Temperature estimation device and temperature estimation method of pm motor

The PM motor temperature estimation device improves accuracy by calculating q-axis voltage errors using identified motor parameters and optimization techniques, addressing the inaccuracies and complexity of existing methods without additional sensors, ensuring consistent magnet temperature estimation.

JP2024167494AActive Publication Date: 2024-12-04MEIDENSHA CORP
View PDF 7 Cites 0 Cited by

Patent Information

Application Number
JP2023083594
Authority / Receiving Office
JP · JP
Patent Type
Applications
Current Assignee / Owner
Filing Date
2023-05-22
Publication Date
2024-12-04
Estimated Expiration
2043-05-22

AI Technical Summary

Technical Problem

Existing methods for estimating the rotor magnet temperature of a PM motor suffer from inaccuracies due to motor parameter errors and require additional components like search coils and torque sensors, leading to suboptimal accuracy and increased complexity.

Method used

A PM motor temperature estimation device that calculates the q-axis voltage error based on d-axis and q-axis currents, using identified motor parameters and voltage error values to determine magnet temperature without the need for additional sensors, and employs optimization techniques to minimize estimation errors.

Benefits of technology

Achieves accurate magnet temperature estimation with reduced complexity and memory requirements by deriving optimal parameters and using voltage error tables or surface fitting equations, ensuring consistent accuracy under varying current conditions.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure 2024167494000001_ABST
    Figure 2024167494000001_ABST
Patent Text Reader

Abstract

To enhance magnet temperature estimation accuracy of a PM motor 1 driven by an inverter 2 provided with a current control system.SOLUTION: A temperature estimation device of a PM motor comprises: a parameter pre-identification device 30 for deriving a reference magnetic flux Φn, a temperature coefficient β, a d-axis inductance Ld, and a resistance Ra being parameters of the motor on the basis of voltage commands Vd* and Vq*, an electric angular velocity ωe, d-axis and q-axis current commands id* and iq*, d-axis and q-axis currents id and iq, and a magnet temperature measured value Tpm obtained by measuring a magnet temperature of the PM motor; a voltage error table calculator for calculating a q-axis voltage error value ΔVq on the basis of the Vd*, Vq*, ωe, id, iq, and Tpm, and the parameters of the motor Φn, β, Ld, and Ra; and a magnet temperature estimation unit for obtaining a magnet temperature estimation value on the basis of the voltage commands Vd* and Vq*, the d-axis and q-axis currents id and iq, the electric angular velocity ωe, and the calculated q-axis voltage error value ΔVq.SELECTED DRAWING: Figure 1
Need to check novelty before this filing date? Find Prior Art

Description

[Technical field]

[0001] The present invention relates to an apparatus and method for estimating a rotor magnet temperature of a permanent magnet (PM) motor from measurement data of the inverter that drives the PM motor. [Background technology]

[0002] As a conventional technology for estimating the rotor magnet temperature of a PM motor, for example, Non-Patent Document 1 describes a method in which the flux linkage caused by the magnet is estimated from the measured current and voltage command of the inverter based on the dq-axis voltage equation of the PM motor, and the magnet temperature is estimated from the change in the estimated magnetic flux.

[0003] Patent document 1 also describes a method of estimating magnet temperature from the induced voltage of a PM motor, in which a search coil is embedded to measure the induced voltage while the motor is running, and a voltage sensor is also added to measure the voltage applied to the motor.

[0004] Furthermore, Patent Document 2 describes a magnet temperature estimation method using a voltage disturbance observer, and Patent Document 3 describes a magnet temperature estimation method using a disturbance observer and a torque meter. [Prior art documents] [Non-patent literature]

[0005] [Non-Patent Document 1] Ishihara, et al.: "Magnet Temperature Estimation Using a Magnetic Flux Simulator for IPMSM", 2022 National Convention of the Institute of Electrical Engineers of Japan, No. 5-079 [Patent documents]

[0006] [Patent Document 1] JP 2021-118652 A [Patent Document 2] Patent Publication No. 2022-178400 [Patent Document 3] Patent No. 7063406 Summary of the Invention [Problem to be solved by the invention]

[0007] In the method of Non-Patent Document 1, fixed values ​​of motor parameters are used, so errors in the motor parameters cause errors in the magnet temperature estimation, and the accuracy of the magnet temperature estimation is poor.

[0008] The method of Patent Document 1 requires the addition of a search coil and a voltage sensor.

[0009] The method of Patent Document 3 requires a torque sensor.

[0010] The present invention is devised to solve the above-mentioned problems, and has an object to provide a temperature estimating device for a PM motor with improved magnet temperature estimation accuracy. [Means for solving the problem]

[0011] In order to solve the above problem, the temperature estimation device for a PM motor according to claim 1 comprises: In a PM motor driven by an inverter equipped with a current control system, a q-axis voltage error calculation unit that calculates a q-axis voltage error value ΔVq, which is a difference between a command value and a detected value of a q-axis voltage, based on a d-axis current id and a q-axis current iq obtained by coordinate-transforming the detected current of the PM motor into a dq axis; The voltage command Vd obtained by controlling the current so that the motor detection current detected by the current flowing through the PM motor is equal to the current command. * ,Vq * and a magnet temperature estimation unit that calculates a magnet temperature estimate based on the d-axis current id, the q-axis current iq, and the electrical angular velocity ωe calculated by detecting the d-axis current id, the q-axis current iq, and the rotation speed of the PM motor, and the q-axis voltage error value ΔVq calculated by the q-axis voltage error calculation unit.

[0012] The temperature estimation device for a PM motor according to claim 2 is the device according to claim 1, The voltage command Vd* ,Vq * , electrical angular velocity ωe, d-axis current command id * , q-axis current command iq * A parameter identification unit derives reference magnetic flux Φn, temperature coefficient β, d-axis inductance Ld, and resistance Ra, which are motor parameters that minimize the error between the estimated value and the measured value of the magnet temperature, based on the d-axis current id, the q-axis current iq, and the magnet temperature measurement value Tpm of the PM motor, The q-axis voltage error calculation unit is * ,Vq * , ωe, id * , iq * , id, iq, and Tpm, and a voltage error table calculator that calculates a q-axis voltage error value ΔVq by calculating equations (1), (2), and (3) based on the motor parameters Φn, β, Ld, and Ra derived by the parameter identification unit,

[0013]

number

[0014]

number

[0015]

number

[0016] (T^ W.0.ΔVq is the magnet temperature estimate when voltage error is not taken into account, T0 is the reference temperature, and Tpm is the measured magnet temperature. The magnet temperature estimation unit is a test data Vd * ,Vq * , id, iq, ωe, the motor parameters Φn, β, Ld, Ra derived by the parameter identification unit, and the q-axis voltage error value ΔVq calculated by the voltage error table calculator, to calculate equation (4) to obtain a magnet temperature estimate.

[0017]

number

[0018] (T^ is the estimated magnet temperature, T0 is the reference temperature) The temperature estimation device for a PM motor according to claim 3 is the device according to claim 1, The q-axis voltage error calculation unit is * ,Vq * , ωe, id * , iq * , id, iq, Tpm, and the PM motor design parameters, that is, reference magnetic flux Φn, temperature coefficient β, d-axis inductance Ld, and resistance Ra, are used to calculate equations (1), (2), and (3) to calculate a q-axis voltage error value ΔVq;

[0019]

number

[0020]

number

[0021]

number

[0022] (T^ W.0.ΔVq is the magnet temperature estimate when voltage error is not taken into account, T0 is the reference temperature, and Tpm is the measured magnet temperature. The magnet temperature estimation unit is a test data Vd * ,Vq * , id, iq, ωe, and the q-axis voltage error value ΔVq calculated by the voltage error table calculator, the magnet temperature estimate is calculated using equation (4).

[0023]

number

[0024] (T^ is the estimated magnet temperature, T0 is the reference temperature) The temperature estimation device for a PM motor according to claim 4 is the device according to claim 1, The q-axis voltage error calculation unit calculates the value of the voltage error table as ΔVq_ value , f is a function of the surface, and a is a parameter vector of the surface equation, ΔVq_ value -Derive the parameter vector of the equation of the surface that minimizes f(Id, Iq, a) by solving the optimization problem using equation (17).

[0025]

number

[0026] (a * is the optimized parameter vector, and argmin is the set of values ​​at which the minimum point is reached) The surface fitting equation is formulated using the parameter vector derived from equation (17), and the d-axis current Id and q-axis current Iq during actual operation are input into equation (17) to calculate the q-axis voltage error value ΔVq.

[0027] The temperature estimation device for a PM motor according to claim 5 is the device according to claim 2, the parameter identification unit calculates a magnetic flux Φ (=Vemf / ωe) using measurement data of an induced voltage Vemf of the PM motor measured during no-load operation of the PM motor and an electrical angular velocity ωe obtained by detecting a rotation speed of the PM motor; identifying Φn and β as motor parameters so that a line Φ=aTpm+b (a is the slope of the line, and b is the y-intercept of the line) obtained by the least squares method, which expresses the relationship between the magnet temperature measurement value Tpm obtained by measuring the magnet temperature of the PM motor and the calculated magnetic flux Φ, corresponds to Φ=Φn(1+β(Tpm-T0)) (Φn is a reference magnetic flux, β is a temperature coefficient, and T0 is a reference temperature) which defines the temperature change of the magnetic flux; determining an operating speed when a magnet temperature estimation value estimated by the magnet temperature estimator when a constant speed operation test of the PM motor is performed with a d-axis current command value Idcmd and a q-axis current command value Iqcmd set to 0 is close to a magnet temperature measurement value as a speed region where a valid temperature estimation is performed; A constant speed operation test is carried out in the determined speed range with the d-axis current command value Idcmd<0 and the q-axis current command value Iqcmd=0, or Idcmd>0 and Iqcmd=0, and Vd * ,Vq * , id, iq, ωe, and Tpm, and based on the acquired steady-state data, an optimization problem is solved using an evaluation equation for finding a combination of parameters, d-axis inductance Ld and q-axis voltage error ΔVq, which minimizes the square error between the measured value and the estimated value of the magnet temperature, as shown in equations (9) to (12), to identify the d-axis inductance Ld;

[0028]

number

[0029] subject to

[0030]

number

[0031]

number

[0032]

number

[0033] (T^ is the estimated magnet temperature, Ldn is the d-axis inductance, and ΔVq is the difference between the q-axis voltage command value and the q-axis voltage detection value.) In the speed range determined above where a reasonable temperature estimation was performed, a constant speed operation test was performed with the d-axis current command value Idcmd=0 and the q-axis current command value Iqcmd>0, or Idcmd=0 and Iqcmd<0, and Vd * ,Vq * , id, iq, ωe, and Tpm are acquired, and based on the acquired steady-state data, an optimization problem is solved using evaluation equations (13) to (16) that find a combination of parameter resistance Ran and q-axis voltage error ΔVq that minimizes the square error between the measured and estimated magnet temperature values, thereby identifying Ran.

[0034]

number

[0035] subject to

[0036]

number

[0037]

number

[0038]

number

[0039] (T^ is the estimated magnet temperature, Ran is the winding resistance of the PM motor, and ΔVq is the difference between the q-axis voltage command value and the q-axis voltage detection value.) The method for estimating the temperature of a PM motor according to claim 6 further comprises: A method for estimating a temperature of a PM motor driven by an inverter having a current control system, comprising: a q-axis voltage error calculation step in which a q-axis voltage error calculation unit calculates a q-axis voltage error value ΔVq, which is a difference between a command value and a detected value of the q-axis voltage, based on a d-axis current id and a q-axis current iq obtained by coordinate-transforming the detected current of the PM motor into the dq-axis; The magnet temperature estimation unit detects the current flowing through the PM motor, which is test data, and performs current control so that the motor detection current becomes equal to the current command. The voltage command Vd * ,Vq * and a magnet temperature estimation step of calculating an estimated magnet temperature value based on the d-axis current id, the q-axis current iq, and the electrical angular velocity ωe calculated by detecting the d-axis current id, the q-axis current iq, and the rotation speed of the PM motor, and the q-axis voltage error value ΔVq calculated by the q-axis voltage error calculation unit. Effect of the Invention

[0040] (1) According to the invention as set forth in claims 1 to 6, magnet temperature estimation can be achieved with a certain degree of accuracy even under current conditions where d-axis current id≠0 and q-axis current iq≠0. (2) According to the invention as recited in claims 2 and 5, the motor parameters derived by the parameter identification unit are used, so that the magnet temperature can be estimated with high accuracy. (3) According to the invention as set forth in claim 3, the effort of deriving motor parameters (optimum parameters) is eliminated, and the configuration is simplified. (4) According to the invention as recited in claim 4, by using an equation for surface fitting, it is not necessary to store a voltage error table in memory, so that the memory capacity can be reduced. [Brief description of the drawings]

[0041] [Figure 1] FIG. 2 is a diagram showing the configuration of an actual device at the time of pre-identification according to an embodiment of the present invention. [Diagram 2] FIG. 2 is a diagram showing an actual configuration of an embodiment of the present invention when estimating a magnet temperature. [Diagram 3] FIG. 2 is a block diagram of a pre-identification phase in the first embodiment of the present invention. [Figure 4] FIG. 2 is a block diagram of an estimation phase in the first embodiment of the present invention. [Diagram 5] FIG. 2 is a block diagram of a voltage error ΔVq table calculation according to the first embodiment of the present invention. [Figure 6] FIG. 4 is a characteristic diagram showing the relationship between magnetic flux and magnet temperature. [Figure 7] FIG. 2 is an explanatory diagram of a step function used in the first embodiment of the present invention. [Figure 8] FIG. 11 is a block diagram of a pre-identification phase in a second embodiment of the present invention. [Figure 9] FIG. 11 is a block diagram of an estimation phase in the second embodiment of the present invention. [Figure 10] FIG. 11 is an explanatory diagram of surface fitting of a voltage error table in the third embodiment of the present invention. [Figure 11] 1 shows the results of actual device verification illustrating the effects of the present invention, where (a) is a graph of temperature estimation values, (b) is a graph of temperature measurement values, and (c) is a graph of temperature estimation errors. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0042] Hereinafter, an embodiment of the present invention will be described with reference to the drawings, but the present invention is not limited to the following embodiment example.

[0043] The PM motor magnet temperature estimation in this embodiment is composed of (1) a pre-identification phase in which motor parameters are identified in advance, and (2) an estimation phase in which the magnet temperature is estimated based on the identified parameters. EXAMPLES

[0044] FIG. 1 shows an actual device configuration in the pre-identification phase in the first embodiment, and FIG. 2 shows an actual device configuration in the estimation phase.

[0045] In FIG. 1 showing the actual machine configuration at the time of pre-identification, reference numeral 1 denotes a PM motor driven by an inverter 2, 3 denotes a load connected to the PM motor 1, and 4 denotes a rotational position sensor that detects the rotation of the PM motor and outputs a phase θ.

[0046] Reference numeral 5 denotes a pole pair number calculator which multiplies the phase θ detected by the rotational position sensor 4 by the number of pole pairs P of the motor to convert it into a phase θe taking the number of pole pairs into consideration.

[0047] A differential calculator 6 calculates an electrical angular velocity ωe by differentiating the phase θe taking into account the number of pole pairs.

[0048] Reference numeral 7 denotes a magnet temperature sensor (eg, a thermocouple) that measures the magnet temperature of the PM motor 1 .

[0049] Reference numeral 8 denotes a current sensor that detects three-phase currents iu, iv, and iw of the PM motor 1 (output currents of the inverter 2).

[0050] Reference numeral 9 denotes a three-phase-dq converter that converts the three-phase currents iu, iv, and iw of the PM motor 1 into d-axis current id and q-axis current iq of a dq coordinate system synchronized with a phase θe that takes into account the number of pole pairs.

[0051] 10, the d-axis current id and the q-axis current iq are determined based on a d-axis current command id * , q-axis current command iq * The current is controlled so that it is equal to the voltage command Vd * , Vq * This is a current controller that calculates

[0052] In addition to the current control, the electrical angular velocity ωe is controlled by the electrical angular velocity command ωe * A configuration for controlling the speed so that the speed becomes equal to the above may be added.

[0053] 11 is a dq axis voltage command Vd output from the current controller 10 * , Vq * The three-phase voltage command vu is calculated using the phase θe that takes into account the number of pole pairs output from the pole pair number calculator 5. * , vv * , VW * and supplies the result to the inverter 2. The inverter 2 applies a voltage to the PM motor 1 according to the voltage command.

[0054] In the first embodiment, a voltage detector (not shown) for detecting the induced voltage Vemf of the PM motor 1 is provided.

[0055] The pole pair number calculator 5, the differential calculator 6, the three-phase-dq converter 9, the current controller 10, and the dq-three-phase converter 11 constitute a controller 20.

[0056] 30 indicates the electrical angular velocity ωe calculated by the differential calculator 6, the d-axis current id and the q-axis current iq converted by the three-phase-dq converter 9, and the dq-axis voltage command Vd calculated by the current controller 10. * , Vq * This is a parameter pre-identification device (parameter identification unit) that identifies the motor parameters, that is, the reference magnetic flux Φn, the temperature coefficient β, the d-axis inductance Ld, and the resistance Ra, based on the induced voltage Vemf of the PM motor 1 measured during no-load operation by a voltage detector not shown, and the magnet temperature measurement value Tpm measured by the magnet temperature sensor 7.

[0057] In the first embodiment, the parameter pre-identifier 30 includes a pre-identifier 31 and a voltage error table calculator 32a (q-axis voltage error calculator) which will be described later (FIG. 3).

[0058] In FIG. 2, which shows the actual machine configuration when estimating the magnet temperature, 40 indicates the electrical angular velocity ωe calculated by the differential calculator 6, the d-axis current id and q-axis current iq converted by the three-phase dq converter 9, and the dq-axis voltage command Vd calculated by the speed / current controller 10. * , Vq * , and the other parts are configured in the same way as in FIG. 1.

[0059] In the first embodiment, the magnet temperature estimator 40 includes a magnet temperature estimator 41, which will be described later (FIG. 4).

[0060] Next, the blocks of the pre-identification phase will be explained with reference to Fig. 3. The pre-identification phase consists of (1) derivation of optimal parameters and (2) derivation of a voltage error table.

[0061] (1) The optimal parameters are derived in the pre-identification device 31 by determining the voltage command value Vd * , Vq * , electrical angular velocity response value ωe, current command value id * , iq *This is done by inputting the current response values ​​id, iq and the magnet temperature response value Tpm, which is the correct data, and outputting the reference magnetic flux Φn, temperature coefficient β, d-axis inductance Ld, and resistance Ra, which are identification parameters that minimize the error between the estimated value and the response value of the magnet temperature.

[0062] The method of deriving the optimal parameters performed by the pre-identification device 31 will be described in detail later with reference to FIGS.

[0063] (2) The voltage error table is derived by calculating the q-axis voltage error ΔVq, which is the difference between the command value and the detected value of the q-axis voltage, in the voltage error table calculator 32a and creating a table.

[0064] The calculation formula by which this voltage error table calculator 32a calculates the ΔVq is shown in the block of FIG.

[0065] In Fig. 3 and Fig. 5, the measured magnet temperature Tpm and the estimated magnet temperature T^ when the voltage error is not taken into account W.0.ΔVq is expressed as equations (1) and (2).

[0066]

number

[0067]

number

[0068] Here, T0 is the reference temperature (the temperature condition where magnetic flux = Φn).

[0069] Then, by taking the difference between equations (1) and (2) and multiplying it by a constant (Φnβωe), we obtain equation (3) for calculating the q-axis voltage error ΔVq.

[0070]

number

[0071] Next, a block diagram of the estimation phase will be described with reference to FIG. 4. In FIG. 4, 41 is a test data (online or offline Vd * , Vq * , id, iq, ωe) and estimates the magnet temperature using the optimal parameters (identification parameters Φn, β, Ld, Ra) derived by the pre-identification device 31 in FIG. 3 and the voltage error ΔVq table calculated by the voltage error table calculator 32a in FIG. 3.

[0072] This magnet temperature estimator 41 calculates the following equation (4) to obtain the magnet temperature estimate value T^.

[0073]

number

[0074] Next, we will explain the derivation of the optimal parameters performed by the pre-identification device 31 in Fig. 3. First, the magnetic flux Φ from the induced voltage Vemf of the PM motor 1 is calculated by equation (5).

[0075]

number

[0076] That is, the pre-identification device 31 obtains the magnetic flux Φ of the equation (5) using the measurement data of the induced voltage Vemf and the electrical angular velocity ωe measured during no-load operation of the PM motor 1.

[0077] The relationship between the magnetic flux Φ obtained by the above equation (5) and the magnet temperature measurement value Tpm measured by the magnet temperature sensor 7 is the straight line shown by the dashed line in Figure 6 (the straight line obtained by the least squares method Φ = aTpm + b; a is the slope of the line, and b is the y-intercept of the line).

[0078] The temperature change of magnetic flux is modeled (defined) as follows:

[0079]

number

[0080] Here, T0 means the reference temperature (Φn temperature condition). The reference magnetic flux Φn and temperature coefficient β are determined so that this equation (6) corresponds to the straight line obtained by the least squares method in Figure 6. If the straight line obtained by the least squares method in Figure 6 is Φ=aTpm+b, then the relationship between Φn and β can be found by comparing the coefficients as shown in the following equation (7).

[0081]

number

[0082] By rearranging the equation, we obtain the following equation (8).

[0083]

number

[0084] The lower part of (8) is a quadratic equation for Φn, and solving this gives us Φn. Substituting this into the upper part of (8) gives us β.

[0085] Then, the following test is carried out using the magnet temperature estimator 41 in FIG. 4 to confirm the validity of Φn, β and the above formula (4).

[0086] In the test, the PM motor 1 is rotated at a constant speed, and the current command values ​​for the d and q axes are set to Idcmd=0 and Iqcmd=0, respectively.

[0087] If the measured data is substituted into equation (4) and the obtained estimated magnet temperature value T^ is close to the measured magnet temperature value Tpm, then the parameters and the estimation equation can be said to be valid.

[0088] Next, a method for identifying the d-axis inductance Ld will be described. First, a constant speed operation test is performed with the d-axis current command value Idcmd<0 and the q-axis current command value Iqcmd=0, or with Idcmd>0 and Iqcmd=0.

[0089] The operating speed is the speed range where a valid temperature estimate was made in the validation check. Then, Idcmd is changed, and the data of each steady-state value, the d-axis voltage reference value Vd * , q-axis voltage reference value vq * , d-axis current id, q-axis current iq, electrical angular velocity ωe, and magnet temperature Tpm are obtained. In this case, for example, the step function shown in Fig. 7 is used. In other words, the command value is increased and maintained for a certain period of time, and this is repeated. If multiple speed conditions are required, the d-axis current command value Idcmd is changed for each speed condition to obtain steady-state value data.

[0090] The steady-state data (vd * , vq * , id, iq, ωe, and steady-state values ​​of Tpm are used to solve the optimization problem using the following evaluation equations (9) to (12).

[0091]

number

[0092] subject to

[0093]

number

[0094]

number

[0095]

number

[0096] T^ is the estimated magnet temperature, Ldn is the d-axis inductance, and ΔVq is the difference between the q-axis voltage command value and the q-axis voltage detection value.

[0097] where equation (9) is an evaluation function that expresses "finding the combination of parameters Ld and ΔVq that minimizes the square error between the actual measured value and the estimated value of the magnet temperature." Equations (10) to (12) are constraints, and respectively represent the temperature estimation formula, the range of Ld, and the range of ΔVq.

[0098] If multiple speed conditions are required, repeat the above process of "obtaining steady-state value data when Idcmd is changed → optimizing" for each speed condition.

[0099] Next, a method for identifying the resistance Ra will be described. First, a constant speed operation test is performed with the d-axis current command value Idcmd=0 and the q-axis current command value Iqcmd>0, or with Idcmd=0 and Iqcmd<0.

[0100] The operating speed is set to the speed range where the above-mentioned appropriate temperature estimation is performed. Then, Iqcmd is changed, and the data of each steady-state value and the d-axis voltage reference value Vd * , q-axis voltage reference value vq * , d-axis current id, q-axis current iq, electrical angular velocity ωe, and magnet temperature Tpm are obtained. In this case, for example, the step function shown in Fig. 7 is used. In other words, the command value is increased and maintained for a certain period of time, and this is repeated. If multiple speed conditions are required, the q-axis current command value Iqcmd is changed for each speed condition to obtain steady-state value data.

[0101] In the case of resistance Ra, the conditions of the operation test are, for example, Id=0, Iq>0, so the steady-state value data vd * , vq * , id, iq, ωe, and Tpm are used to solve the optimization problem using the following evaluation equations (13) to (16).

[0102]

number

[0103] subject to

[0104]

number

[0105]

number

[0106]

number

[0107] T^ is the estimated magnet temperature, Ran is the winding resistance, and ΔVq is the difference between the q-axis voltage command value and the q-axis voltage detection value.

[0108] where equation (13) is an evaluation function that expresses "finding the combination of parameters Ran and ΔVq that minimizes the square error between the actual measured value and the estimated value of the magnet temperature." Equations (14) to (16) are constraints, and respectively represent the temperature estimation formula, the range of Ran, and the range of ΔVq.

[0109] If multiple speed conditions are required, repeat the above process of "obtaining steady-state value data when Iqcmd is changed → optimization" for each speed condition.

[0110] As described above, according to the first embodiment, magnet temperature estimation with a certain degree of accuracy is achieved, including the current condition where the d-axis current id ≠ 0 and the q-axis current iq ≠ 0. In addition, since the motor parameters derived by the pre-identification device 31 are used, magnet temperature estimation with high accuracy can be performed. EXAMPLES

[0111] A block diagram of the pre-identification phase in the second embodiment is shown in Fig. 8, and a block diagram of the estimation phase is shown in Fig. 9. The actual device configuration is similar to that in Figs. 1 and 2, but in the pre-identification phase of the second embodiment, it is not necessary to derive optimal parameters by the pre-identification device 31, and only the derivation of a voltage error table is performed as in the voltage error table calculator 32b in Fig. 8.

[0112] The voltage error table calculator 32b calculates the voltage command value Vd* , Vq * , electrical angular velocity response value ωe, current command value id * , iq * The current response values ​​id, iq, and the magnet temperature response value Tpm, which is the correct data, are input, and the reference magnetic flux Φn, temperature coefficient β, d-axis inductance Ld, and resistance Ra, which are the design parameters of the PM motor, are used to calculate the q-axis voltage error value ΔVq using the above equations (1), (2), and (3). The calculated value is tabulated.

[0113] 42 in FIG. 9 shows the test data (online or offline Vd * , Vq * , id, iq, ωe) and calculates the above equation (4) using the voltage error ΔVq table derived by the voltage error table calculator 32b in FIG. 8 to estimate the magnet temperature.

[0114] As described above, according to the second embodiment, the same effects as those of the first embodiment can be obtained, and further, the effort of deriving optimal parameters is unnecessary, and the configuration is simplified. EXAMPLES

[0115] In the present embodiment 3, the configuration of the q-axis voltage error calculation unit is different as described below. That is, the voltage error ΔVq table substituted in the estimation phase (magnet temperature estimators 41, 42) is not the voltage error ΔVq table calculated in the pre-identification phase (voltage error table calculators 32a, 32b), but uses an equation of a curved surface derived by curved surface fitting, and other parts are the same as those in the first and second embodiments. The equation for curved surface fitting is obtained in advance as follows: Using the voltage error table as input data, the optimization problem of the following equation (17) is solved to obtain each parameter of the curved surface (the mesh surface in FIG. 10).

[0116]

number

[0117] a is the parameter vector of the surface equation, a *is the optimized parameter vector, ΔVq_ value is the value in the voltage error table, f is a function of the surface, and argmin is the set of values ​​at which the minimum point is reached. According to equation (17), ΔVq_ value -Derive the parameter vector of the surface equation that minimizes f(Id, Iq, a).Then, use the derived parameter vector to formulate the surface fitting equation as shown in Figure 10.

[0118] During actual operation, Id and Iq are input into this surface fitting equation to calculate the q-axis voltage error ΔVq online.

[0119] FIG. 10 shows an example of fitting implementation, where the dots represent table data and the mesh-like surface represents the fitted surface.

[0120] As described above, according to the third embodiment, by using the equation for surface fitting, it is not necessary to hold the voltage error table in memory, and therefore the memory capacity can be reduced.

[0121] The results of the actual verification of the present invention are shown in Fig. 11. In Fig. 11, (a) shows the estimated temperature value, (b) shows the measured temperature value, and (c) shows the temperature estimation error.

[0122] According to FIG. 11, it can be seen that magnet temperature estimation with a certain degree of accuracy is achieved under current conditions including Id≠0 and Iq≠0. [Explanation of symbols]

[0123] 1...PM motor 2...Inverter 3. Load 4...Rotational position sensor 5...Pole pair calculator 6... Differential calculator 7…Magnet temperature sensor 8…Current sensor 9...3 phase-dq converter 10...Current controller 11…dq-3 phase converter 20…Controller 30...Parameter pre-identification device 31...Pre-identification device 32a, 32b...Voltage error table calculator 40,41,42...Magnet temperature estimator

Claims

1. In a PM motor driven by an inverter equipped with a current control system, a q-axis voltage error calculation unit that calculates a q-axis voltage error value ΔVq, which is a difference between a command value and a detected value of a q-axis voltage, based on a d-axis current id and a q-axis current iq obtained by converting the detected current of the PM motor into a d-axis and q-axis coordinate system; A voltage command Vd obtained by performing current control so that the motor detection current, which is the current flowing through the PM motor, becomes equal to the current command. * , Vq * a magnet temperature estimation unit that calculates a magnet temperature estimate based on the d-axis current id, the q-axis current iq, and an electrical angular velocity ωe calculated by detecting the d-axis current id, the q-axis current iq, and the rotation speed of the PM motor, and a q-axis voltage error value ΔVq calculated by the q-axis voltage error calculation unit.

2. The voltage command Vd * , Vq * , electrical angular velocity ωe, d-axis current command id * , q-axis current command iq * a parameter identification unit that derives a reference magnetic flux Φn, a temperature coefficient β, a d-axis inductance Ld, and a resistance Ra, which are motor parameters that minimize an error between an estimated value and a measured value of the magnet temperature, based on a d-axis current id, a q-axis current iq, and a magnet temperature measurement value Tpm obtained by measuring the magnet temperature of the PM motor; The q-axis voltage error calculation unit * , Vq * , ωe, i.d * , iq * , id, iq, Tpm, and the motor parameters Φn, β, Ld, Ra derived by the parameter identification unit, by calculating equations (1), (2), and (3) to calculate a q-axis voltage error value ΔVq, [0010] [0025] [0030] (T^ W.0.ΔVq is the magnet temperature estimate without considering the voltage error, T 0 is the reference temperature, Tpm is the measured magnet temperature) The magnet temperature estimation unit is a test data Vd * , Vq * , id, iq, ωe, the motor parameters Φn, β, Ld, Ra derived by the parameter identification unit, and the q-axis voltage error value ΔVq calculated by the voltage error table calculator, to calculate equation (4) to determine a magnet temperature estimate. [0045] (T^ is the estimated magnet temperature, T 0 is the reference temperature)

3. The q-axis voltage error calculation unit * , Vq * , ωe, i.d * , iq * , id, iq, and Tpm, and a voltage error table calculator that calculates a q-axis voltage error value ΔVq by calculating equations (1), (2), and (3) based on the reference magnetic flux Φn, temperature coefficient β, d-axis inductance Ld, and resistance Ra, which are design parameters of the PM motor; [0010] [0025] [0030] (T^ W.0.ΔVq is the magnet temperature estimate without considering the voltage error, T 0 is the reference temperature, Tpm is the measured magnet temperature) The magnet temperature estimation unit is a test data Vd * , Vq * , id, iq, ωe, and the q-axis voltage error value ΔVq calculated by the voltage error table calculator, to calculate equation (4) to determine an estimated magnet temperature. [0045] (T^ is the estimated magnet temperature, T 0 is the reference temperature)

4. The q-axis voltage error calculation unit calculates the value of the voltage error table as ΔVq_ value , f is a function of the surface, and a is a parameter vector of the surface equation, ΔVq_ value -Derive the parameter vector of the equation of the surface that minimizes f(Id, Iq, a) by solving the optimization problem using equation (17). ##EQU00017## (a * is the optimized parameter vector, argmin is the set of values ​​at which the minimum point is reached) The temperature estimation device for a PM motor according to claim 1, characterized in that a surface fitting equation is formulated using a parameter vector derived by equation (17), and the d-axis current Id and the q-axis current Iq during actual operation are input into equation (17) to calculate the q-axis voltage error value ΔVq.

5. the parameter identification unit calculates a magnetic flux Φ (=Vemf / ωe) using measurement data of an induced voltage Vemf of the PM motor measured during no-load operation of the PM motor and an electrical angular velocity ωe obtained by detecting a rotation speed of the PM motor; a line Φ=aTpm+b (a is the slope of the line, and b is the y-intercept of the line) obtained by the least squares method, which expresses the relationship between the magnet temperature measurement value Tpm obtained by measuring the magnet temperature of the PM motor and the calculated magnetic flux Φ, corresponds to Φ=Φn(1+β(Tpm-T0)) (Φn is a reference magnetic flux, β is a temperature coefficient, and T0 is a reference temperature), which defines the temperature change of the magnetic flux, and Φn and β are identified as motor parameters; a d-axis current command value Idcmd and a q-axis current command value Iqcmd are each set to 0, and a constant speed operation test of the PM motor is performed; and the operating speed when the magnet temperature estimated value estimated by the magnet temperature estimator is close to the magnet temperature measurement value is determined to be a speed region where a valid temperature estimation is performed; A constant speed operation test is carried out in the determined speed range with the d-axis current command value Idcmd<0 and the q-axis current command value Iqcmd=0, or Idcmd>0 and Iqcmd=0, and Vd at each Idcmd when Idcmd is changed stepwise is calculated. * , Vq * , id, iq, ωe, and Tpm, and based on the acquired steady-state data, an optimization problem is solved using an evaluation formula for finding a combination of parameters, d-axis inductance Ld and q-axis voltage error ΔVq, which minimizes the square error between the measured value and the estimated value of the magnet temperature, as shown in formulas (9) to (12), to identify the d-axis inductance Ld; [0090] subject to [0010] ##EQU00011## ##EQU00012## (T^ is the estimated magnet temperature, Ldn is the d-axis inductance, and ΔVq is the difference between the q-axis voltage command value and the q-axis voltage detection value.) A constant speed operation test is performed in the speed range where the above-mentioned determined appropriate temperature estimation is performed, with the d-axis current command value Idcmd=0 and the q-axis current command value Iqcmd>0, or Idcmd=0 and Iqcmd<0, and Vd at each Iqcmd when Iqcmd is changed in a stepwise manner is calculated. * , Vq * , id, iq, ωe, and Tpm, and based on the acquired steady-state data, Ran is identified by solving an optimization problem using evaluation equations shown in (13) to (16) that find a combination of parameter resistance Ran and q-axis voltage error ΔVq that minimizes the square error between the measured value and the estimated value of the magnet temperature. ##EQU00013## subject to ##EQU00014## ##EQU00015## ##EQU00016## (T^ is the estimated magnet temperature, Ran is the winding resistance of the PM motor, and ΔVq is the difference between the q-axis voltage command value and the q-axis voltage detection value.)

6. A method for estimating a temperature of a PM motor driven by an inverter having a current control system, comprising: a q-axis voltage error calculation step in which a q-axis voltage error calculation unit calculates a q-axis voltage error value ΔVq, which is a difference between a command value and a detected value of the q-axis voltage, based on a d-axis current id and a q-axis current iq obtained by coordinate-transforming the detected current of the PM motor into the dq axes; The magnet temperature estimation unit detects the current flowing through the PM motor, which is test data, and performs current control so that the motor detection current becomes equal to the current command. * , Vq * a magnet temperature estimation step of calculating an estimated magnet temperature value based on the d-axis current id, the q-axis current iq, and the electrical angular velocity ωe calculated by detecting the d-axis current id, the q-axis current iq, and the rotation speed of the PM motor, and the q-axis voltage error value ΔVq calculated by the q-axis voltage error calculation unit.

Citation Information

Patent Citations

  • Controller and control method for electric vehicle

    JP1999018496A

  • Driving method and temperature-estimating method of pm motor

    JP2005218215A

  • Estimation device of temperature of rotor magnet in permanent magnet synchronous motor

    JP2021158868A

  • Power conversion device

    JP2022178400A

  • Methods of estimating rotor magnet temperature and systems thereof

    US20160091568A1