, when the current flowing through coil 26 is higher than a predetermined allowable current value Is, magnetic saturation occurs in the linked magnetic flux. In this case, the linked magnetic flux does not increase linearly but increases more slowly. This can be regarded as a change in the inductance L of coil 26 or as causing a relatively large error in the aforementioned linear calculation.

[0106] If the initial current I0 is larger than the allowable current value Is, the accuracy of the model formula represented by Equation (2) will change according to whether the value of the forward current is below the allowable current value Is.

[0107] The magnet temperature estimation unit 34 according to this embodiment, according to the level of the load of the motor 2, the current value V detected after the value of the forward current falls below the allowable current value Is p and Equation (2) are used to estimate the induced voltage V emf or the current value V detected before the value of the forward current falls below the allowable current value Is p and Equation (2) are used to estimate the induced voltage V emf can be selectively used.

[0108] (2-2) Specific Configuration Similar to other elements constituting the magnet temperature estimation device 30, the magnet temperature estimation unit 34 is electrically and directly or indirectly connected to the inverter 10. The magnet temperature estimation unit 34 estimates the magnet temperature of the motor 2 based on the voltage value corresponding to the current flowing through the inverter 10.

[0109] Specifically, as shown in FIG. 6, the magnet temperature estimation unit 34 according to this embodiment includes a motor load determination unit 341, a standby time determination unit 342, a power supply cutoff unit 343, an induced voltage estimation unit 344, an induced voltage conversion unit 345, a power supply restart unit 346, and a model correction unit 347.

[0110] (2-2-1) Motor Load Determination Unit 341 The motor load determination unit 341 is connected to be able to transmit and receive signals to and from the controller 7 of the automobile 1. The motor load determination unit 341 receives a signal indicating the level of the load on the motor 2.

[0111] Specifically, the motor load determination unit 341 determines whether the load on the motor 2 is higher than no load or is no load. This determination can be made, for example, based on a signal generated by the controller 7. This signal is generated, for example, based on the required torque of the automobile 1, the driving state, the driving situation, the accelerator pedal opening, the driver's manual operation, etc. Specifically, the motor load determination unit 341 determines that the motor 2 is under no load when the automobile 1 is idling, coasting, or shifting gears.

[0112] More specifically, when the load on the motor 2 is higher than no load, the motor load determination unit 341 determines whether the load is relatively high or relatively low. This determination is made, for example, based on the signal generated as described above by the controller 7.

[0113] Specifically, the motor load determination unit 341 determines that the load on the motor 2 is relatively high when the driving torque of the automobile 1 is output only from the motor 2. Also, the motor load determination unit 341 determines that the load on the motor 2 is relatively low when the driving torque of the automobile 1 is output from both the motor 2 and the engine 3.

[0114] In addition, the motor load determination unit 341 may determine that the load on the motor 2 is relatively high when the automobile 1 is climbing a slope, towing, or traveling at high speed.

[0115] In this embodiment, the motor load determination unit 341 classifies the states into three categories: when the motor 2 is unloaded, when the load on the motor 2 is higher than no load and relatively low, and when the load on the motor 2 is higher than no load and relatively high. Hereinafter, these three classifications are referred to as the first region R1, the second region R2, and the third region R3, respectively. Note that the three classifications are not essential. For example, any two of the first region R1, the second region R2, and the third region R3 may be used.

[0116] As shown in FIG. 4, these classifications affect the selection of the detection timing (waiting time) of the current value.

[0117] (2-2-2) Waiting time determination unit 342 The waiting time determination unit 342 sets the waiting time so as to increase or decrease according to the load on the motor 2. This waiting time is the elapsed time from when all the switching elements 41 of the plurality of power devices 40 are turned off until the forward voltage is acquired. This is equal to the elapsed time from T0 in FIG. 4. For example, the second waiting time Tw2 described later is the time interval from time T0 in FIG. 4 to the arrow Tw2. Regarding the timing (time T0) when all the switching elements 41 are turned off as a reference, the same applies to the “saturation time Ts” described later.

[0118] Note that the waiting time may be made to coincide with the off period of the switching element 41. In that case, as soon as the forward voltage is acquired, the power supply to the motor 2 is immediately restarted. The waiting time may be at least a time interval equal to or less than the off period.

[0119] Specifically, when the load on the motor 2 is high, the waiting time determination unit 342 sets a shorter waiting time than when the load is low. Specifically, the waiting time determination unit 342 sets the waiting time so that it is selectively determined whether or not the back electromotive force is considered in the acquired value (voltage value) of the forward voltage according to the level of the load on the motor 2.

[0120] More specifically, the standby time determination unit 342 sets the standby time such that only the induced voltage is considered in the first region R1, while both the induced voltage and the counter electromotive force are considered in the second region R2 and the third region R3.

[0121] For example, in the case of the first region R1 corresponding to the no-load state of the motor 2, the standby time determination unit 342 selects, as the standby time, a relatively long first standby time Tw1 that matches the elapsed time until the forward current decreases to zero (until the forward voltage becomes flat) after turning off all the switching elements 41, or is longer than the elapsed time.

[0122] The first standby time Tw1 selected in the first region R1 may be set to be longer than a time constant τ determined based on, for example, the winding resistance R of the coil 26 and the inductance L of the coil 26 in the linear region described later. The winding resistance R of the coil 26 is determined according to the specifications of the motor 2 and is stored in advance in the magnet temperature estimation unit 34.

[0123] On the other hand, in the case of the second region R2 or the third region R3 corresponding to the non-no-load state of the motor 2, the standby time determination unit 342 sets the standby time to be relatively short so that it is shorter than the elapsed time from when all the switching elements 41 are turned off until the forward current decreases to zero.

[0124] The standby time (second standby time Tw2) selected in the second region R2 is set to be, for example, at least longer than a saturation time Ts that is the boundary at which the coil 26 of the motor 2 undergoes magnetic saturation, as shown in FIG. 4. This saturation time Ts is shorter than the aforementioned time constant τ. This saturation time Ts can also be regarded as a boundary value at which the forward current falls below the allowable current value Is.

[0125] On the other hand, the standby time (third standby time Tw3) in the third region R3 is set to be, for example, at least shorter than the saturation period Ts, as shown in FIG. 4.

[0126] (2-2-3) Power supply cut-off unit 343 When estimating the magnet temperature, the power supply cut-off unit 343 turns off all the switching elements 41 of the plurality of power devices 40 so as to cut off the power supply to the motor 2. The power supply cut-off unit 343 counts the elapsed time from the time point (time T0) when it is turned off by a timer (not shown) or the like. At that time, the magnet temperature estimation unit 34 acquires the current value of the forward current if at the time point when it is turned off, that is, the value of the initial current I0.

[0127] (2-2-4) Induced voltage estimation unit 344 The induced voltage estimation unit 344 estimates the induced voltage associated with the rotation of the motor 2 based on the voltage value (forward voltage V p ) during the period from when the switching element 41 is turned off until a predetermined standby time elapses.

[0128] Specifically, the induced voltage estimation unit 344 detects the forward voltage V p during the standby time, and estimates the induced voltage based on the forward voltage V p . Note that, before the standby time elapses, a plurality of forward voltages V p from when the standby time is reached may be detected, and the induced voltage may be estimated based on their moving average. When using the moving average, among the first region R1, the second region R2, and the third region R3, the forward voltage V p within the region to which the standby time determined by the standby time determination unit 342 belongs may be used.

[0129] Also, when the load of the motor 2 belongs to the first region R1, that is, when the load of the motor 2 is no load, the induced voltage estimation unit 344 uses the first standby time Tw1 set to be longer than other regions as described above. In this case, the induced voltage estimation unit 344 detects the voltage value after transitioning from the transient state to the steady state after the switching element 41 is turned off as the forward voltage V p . The induced voltage estimation unit 344 regards that voltage value as the induced voltage based on the above formula (1).

[0130] Further, when the load of the motor 2 belongs to the second region R2 or the third region R3, that is, when the load of the motor 2 is higher than no load, the induced voltage estimation unit 344 uses the second standby time Tw2 set shorter than the first region R1 as described above. In this case, the induced voltage estimation unit 344 uses the forward voltage V p as the voltage value in the transient state will be detected. The induced voltage estimation unit 344 converts the detected voltage value into an induced voltage based on a model depending on the inductance L of the motor 2 represented by the above formula (2).

[0131] Specifically, when the load of the motor 2 belongs to the third region R3 where the load is relatively high, the induced voltage estimation unit 344 uses the third standby time Tw3 set shorter than the first region R1 and the second region R2 as described above. In this case, the induced voltage estimation unit 344 uses the forward voltage V p as the voltage value before the transient current accompanying the turn-off of the switching element 41 decreases to the allowable current value Is will be detected. Magnetic saturation affects this voltage value via the current value. The induced voltage estimation unit 344 estimates the induced voltage based on the detected voltage value and the model represented by the formula (2).

[0132] On the other hand, when the load of the motor 2 belongs to the second region R2 where the load is relatively low, the induced voltage estimation unit 344 uses the second standby time Tw2 set shorter than the first region R1 and longer than the third region R3 as described above. In this case, the induced voltage estimation unit 344 uses the forward voltage V p as the voltage value after the transient current has decreased to the allowable current value Is will be detected. Magnetic saturation does not affect this voltage value. The induced voltage estimation unit 344 estimates the induced voltage based on the detected voltage value and the model represented by the formula (2).

[0133] (2-2-5) Induced voltage conversion unit 345 The induced voltage conversion unit 345 estimates the magnet temperature based on the correlation between the induced voltage and the magnet temperature. Specifically, the induced voltage conversion unit 345 according to the present embodiment estimates the magnet temperature based on the frequency spectrum corresponding to a predetermined frequency among the frequency components constituting the induced voltage.

[0134] FIG. 7 is a diagram illustrating the amplitude and phase obtained by performing a fast Fourier transform on the induced voltage. FIG. 8 is a graph illustrating the relationship between the magnetic flux density of the magnet 25 and the magnet temperature.

[0135] Hereinafter, the induced voltage Vu generated in the U-phase coil 26 will be taken as an example for explanation. However, the following explanation is the same for the V-phase coil 26 and the W-phase coil 26.

[0136] The graph G4 shown in the left part of FIG. 7 shows the induced voltage Vu generated in the U-phase coil 26. On the other hand, the bar graph G5 shown in the right part of FIG. 7 shows the absolute value (amplitude) of the frequency spectrum obtained by performing a fast Fourier transform (FFT) on the induced voltage Vu. Also, the line graph G6 shown in the right part of FIG. 7 shows the argument (phase) of the frequency spectrum obtained by FFT.

[0137] Also, the line graph G6' shown in the right part of FIG. 7 shows the case where the temperature difference between the S magnet 25a and the N magnet 25b is the same as that of the graph G6, and the magnitude relationship between the temperatures of the S magnet 25a and the N magnet 25b is reversed from the situation shown in the graph G6.

[0138] As shown in the graph G4, the waveform of the induced voltage Vu is different from a sine wave. This is due to the change in the positional relationship between the magnet 25 and the coil 26 as the rotor 23 rotates. Looking at it from another perspective, the waveform shown in the graph G4 can also be understood as being realized by superimposing frequency components (harmonics) that are integer multiples of the predetermined frequency on a sine wave (fundamental wave) having the predetermined frequency.

[0139] Here, if the frequency of the fundamental wave is f0, the frequency of the lowest-order harmonic is f1, the number of poles of the motor 2 is Np, the motor rotation speed is Rm, and the number of magnets 25 of one pole (the number of magnets 25 per pole pair) is Nm, then f0 = Np × Rm...(3) f1 = Nm × f0...(4) The relationship is satisfied. Assume that Np = 8 and Nm = 2. The lowest-order harmonic becomes the second harmonic having a frequency twice that of the fundamental wave.

[0140] As a result of intensive studies, the inventors of the present application have found that, among the frequency spectra obtained by applying FFT to the induced voltage Vu, the frequency spectrum of the lowest-order harmonic includes information indirectly indicating the temperature difference between the S magnet 25a as the first magnet 25a and the N magnet 25b as the second magnet 25b.

[0141] According to the findings obtained by the inventors of the present application, the amplitude of the frequency spectrum of the lowest-order harmonic is proportional to the temperature difference between the S magnet 25a and the N magnet 25b. For example, the amplitude related to the harmonic becomes larger when the temperature difference between the magnets is large than when it is small. In the example shown in FIG. 7, as the temperature difference between the S magnet 25a and the N magnet 25b increases, the amplitude shown in the enclosed portion C4 becomes larger.

[0142] Further, according to the inventors of the present application, the phase of the frequency spectrum of the harmonic is inverted when the N magnet 25b is at a higher temperature than the S magnet 25a and when the S magnet 25a is at a higher temperature than the N magnet 25b. The example shown in FIG. 7 shows the case where the S magnet 25a is at a higher temperature than the N magnet 25b. In this case, the sign of the phase of the harmonic is positive as shown in the enclosed portion C5. On the other hand, even if the temperature difference between the magnets 25 is the same as in the example shown in FIG. 7, when the N magnet 25b is at a higher temperature than the S magnet 25a, the sign of the phase at the same frequency as the enclosed portion C5 is inverted and becomes negative as shown in the enclosed portion C5'.

[0143] Furthermore, according to the inventors of the present application, the amplitude of the fundamental wave frequency spectrum is related to the average temperature of the S magnet 25a and the N magnet 25b. Specifically, the amplitude related to the fundamental wave becomes larger when the average value of the magnetic flux density emitted by the S magnet 25a and the N magnet 25b is large than when the average value is small. On the other hand, as shown in the graph G7 of FIG. 8, the magnitude of the magnetic flux density has a linear relationship with the magnet temperature. Therefore, by using the amplitude related to the fundamental wave, the average temperature that each magnet 25 can take can be estimated.

[0144] Specifically, the induced voltage conversion unit 345 performs an FFT on the induced voltage in order to obtain a frequency spectrum corresponding to a predetermined frequency. Here, the predetermined frequency is the frequency of the lowest-order harmonic among the n-th harmonics (n is an integer) defined according to the number of magnets 25.

[0145] Next, the induced voltage conversion unit 345 estimates the magnetic flux density emitted by the magnet 25 based on the amplitude of the frequency spectrum obtained through the FFT. Specifically, the induced voltage conversion unit 345 adds the amplitude related to the fundamental wave and the amplitude related to the lowest-order harmonic. Thereby, the maximum value of the amplitude (maximum amplitude) considering the temperature difference between the magnets 25 can be calculated. In order to perform this calculation, the induced voltage conversion unit 345 according to the present embodiment includes a band-pass filter to which the frequency spectrum obtained through the FFT is input. The configuration of this band-pass filter can be appropriately changed according to the frequency of the lowest-order harmonic.

[0146] Then, the induced voltage conversion unit 345 calculates the magnetic flux density based on the maximum amplitude thus obtained, and estimates the magnet temperature based on the magnetic flux density. The maximum amplitude indicates the maximum value of the induced voltage Vu considering the temperature difference between the magnets 25. On the other hand, the induced voltage Vu and the magnetic flux density emitted by the magnet 25 are in a proportional relationship as is well known. Also, the magnetic flux density emitted by the magnet 25 and the magnet temperature have a linear relationship. Therefore, by detecting the induced voltage Vu, the magnet temperature can be calculated via the magnetic flux density.

[0147] Specifically, the induced voltage conversion unit 345 reads from memory a table corresponding to graph G7 in FIG. 8, and compares the table with the magnetic flux density to estimate the magnet temperature.

[0148] The induced voltage conversion unit 345 can also identify the magnet 25 that is relatively hotter or colder between the north magnet 25b and the south magnet 25a based on the phase of the frequency spectrum obtained through FFT. This function can be used for motor control using magnet temperature, which will be described later.

[0149] (2-2-6) Power supply restart section 346 The power supply restart unit 346 restarts the forward voltage V p After acquiring the forward voltage V p When the power supply is resumed, the power supply is immediately resumed. At this time, the plurality of switching elements 41 are turned on and off so as to realize normal PWM control.

[0150] (2-2-7) Model Modification Section 347 The model correction unit 347 can estimate the magnet temperature in the second region R2 in parallel with estimating the magnet temperature in the first region R1. The estimation in the first region R1 is performed by using the parameter K a , K. b , K. c Since this does not take into account the above, the estimation is more accurate than in the second region R2.

[0151] The model correction unit 347 corrects the estimation result in the second region R2 so that it matches the estimation result in the first region R1. This correction is performed by using the parameter K a , K. b , K. c This is done by adjusting one or more values of

[0152] Parameter K by the model correction unit 347 a , K. b , K. cThe adjustment may be executed periodically, may be executed irregularly, or may be executed each time an estimation is made in the first region R1.

[0153] (3) Method for reducing noise Incidentally, the estimation of the magnet temperature by the magnet temperature estimation unit 34 is realized by detecting the forward current if and the forward voltage Vf. These detections are performed during the period (standby time) when the switching element 41 of the power device 40 to be controlled is off. However, the switching element 41 is turned on and off at high speed, and the magnitude of the off period also changes. Therefore, the detection is easily affected by noise (switching noise) caused by switching.

[0154] For example, since the smaller the off period, the more affected by the switching noise, the influence of the switching noise can be reduced by detecting during a large off period. However, increasing the off period may prevent the motor 2 from being appropriately driven, and if detection is limited to a large off period, the magnet temperature cannot be appropriately estimated.

[0155] On the other hand, since a large amount of switching noise is superimposed on the forward current if and the forward voltage Vf detected during a small off period, the estimation accuracy of the magnet temperature decreases. That is, since the magnet temperature (estimated magnet temperature) estimated by the magnet temperature estimation unit 34 includes switching noise, it is preferable to reduce its influence in order to accurately estimate the magnet temperature.

[0156] Therefore, in this temperature estimation device 30, by combining the forward voltage and forward current of the above-described freewheeling diode 42, the estimation of the magnet temperature based on Equation (1) or Equation (2), and the calculation of the magnet temperature using the dynamic thermal model, the influence of the switching noise is reduced, and the estimation accuracy of the magnet temperature is improved.

[0157] Fig. 5 shows a block diagram regarding the estimation of the magnet temperature. The magnet temperature estimation unit 34 further includes a dynamic thermal model 38. And the magnet temperature estimation unit 34 further has extended software that combines the dynamic thermal model 38 with the above-described estimation of the magnet temperature. This extended software is configured based on a known Kalman filter.

[0158] The dynamic thermal model 38 can be designed using known thermal fluid analysis software or the like based on the peripheral structure of the magnet 25 such as the motor cooling system 2a and each factor contributing to the temperature change of the magnet 25 such as the current and voltage output by the inverter 10.

[0159] During the operation of the inverter 10, current and voltage are constantly output to the motor 2. At the same time, a refrigerant is supplied to the motor 2 by the motor cooling system 2a. In the thermal model 38, the temperature of the magnet 25 is dynamically calculated by detecting and introducing these factors that affect the heat generation and cooling of the magnet 25.

[0160] The dynamic thermal model 38 may be stored in the memory of the AD converter 35 as an arithmetic program. Further, in the memory of the AD converter 35, a heat generation map set based on the relationship between the current and voltage (power) output by the inverter 10 and the heating amount of the magnet 25, and a heat dissipation map set based on the relationship between the flow rate and temperature of the refrigerant and the heat dissipation amount of the magnet 25 may be stored and referred to as appropriate.

[0161] The magnet temperature estimation unit 34 identifies the current and voltage (power) output by the inverter 10 from its command value, measured value, etc. At the same time, the flow rate and temperature of the refrigerant supplied to the motor cooling system 2a are also identified from its command value, measured value, etc. Then, by introducing these factors that affect both the heat generation and heat dissipation of the magnet 25 into the dynamic thermal model 38, the temperature of the magnet 25 is calculated sequentially.

[0162] The magnet temperature estimation unit 34 uses a known Kalman filter to calculate the error between the temperature of the magnet 25 (calculated magnet temperature) calculated by the dynamic thermal model 38 and the estimated magnet temperature based on Equation (1) or Equation (2), and calculates the optimal gain K (Kalman gain K) at which the error is minimized. Then, the calculated error and Kalman gain K are introduced into the dynamic thermal model 38 to change the correction term related to the noise of the dynamic thermal model 38. By doing so, the magnet temperature estimation unit 34 corrects the dynamic thermal model 38. After that, the calculated value obtained by the corrected dynamic thermal model 38 is adopted as the temperature of the magnet 25.

[0163] As a result, the dynamic thermal model 38 is sequentially corrected, and the calculated magnet temperature converges to the true value. Therefore, the magnet temperature estimation unit 34 can estimate the temperature of the magnet 25 with reduced switching noise. Accordingly, the estimation accuracy of the temperature of the magnet 25 is improved.

[0164] Since the dynamic thermal model 38 is not affected by switching, the calculated magnet temperature does not include switching noise. On the other hand, as described above, the measured estimated magnet temperature includes switching noise. Therefore, the error between these calculated magnet temperature and estimated magnet temperature consists of a model error caused by the dynamic thermal model 38 and a noise error caused by the switching noise.

[0165] At this time, in the case of the Kalman filter, the model error is minimized while minimizing the noise error. Specifically, the covariance of these errors is calculated, and the gain K at which the value is minimized is set as the optimal gain K, that is, the Kalman gain K.

[0166] However, in this case, the calculation of the covariance has a large computational load. Therefore, there is a possibility that the responsiveness may be delayed. There is also a possibility that the hardware becomes large and cannot be incorporated into the AD converter 35.

[0167] Therefore, it may be configured to predict in advance the change in switching noise and thereby change the gain K (so-called feedforward control). By doing so, the calculation of the Kalman gain K becomes unnecessary, and along with the improvement in the estimation accuracy of the temperature of the magnet 25, the computational load can be reduced and high responsiveness can also be achieved.

[0168] (4) Combination of feedforward control Specifically, as shown by the dashed line in FIG. 9, an appropriate gain K is predicted based on the current and voltage output by the inverter 10, and the magnitude of the gain K is changed.

[0169] The magnet temperature estimation unit 34 further includes predetermined gain setting data. And the above-described extended software is configured to use this gain setting data. The gain setting data is provided with data regarding the correlation between the noise of the inverter 10, that is, a specific output related to the switching noise, and the gain K (also referred to as the predictive gain K in order to distinguish it from the Kalman gain K) corresponding to the dynamic thermal model 38.

[0170] For example, as the current output by the inverter 10 increases, the switching noise increases, and as the current output by the inverter 10 decreases, the switching noise decreases. Therefore, the current value output by the inverter 10 is correlated with the change in the switching noise (corresponding to the specific output).

[0171] The magnet temperature estimation unit 34 includes gain setting data (first gain setting data) regarding the correlation between the current value i output by the inverter 10 and the predictive gain K as shown in FIG. 10. In the first gain setting data, it is set such that the predictive gain K decreases as the current value i output by the inverter 10 increases.

[0172] Also, if the carrier frequency of the inverter 10 decreases, the switching noise increases, and if the carrier frequency of the inverter 10 increases, the switching noise decreases. Therefore, the carrier frequency of the inverter 10 is correlated with the change in the switching noise (corresponding to a specific output).

[0173] The magnet temperature estimation unit 34 includes gain setting data (second gain setting data) regarding the correlation between the carrier frequency fr of the inverter 10 and the predictive gain K as shown in FIG. 11. In the second gain setting data, the predictive gain K is set to decrease as the carrier frequency fr of the inverter 10 decreases.

[0174] The magnet temperature estimation unit 34 may also be provided with gain setting data (third gain setting data) regarding the correlation between both the current value and the carrier frequency output by the inverter 10 and the predictive gain K. In that case, although the amount of data increases, the predictive gain K corresponding to the changes in both of these specific outputs can be obtained.

[0175] The magnet temperature estimation unit 34 predicts an appropriate gain K based on the current value and the carrier frequency output by the inverter 10 and the first gain setting data or the second gain setting data (or the third gain setting data), and obtains the predictive gain K.

[0176] For example, when the current value output by the inverter 10 is input, the magnet temperature estimation unit 34 obtains the predictive gain K corresponding to the current value from the first gain setting data. Thereby, if the current value output by the inverter 10 decreases, the predictive gain K increases accordingly, and if the current value output by the inverter 10 increases, the predictive gain K decreases accordingly.

[0177] Similarly, when the carrier frequency of the inverter 10 is input, the magnet temperature estimation unit 34 obtains a predictive gain K corresponding to the carrier frequency from the second gain setting data. As a result, if the carrier frequency of the inverter 10 increases, the predictive gain K increases accordingly, and if the carrier frequency of the inverter 10 decreases, the predictive gain K decreases accordingly.

[0178] The magnet temperature estimation unit 34 introduces the thus obtained predictive gain K and the error between the estimated magnet temperature and the arithmetic magnet temperature into the dynamic thermal model 38, and changes the correction term related to the noise of the dynamic thermal model 38. By doing so, the magnet temperature estimation unit 34 corrects the dynamic thermal model 38. After that, the calculated value obtained by the corrected dynamic thermal model 38 is adopted as the temperature of the power device 40.

[0179] As a result, the dynamic thermal model 38 is sequentially corrected, and the calculated magnet temperature converges to the true value. Therefore, the magnet temperature estimation unit 34 can estimate the temperature of the power device 40 with reduced switching noise. Accordingly, the estimation accuracy of the temperature of the power device 40 is improved. Moreover, since the calculation of the Kalman gain K becomes unnecessary, the calculation load can be reduced and high responsiveness can be realized. As a result, the hardware can be reduced, so that it can be easily incorporated into the AD converter 35.

[0180] (5) Specific example of magnet temperature estimation process by temperature estimation device FIGS. 12 and 13 show an example of temperature estimation control of the magnet 25 by the temperature estimation device 30 with noise countermeasures. In this temperature estimation device 30, noise countermeasures are taken by the combination of the feedforward control described above.

[0181] When the vehicle 1 is keyed in and power is supplied to the on-board electrical equipment, the temperature estimation device 30 starts reading related data such as the current output by the inverter 10 (step S1). While the inverter 10 is operating, that is, while power is being supplied to the on-board electrical equipment, the temperature estimation device 30 always executes the temperature estimation control process (Yes in step S2). When the inverter 10 stops, the temperature estimation device 30 stops the temperature management control process (No in step S2).

[0182] When the inverter 10 is operating and power is being output to the motor 2, the temperature estimation device 30 (magnet temperature estimation unit 34) preferably executes the magnet 25 temperature estimation process using the off period of the switching element 41, the magnet 25 temperature calculation process using the dynamic thermal model 38, and the acquisition process of the predictive gain K in parallel. By doing so, the temperature can be estimated in a short time, resulting in excellent responsiveness.

[0183] (Magnet temperature estimation process using the off period of the switching element) FIG. 14 shows an example of the magnet 25 temperature estimation process using the off period of the switching element 41. FIG. 15 shows an example of the process related to the classification of the load of the motor 2. The processes of FIGS. 14 and 15 illustrate step S3 of FIG. 12 in detail.

[0184] First, in step S31, the motor load determination unit 341 determines the load of the motor 2. The motor load determination unit 341 determines whether the load of the motor 2 belongs to the first region R1, the second region R2, or the third region R3.

[0185] FIG. 15 shows a specific example of the process performed in step S31. First, in step S311, the motor load determination unit 341 acquires the load of the motor 2. This process may be performed based on, for example, the target load of the motor 2 calculated by the controller 7.

[0186] In step S312, the motor load determination unit 341 determines whether the motor 2 is unloaded. This determination can be YES, for example, when the vehicle 1 is being driven only by the driving torque output by the engine 3. If this determination is YES, the control process proceeds to step S313. The motor load determination unit 341 determines that the load on the motor 2 belongs to the first region R1. The first region R1 is a region where the estimation accuracy of the magnet temperature is the best, although the OFF period (standby time) of the switching element 41 is the longest first standby time Tw1. In the first region R1, Equation (1) is used as described above.

[0187] If the determination in step S312 is NO, the control process proceeds to step S314. In this case, the motor load determination unit 341 determines whether driving torque is being output from both the motor 2 and the engine 3. If this determination is YES, the control process proceeds to step S315. The motor load determination unit 341 determines that the load on the motor 2 belongs to the second region R2. The second region R2 is a region where the OFF period of the switching element 41 is the intermediate second standby time Tw2, and it is a region where the influence on the drive control of the motor 2 due to its OFF period and the estimation accuracy of the magnet temperature can be balanced well. In the second region R2, Equation (2) is used as described above.

[0188] If the determination in step S314 is NO, the control process proceeds to step S316. In this case, the motor load determination unit 341 determines that, for example, driving torque is being output only from the motor 2. The motor load determination unit 341 determines that the load on the motor 2 belongs to the third region R3. The third region R3 is a region where the OFF period of the switching element 41 is the shortest third standby time Tw3, and it is a region where the influence on the drive control of the motor 2 due to its OFF period can be minimized. In the third region R3, Equation (2) is used as described above.

[0189] When any one of step S313, step S315, or step S316 ends, the control process shifts from step S31 in FIG. 14 to step S32 in the same figure.

[0190] In the subsequent step S32, the model modification unit 347 determines whether it is necessary to modify the model shown in Equation (2). Regarding this process, for example, it may be configured to determine "model modification is necessary" periodically.

[0191] In the subsequent step S33, the standby time determination unit 342 sets a standby time corresponding to the load of the motor 2. As illustrated in FIG. 14, this standby time may be selected from the first standby time Tw1, the second standby time Tw2, and the third standby time Tw3 according to whether the load of the motor 2 belongs to any of the first region R1 to the third region R3.

[0192] In the subsequent step S34, the power supply cut-off unit 343 turns off all the switching elements 41. In the subsequent step S35, the magnet temperature estimator 34 acquires the current value of the forward current if at the off time, that is, the value of the initial current I0.

[0193] In the subsequent step S36, the induced voltage estimator 344 detects, for example, the value of the forward voltage Vf at the time when the standby time has elapsed. As described above, instead of the value of the forward voltage Vf at the time when the standby time has elapsed, a moving average including the value may be used. When the model modification unit 347 performs model modification, in addition to the value of the forward voltage Vf in the first region R1, the value of the forward voltage Vf in the second region R2 is also detected.

[0194] In the subsequent step S37, the induced voltage estimator 344 estimates the value of the induced voltage based on the value of the forward voltage Vf. This estimation is performed with reference to Equation (1) when the load of the motor 2 belongs to the first region R1, and is performed with reference to Equation (2) when the load of the motor 2 belongs to the second region R2 or the third region R3.

[0195] When the model modification unit 347 performs model modification, the induced voltage is estimated based on each of the detected value of the forward voltage Vf in the first region R1 and Equation (1), and the detected value of the forward voltage Vf in the second region R2 and Equation (2), respectively.

[0196] In the subsequent step S38, the induced voltage conversion unit 345 converts the estimated value of the induced voltage into the magnet temperature Tsw. For the conversion of the induced voltage into the magnet temperature Tsw by the induced voltage conversion unit 345, for example, the method described in Japanese Patent Application Laid-Open No. 2021-118653 can be used. Therefore, details of the conversion method are omitted.

[0197] Note that when the model correction unit 347 performs model correction, conversion to the magnet temperature Tsw is performed for each of the estimated value based on the forward voltage Vf in the first region R1 and the estimated value based on the forward voltage Vf in the second region R2.

[0198] In the subsequent step S39, the power supply restart unit 346 restarts the power supply to the motor 2. This restart may be performed immediately when the value of the forward voltage Vf is detected. As a result, as shown in the upper diagram of FIG. 17, the current decrease is temporary. In FIG. 17, the section indicated by a pair of arrows indicates the period during which all the switching elements 41 are turned off. The current value will temporarily decrease during the period from when all the switching elements 41 are turned off until the power supply is restarted.

[0199] Also, regarding the lower diagram of FIG. 17, when all the switching elements 41 are turned off, although the voltage value significantly decreases, it temporarily becomes a constant value within a short period. This means that the voltage value has changed to the induced voltage. After that, when the power supply is restarted, the voltage value will resume a trigonometric change.

[0200] In the subsequent step S391, the model correction unit 347 corrects the model shown in Equation (2) based on the determination in step S32. When correcting the model, one or more parameters in Equation (2) are corrected so that the magnet temperature based on the forward voltage Vf in the second region R2 approaches the magnet temperature based on the forward voltage Vf in the first region R1. If model correction is not required, step S391 is skipped.

[0201] When the process of step S39 or step S391 is completed, the control process proceeds from step S3 to step S4 in FIG. 12. In this step S4, the temperature estimation device 30 acquires the estimated magnet temperature Tsw as the estimated magnet temperature Tsw1.

[0202] (Calculation process of magnet temperature using dynamic thermal model) In the process of calculating the temperature of the magnet 25 using the dynamic thermal model 38, the temperature estimation device 30 calculates the magnet temperature Tsw using the dynamic thermal model 38 (step S5). That is, the current and voltage (power) output by the inverter 10 are specified from their command values, measured values, etc. At the same time, the flow rate and temperature of the refrigerant supplied from the motor cooling system 2a are also specified from their command values, measured values, etc. Then, these factors that affect both the heat generation and heat dissipation of the magnet 25 are introduced into the dynamic thermal model 38. By doing so, the temperature estimation device 30 calculates the temperature of the magnet 25.

[0203] The temperature estimation device 30 acquires the magnet temperature Tsw estimated in this way as the calculated magnet temperature Tsw2 (step S6).

[0204] The temperature estimation device 30 calculates the error between these by subtracting the acquired estimated magnet temperature Tsw1 and the calculated magnet temperature Tsw2 (step S7). Since it is only a subtraction, the computational load is almost zero.

[0205] (Acquisition process of predictive gain K) The temperature estimation device 30 predictively acquires the gain K based on the specific output (current value, carrier frequency) related to the noise from the current and voltage (power) output by the inverter 10 (step S8).

[0206] Specifically, when the current value output by the inverter 10 is input, the magnet temperature estimation unit 34 acquires a predicted gain K corresponding to the current value from the first gain setting data. Alternatively, when the carrier frequency of the inverter 10 is input, the magnet temperature estimation unit 34 acquires a predicted gain K corresponding to the carrier frequency from the second gain setting data. The temperature estimation device 30 preferentially adopts the predicted gain K based on either the current value or the carrier frequency according to the situation.

[0207] When the magnet temperature estimation unit 34 includes third gain setting data, when the current value and the carrier frequency output by the inverter 10 are input, the magnet temperature estimation unit 34 acquires a predicted gain K corresponding to the current value and the carrier frequency from the third gain setting data.

[0208] When the calculation process of the error between the estimated magnet temperature Tsw1 and the calculated magnet temperature Tsw2 and the acquisition process of the predicted gain K are completed, as shown in FIG. 13, the temperature estimation device 30 introduces the calculated error and the acquired predicted gain K into the dynamic thermal model 38 (step S9).

[0209] Thus, as described above, the dynamic thermal model 38 is corrected, and the corrected dynamic thermal model 38 is used to estimate the magnet temperature Tsw (steps S10, S11).

[0210] Specifically, the dynamic thermal model 38 is corrected by changing the correction term related to the noise of the dynamic thermal model 38 using the predicted gain K and the error. After that, the calculated value obtained by the corrected dynamic thermal model 38 is adopted as the finally estimated magnet temperature Tsw of the magnet 25.

[0211] The temperature estimation device 30 determines whether or not the magnet temperature Tsw estimated in this way exceeds a predetermined temperature Ts (step S12). Specifically, data (current limit data) showing the relationship between the maximum current value that the inverter 10 can output and the temperature of the magnet temperature Tsw, as shown in FIG. 16, is preset in the temperature estimation device 30 (magnet temperature estimation unit 34).

[0212] Then, when the temperature estimation device 30 refers to its current limit data and determines that the estimated magnet temperature Tsw exceeds the temperature Ts (Yes in step S12), it decreases the maximum current value that the inverter 10 can output according to the magnet temperature Tsw, and manages the current so that it does not exceed that value (step S13).

[0213] When the estimated magnet temperature Tsw is equal to or higher than Tmax, the operation of the inverter 10 is stopped so that the inverter 10 does not output current. On the other hand, when the temperature estimation device 30 refers to the current limit data and determines that the estimated magnet temperature Tsw does not exceed the temperature Ts (No in step S14), the maximum current value is set to the normal value (a constant value), and the current is managed so that it does not exceed that value. Then, the process returns to step S2, and the estimation process of the magnet temperature Tsw is repeated.

[0214] Thus, with this temperature estimation device 30, the magnet temperature can be estimated with high accuracy. Therefore, by using this temperature estimation device 30 to manage the temperature of the magnet 25, the thermal influence on the motor 2 can be effectively suppressed.

[0215] (6) Regarding the balance between the estimation accuracy of the magnet temperature and the driving force control of the motor When the switching element 41 is turned off, the voltage value of the forward voltage Vf changes to an induced voltage suitable for estimating the magnet temperature. The changing voltage value includes not only the induced voltage but also those caused by the back electromotive force. On the other hand, the voltage value after the change can be regarded as the induced voltage itself.

[0216] Using the former voltage value can minimize the influence on the drive control of the motor 2. However, since it is not the induced voltage itself, it is inconvenient for improving the estimation accuracy of the magnet temperature.

[0217] Using the latter voltage value is convenient for improving the estimation accuracy of the magnet temperature. However, it is inconvenient for suppressing the influence on the drive control of the motor 2.

[0218] However, generally, whether to prioritize the estimation accuracy of the magnet temperature or to prioritize suppressing the influence on the drive control of the motor 2 may vary depending on the operating state of the vehicle 1 or the like. For example, when the motor load is relatively high, suppressing the influence on the drive control of the motor 2 should be prioritized. On the other hand, when the motor load is relatively low (e.g., at no load), suppressing the influence on the drive control of the motor is not much of a problem, and improving the estimation accuracy of the magnet temperature is prioritized.

[0219] Therefore, as illustrated in FIG. 15, by increasing or decreasing the waiting time indicating the acquisition timing of the voltage value according to the load of the motor 2, the balance between the estimation accuracy of the magnet temperature and the drive force control of the motor 2 can be made appropriate according to the motor load at that time.

[0220] Also, as illustrated in step S39 of FIG. 14, as soon as the voltage value for estimating the magnet temperature is acquired, the power supply to the motor 2 is immediately restarted. Thereby, the influence on the drive force control of the motor 2 can be suppressed as much as possible.

[0221] Moreover, considering the back electromotive force in the voltage value corresponds to detecting the voltage value in the transient state before reaching the induced voltage. In this case, although it is advantageous for suppressing the influence on the drive force control of the motor 2, it is disadvantageous for improving the estimation accuracy of the magnet temperature.

[0222] Therefore, as described with reference to FIGS. 4 and 15, by properly using or not using the back electromotive force according to the load of the motor 2, the balance between the estimation accuracy of the magnet temperature and the drive force control of the motor 2 can be made appropriate according to the motor load at that time.

[0223] Also, using the voltage value in the transient state is advantageous for suppressing the influence on the drive force control of the motor 2 as described above, but in terms of using the model depending on the inductance L as in Equation (2), it may be disadvantageous for improving the estimation accuracy of the magnet temperature.

[0224] However, when the motor 2 is under no load, there is no need to consider its influence on the driving force control anymore. Therefore, as illustrated in FIGS. 4 and 15, by properly selecting the voltage value according to whether the motor 2 is under no load, it is possible to appropriately balance the estimation accuracy of the magnet temperature and the driving force control of the motor 2.

[0225] Also, although using the voltage value detected before it drops to the allowable current value Is is particularly advantageous in suppressing the influence on the driving force control of the motor 2, it is disadvantageous for improving the estimation accuracy of the magnet temperature due to the influence of magnetic saturation on the value of the inductance L.

[0226] For example, when the load of the motor 2 is relatively high, it is considered that more suppression of the influence on the driving force control of the motor 2 is required compared to when it is relatively low.

[0227] Therefore, as illustrated in the graph of FIG. 4 and step S314 of FIG. 15, by properly selecting the detection timing of the voltage value, it is possible to make the balance between the estimation accuracy of the magnet temperature and the driving force control of the motor 2 more appropriate.

[0228] Also, when the driving torque of the hybrid vehicle is output only from the motor 2, it is considered that the influence on the driving force control of the motor 2 should be suppressed as much as possible. On the other hand, when the driving torque of the hybrid vehicle is output from the engine 3 and the motor 2, it is considered that less suppression of the influence on the driving force control of the motor 2 is required compared to the former case.

[0229] Therefore, as illustrated in step S314 of FIG. 15, by having the magnet temperature estimation unit 34 make a determination, it is possible to make the balance between the estimation accuracy of the magnet temperature and the driving force control of the motor 2 more appropriate.

[0230] Also, as illustrated in step S40 of FIG. 14, by appropriately correcting the model of Equation (2), the estimation accuracy of the magnet temperature can be improved.

Description of Reference Numerals

[0231] 1 Automobile 2 Motor 25 Magnet 26 Coil 3 Engine 4 Battery 10 Inverter 30 Magnet Temperature Estimation Device 31 Diode Current Detection Unit 32 Voltage Dividing Unit 32a Voltage Dividing Diode (First Element) 32b Voltage Dividing Resistor (Second Element) 33 Diode Voltage Detection Unit 33a Protection Diode 33b Transformer Unit 34 Magnet Temperature Estimation Unit 341 Motor Load Judgment Unit 342 Standby Time Determination Unit 35 AD Converter 38 Dynamic Thermal Model 40 Power Device 41 Switching Element 42 Freewheeling Diode R1 First Region R2 Second Region R3 Third Region Tw1 First Standby Time Tw2 Second Standby Time Tw3 Third Standby Time

Claims

1. An inverter in which a plurality of power devices including a switching element and a freewheeling diode connected in anti-parallel with the switching element are mounted, A motor electrically connected to the inverter, A battery that supplies power to the motor via the inverter, A magnet temperature estimation unit that is electrically connected to the inverter and estimates the magnet temperature of the motor based on a voltage value corresponding to the current flowing through the inverter, When estimating the magnet temperature, the magnet temperature estimation unit Turns off all the switching elements of the plurality of power devices so as to cut off the power supply to the motor, Estimates the induced voltage associated with the rotation of the motor based on the voltage value within a period from when the switching element is turned off until a predetermined standby time has elapsed, Estimates the magnet temperature based on the correlation between the induced voltage and the magnet temperature, The magnet temperature estimation unit also sets the standby time so as to increase or decrease according to the load of the motor A magnet temperature estimation device for a motor, characterized in that.

2. In the magnet temperature estimation device for a motor according to claim 1, After acquiring the voltage value, the magnet temperature estimation unit turns on a plurality of the switching elements so as to resume power supply to the motor A magnet temperature estimation device for a motor, characterized in that.

3. In the magnet temperature estimation device for a motor according to claim 1, When the load of the motor is high, the magnet temperature estimation unit shortens the standby time compared to when the load is low A magnet temperature estimation device for a motor, characterized in that.

4. In the magnet temperature estimation device for a motor according to claim 3, The magnet temperature estimation unit selectively estimates the induced voltage by taking into account the counter electromotive force in the voltage value according to the level of the load of the motor A magnet temperature estimation device for a motor, characterized in that. 【Claim ​ ​ ​ When the load on the motor is no load, the voltage value used is the voltage value after the transition from the transient state to the steady state, and this voltage value is regarded as the induced voltage. A magnet temperature estimation device for a motor, characterized in that.

6. In the magnet temperature estimation device for a motor according to claim 5, If the current value that is set corresponding to the specifications of the motor and serves as an index for the coil of the motor to cause magnetic saturation is defined as the allowable current value, when the load on the motor is higher than no load, when the load on the motor is relatively high, the induced voltage is estimated based on the voltage value acquired before the transient current accompanying the turn-off of the switching element drops to the allowable current value and the model; when the load on the motor is relatively low, the induced voltage is estimated based on the voltage value acquired after the transient current drops to the allowable current value and the model. A magnet temperature estimation device for a motor, characterized in that.

7. In the magnet temperature estimation device for a motor according to claim 6, the motor is mounted on a hybrid vehicle equipped with an engine that cooperates with the motor, the magnet temperature estimation unit, when the driving torque of the hybrid vehicle is output only from the motor, determines that the load on the motor is relatively high; when the driving torque is output from both the motor and the engine, determines that the load on the motor is relatively low. A magnet temperature estimation device for a motor, characterized in that.

8. In the magnet temperature estimation device for a motor according to claim 5, when the load on the motor is no load, the magnet temperature estimation unit corrects the model so that the induced voltage based on the model approaches the induced voltage based on the voltage value acquired after the transition to the steady state. A magnet temperature estimation device for a motor, characterized in that.

9. In the magnet temperature estimation device for a motor according to claim 1, the magnet temperature estimation unit further has a dynamic thermal model for calculating the magnet temperature, calculates the error and the Kalman gain between the estimated magnet temperature, which is the magnet temperature estimated by the magnet temperature estimation unit, and the calculated magnet temperature, which is the magnet temperature calculated by the dynamic thermal model, and corrects the dynamic thermal model by introducing the error and the Kalman gain, and adopts the calculated value by the corrected dynamic thermal model as the magnet temperature. A magnet temperature estimation device for a motor, characterized in that.

10. In the magnet temperature estimation device for a motor according to Claim 1, the magnet temperature estimation unit further has a dynamic thermal model for calculating the magnet temperature, specific output related to the noise of the inverter, and gain setting data related to a gain corresponding to the dynamic thermal model, calculates an error between an estimated magnet temperature, which is the magnet temperature estimated by the magnet temperature estimation unit, and a calculated magnet temperature, which is the magnet temperature calculated by the dynamic thermal model, obtains the gain from the gain setting data, corrects the dynamic thermal model by introducing the error and the gain, and adopts a calculated value by the corrected dynamic thermal model as the magnet temperature A magnet temperature estimation device for a motor, characterized by the above.

11. In the magnet temperature estimation device for a motor according to Claim 10, the specific output is a current value output by the inverter, and the gain setting data is set such that the gain decreases as the current value increases A magnet temperature estimation device for a motor, characterized by the above.

12. In the magnet temperature estimation device for a motor according to Claim 10, the specific output is the carrier frequency of the inverter, and the gain setting data is set such that the gain decreases as the carrier frequency decreases A magnet temperature estimation device for a motor, characterized by the above.

13. A hybrid vehicle, comprising the magnet temperature estimation device for a motor according to Claim 1, the motor, and an engine that cooperates with the motor characterized by the above.

Images