JP7632245B2 - Magnet temperature estimation device - Google Patents

Description

The present invention relates to a magnet temperature estimation device.

For example, the following non-patent document 1 describes estimating the temperature of a motor's permanent magnets using machine learning. In particular, the document describes that, in addition to current and voltage, the input variables to the trained model include their exponential moving average values (paragraph before equation (5)).

Wilhelm Kirchg¨assner, Member, IEEE, Oliver Wallscheid, Member, IEEE and Joachim B¨ocker, Senior Member, IEEE‘Data-Driven Permanent Magnet Temperature Estimation in Synchronous Motors with Supervised Machine Learning’arXiv:2001.06246v1

The inventors considered using the detection value of a temperature sensor placed near the motor to estimate the temperature of the motor's permanent magnets. They found that in this case, the output value of the trained model fluctuates significantly.

Means for solving the above problems and their effects will be described below.
The device is equipped with an execution device and a memory device, and estimated mapping data that defines an estimated mapping is stored in the memory device, and the estimated mapping is a mapping that outputs a magnet temperature of a rotating electric machine in response to the input of input variables. The execution device executes an acquisition process that acquires detection values of a temperature sensor provided in the rotating electric machine, an average calculation process that calculates the average value of a plurality of the detected values, and a calculation process that calculates the magnet temperature by inputting the input variables into the estimated mapping, and the input variables include the average value. This is a magnet temperature estimation device.

The temperature detected by the temperature sensor installed in the rotating electric machine has a strong positive correlation with the temperature of the magnet of the rotating electric machine. However, the temperature sensor itself tends to have a smaller heat capacity than the magnet of the rotating electric machine. Therefore, the detected value of the temperature sensor reacts sensitively to thermal disturbances, and the temperature change on a small time scale is larger than the temperature of the magnet. Therefore, when the detected value of the temperature sensor is used as the input to the estimation mapping, the value of the output variable of the estimation mapping tends to change larger on a small time scale, just like the detected value of the temperature sensor. Therefore, in the above configuration, the average value of the detected value of the temperature sensor is used as the input to the estimation mapping. The average value is a variable whose change on a small time scale is suppressed compared to the detected value. Therefore, in the above configuration, the magnet temperature can be estimated with high accuracy.

FIG. 1 is a diagram illustrating a configuration of a vehicle according to an embodiment. 4 is a flowchart showing a procedure of a process executed by a control device according to the embodiment. 5A and 5B are diagrams illustrating an example of a transition between the detected rotor temperature and the magnet temperature according to the embodiment; 6A and 6B are diagrams illustrating the transition of magnet temperature in the same embodiment and a comparative example.

Hereinafter, an embodiment will be described with reference to the drawings.
1 shows the configuration of a vehicle according to this embodiment. A crankshaft 12 of an internal combustion engine 10 is mechanically connected to a carrier C of a planetary gear mechanism 20 serving as a power split device. A rotating shaft 32 of a first motor generator 30 is mechanically connected to a sun gear S of the planetary gear mechanism 20. An output voltage of a first inverter 34 is applied to a terminal of the first motor generator 30. A rotating shaft 42 of a second motor generator 40 is mechanically connected to a ring gear R of the planetary gear mechanism 20. The second motor generator 40 includes a permanent magnet 44 attached to a rotor and a stator coil 46. An output voltage of a second inverter 48 is applied to the stator coil 46. A driving wheel 50 is further mechanically connected to the ring gear R of the planetary gear mechanism 20.

The control device 60 operates the operating section of the internal combustion engine 10 to control the torque, exhaust component ratio, and other controlled variables of the internal combustion engine 10 as a controlled object. The control device 60 also operates the first inverter 34 to control the torque, which is a controlled variable of the first motor generator 30 as a controlled object. The control device 60 also operates the second inverter 48 to control the torque, which is a controlled variable of the second motor generator 40 as a controlled object.

When controlling the controlled variable, the control device 60 refers to the stator temperature Tst, which is the temperature of the stator coil 46 of the second motor generator 40 detected by the stator temperature sensor 70. The control device 60 also refers to the output signal Sm2 of the rotation angle sensor 72 that detects the rotation angle of the rotating shaft 42 of the second motor generator 40. The control device 60 also refers to the currents iu, iv, and iw flowing through the second motor generator 40 detected by the current sensor 74. The control device 60 also refers to the ATF temperature Tatf, which is the temperature of the lubricating oil detected by the ATF temperature sensor 76. The lubricating oil circulates through the sliding parts of the drive system. In detail, the lubricating oil comes into contact with the planetary gear mechanism 20 and the rotating shaft 42 of the second motor generator 40. The control device 60 also refers to the water temperature Tcl, which is the temperature of the cooling water detected by the water temperature sensor 78. The cooling water exchanges heat with at least one of the second inverter 48 and the lubricating oil.

The control device 60 includes a PU 62 and a storage device 64. The PU 62 is a software processing device that includes at least one of a CPU, a GPU, and a TPU. A demagnetization determination program 64a is stored in the storage device 64. The demagnetization determination program 64a has instructions to execute a process to determine whether or not there is a risk of so-called demagnetization occurring, in which the permanent magnet 44 becomes too hot and the magnetic force is reduced.

Figure 2 shows the procedure for the process related to demagnetization judgment. The process shown in Figure 2 is realized by the PU 62 repeatedly executing the demagnetization judgment program 64a. Note that, below, the step number of each process is represented by a number preceded by "S".

In the series of processes shown in FIG. 2, the PU 62 first acquires the stator temperature Tst, the ATF temperature Tatf, the water temperature Tcl, the torque Trq, and the rotation speed ωm2 (S10). The torque Trq is calculated by the PU 62 based on the currents iu, iv, and iw. The rotation speed ωm2 is calculated by the PU 62 based on the output signal Sm2.

Next, PU 62 updates the sampling numbers of stator temperature Tst, ATF temperature Tatf, and water temperature Tcl (S12). That is, for example, when variable i is an integer equal to or greater than 0, PU 62 sets "Tatf(i)" to "Tatf(i+1)" and then substitutes the ATF temperature Tatf obtained by the processing of S10 for "Tatf(0)". This is a process to ensure that the more recent the sampling timing, the smaller the sampling number indicated by the number in parentheses.

Next, the PU 62 calculates the stator temperature average value Tstave, the ATF temperature average value Tatfave, and the water temperature average value Tclave (S14). Each of these average values is a simple average of the most recent "N" sampled values.

Next, the PU 62 determines whether the average stator temperature Tstave is equal to or greater than the high temperature threshold TthL (S16). This process is for determining whether or not the temperature of the permanent magnet 44 is in a state where there is a possibility that it will rise to a temperature at which demagnetization may occur. The high temperature threshold TthL itself is set to a temperature lower than the temperature at which demagnetization may occur. The value obtained by adding the high temperature threshold TthL to the maximum value assumed for the difference between the average stator temperature Tstave and the actual temperature of the permanent magnet 44 is a value that corresponds to the lower limit of the temperature at which demagnetization may occur. In more detail, it is desirable for this added value to be a value obtained by subtracting a predetermined margin from the lower limit.

When the PU 62 determines that the high temperature threshold TthL is equal to or higher than the high temperature threshold TthL (S16: YES), the PU 62 assigns the value calculated by the process of S14 to the input variable for the estimated mapping (S18). The estimated mapping is a mapping defined by the estimated mapping data 64b stored in the storage device 64 shown in FIG. 1. The PU 62 assigns the stator temperature average value Tstave to the input variable x(1). The PU 62 also assigns the ATF temperature average value Tatfave to the input variable x(2). The PU 62 also assigns the water temperature average value Tclave to the input variable x(3). The PU 62 also assigns the torque Trq to the input variable x(4). The PU 62 also assigns the rotation speed ωm2 to the input variable x(5).

Then, PU 62 calculates the value of output variable y by substituting input variables x(1) to x(5) into the estimated mapping (S20). In this embodiment, the estimated mapping is a forward propagation fully connected neural network with one intermediate layer. The output variable y is an estimated value of the temperature of permanent magnet 44. The training data used when generating estimated mapping data 64b is the data used in the process of S18 and the temperature of permanent magnet 44. The temperature of permanent magnet 44 is calculated using a heat diffusion model that inputs more variables than the variables used in the process of S18.

The PU 62 determines whether the value of the output variable y is equal to or greater than the demagnetization threshold TthH (S22). The demagnetization threshold TthH is a value that corresponds to the lower limit of the temperature at which demagnetization occurs. In particular, it is desirable to set the value to the lower limit minus a predetermined margin.

When the PU 62 determines that the temperature of the permanent magnet 44 is equal to or higher than the demagnetization threshold TthH (S22: YES), the PU 62 determines that the temperature of the permanent magnet 44 is in a state in which demagnetization may occur (S24). Note that when the PU 62 completes the process of S24 or when the process of S16 or S22 is determined to be negative, the PU 62 temporarily ends the series of processes shown in FIG. 2.

When it is determined that a situation exists in which demagnetization may occur, the PU 62 executes a process of limiting the output of the second motor-generator 40 to a smaller output.
Here, the operation and effects of this embodiment will be described.

Figure 3(a) shows the progress of the stator temperature Tst and the temperature Tmg of the permanent magnet 44. As shown in Figure 3(a), the stator temperature Tst has a strong positive correlation with the temperature Tmg of the permanent magnet 44. However, compared to the temperature Tmg of the permanent magnet 44, the stator temperature Tst fluctuates greatly on a short time scale. This is thought to be because the heat capacity of the stator temperature sensor 70 is smaller than the heat capacity of the permanent magnet 44, and therefore the temperature of the stator temperature sensor 70 is easily affected by thermal disturbances. In contrast, as shown in Figure 3(b), the fluctuation of the stator temperature average value Tstave on a short time scale is suppressed compared to the stator temperature Tst.

When estimating the temperature Tmg of the permanent magnet 44, the PU 62 uses the average stator temperature Tstave instead of the stator temperature Tst. This allows the temperature of the permanent magnet 44 to be estimated with high accuracy, as shown in FIG. 4(a). In contrast, when the input to the estimation map is the stator temperature Tst, the output variable y0 of the estimation map fluctuates significantly on a short time scale, as shown in FIG. 4(b).

In this manner, in this embodiment, the temperature Tmg of the permanent magnet 44 can be estimated with high accuracy while utilizing the stator temperature Tst.
According to the present embodiment described above, the following actions and effects can be obtained.

(1) The torque Trq of the second motor generator 40 is included as an input variable to the estimation mapping. The torque Trq is a variable that has a strong positive correlation with the current flowing through the second motor generator 40. On the other hand, the current is a variable that has a positive correlation with the amount of heat generated by the second motor generator 40. Therefore, by including the torque Trq as an input variable to the estimation mapping, the temperature of the permanent magnet 44 can be estimated with higher accuracy than when it is not included.

(2) In addition to torque Trq, rotation speed ωmg is included as an input variable to the estimation mapping. The two paired variables of torque Trq and rotation speed ωmg contain information on the output of the second motor generator 40. Furthermore, the efficiency of the second motor generator 40 is determined according to the operating point of the second motor generator 40 defined by torque Trq and rotation speed ωmg. Therefore, the above two variables contain information on the efficiency with which the output of the second motor generator 40 is obtained. Therefore, information on the amount of heat generated to drive the second motor generator 40 is included. Therefore, by including the above two variables as input variables to the estimation mapping, the temperature of the permanent magnet 44 can be estimated with higher accuracy than when they are not included.

(3) The ATF temperature average value Tatfave was included in the input variables to the estimation mapping. Because the lubricating oil is a fluid that exchanges heat with components near the second motor-generator 40, the temperature of the lubricating oil is useful information for understanding the temperature of the permanent magnet 44. Therefore, the temperature of the permanent magnet 44 can be estimated with higher accuracy compared to a case in which the ATF temperature average value Tatfave is not included in the input variables to the estimation mapping.

(4) The input variables to the estimation mapping include the average water temperature Tclave. Because the cooling water is a member that exchanges heat with at least one of the hydraulic oil and the second inverter 48, the temperature of the cooling water is useful information for understanding the temperature of the permanent magnet 44. Therefore, the temperature of the permanent magnet 44 can be estimated with higher accuracy compared to a case in which the input variables to the estimation mapping do not include the average water temperature Tclave.

<Correspondence>
The correspondence between the matters in the above embodiment and the matters described in the "Means for solving the problem" section above is as follows: The execution device corresponds to the PU 62. The storage device corresponds to the storage device 64. The estimated mapping data corresponds to the estimated mapping data 64b. The temperature sensor corresponds to the stator temperature sensor 70. The acquisition process corresponds to the process of S10. The average value calculation process corresponds to the process of S14. The calculation process corresponds to the processes of S18 and S20.

<Other embodiments>
This embodiment can be modified as follows: This embodiment and the following modifications can be combined with each other to the extent that no technical contradiction occurs.

The neural network is not limited to a neural network with one intermediate layer. For example, the neural network may have two or more intermediate layers.
The neural network is not limited to a forward propagation fully connected network. For example, a recurrent neural network (RNN) may be used. In this case, it is preferable to delete the process of S16.

The estimation map is not limited to a neural network. For example, a linear regression model may be used.
The state variable indicating the state of the second motor generator 40 is not limited to the torque Trq and the rotation speed ωm2 as variables indicating the load. For example, instead of the torque Trq, the current flowing through the second motor generator 40 may be used as the variable indicating the load. Also, for example, the variable indicating the load may be a command value for the torque Trq or a command value for the current.

Also, for example, it may be a pair of the output of the second motor generator 40 and the rotational speed ωm2. Here, the torque of the second motor generator 40 can be calculated from the output and the rotational speed ωm2. Therefore, the output and the rotational speed ωm2 include information about the operating point of the second motor generator 40. Also, for example, the state variable indicating the state of the second motor generator 40 may be a pair of a variable indicating a load and the output. Here, the rotational speed ωm2 can be calculated from the variable indicating the load and the output. Therefore, the variable indicating the load and the output include information about the operating point of the second motor generator 40.

However, it is not essential that the state variables indicating the state of the second motor-generator 40 include any two of the three variables: the variable indicating the load, the rotational speed ωm2, and the output. For example, with regard to those three variables, it may include only the variable indicating the load. Also, with regard to those three variables, it may include only the rotational speed ωm2.

The temperature sensor provided in the second motor generator 40 is not limited to the stator temperature sensor 70. For example, it may be a temperature sensor attached to the outer circumferential surface of the housing of the second motor generator 40 to detect the temperature of the housing.

- In the above embodiment, the sampling period of the stator temperature Tst etc. for calculating the stator temperature average value Tstave etc. is made to match the calculation period of the magnet temperature estimate value, but this is not limited thereto. For example, the sampling period may be shorter than the calculation period.

In the above embodiment, the stator temperature average value Tstave is a simple average value of the stator temperature Tst, but the present invention is not limited to this. For example, the stator temperature average value Tstave may be an exponential moving average value.
In the above embodiment, the ATF temperature average value Tatfave is included in the input variable x to the estimation mapping, but this is not limiting. For example, the ATF temperature Tatf may be included.

In the above embodiment, the average water temperature Tclave is included in the input variable x to the estimation mapping, but this is not limiting. For example, the water temperature Tcl may be included.
The inputs to the estimation mapping are not limited to the variables exemplified in the above embodiment and the above modified example. For example, variables related to the temperature of the lubricating oil, such as the ATF temperature average value Tatfave, may be deleted. Also, for example, variables related to the temperature of the coolant, such as the water temperature average value Tclave, may be deleted. Also, for example, if the permanent magnet 44 is significantly affected by the heat from the internal combustion engine 10, a variable indicating the state of the internal combustion engine 10 may be added to the input variables to the estimation mapping. Also, for example, if the permanent magnet 44 is significantly affected by the vehicle speed, a variable indicating the vehicle speed may be added to the input variables to the estimation mapping.

-Vehicles are not limited to series-parallel hybrid vehicles.

REFERENCE SIGNS LIST 10 internal combustion engine 12 crankshaft 20 planetary gear mechanism 30 first motor generator 34 first inverter 40 second motor generator 44 permanent magnet 46 stator coil 48 second inverter 50 drive wheels 60 control device

Claims (1)

An execution device and a storage device,
The storage device stores estimated mapping data that defines an estimated mapping,
The estimation mapping is a mapping that outputs a magnet temperature of a rotating electric machine in response to an input of an input variable,
The execution device is
an acquisition process for acquiring a detection value of a temperature sensor provided in the rotating electric machine and a rotation speed of the rotating electric machine ;
an average value calculation process for calculating an average value of the plurality of detection values;
a calculation process for calculating the magnet temperature by inputting the input variables into the estimation mapping;
The magnet temperature estimator, wherein the input variables include the average value and the rotational speed .

Images