Magnet temperature estimation device

Abstract

To provide a magnet temperature estimation device capable of estimating a temperature of a magnet installed in a rotor of a power generator in a wind power generator.SOLUTION: A magnet temperature estimation device 50 includes a processing device 51 and a storage device 52. The storage device 52 of the magnet temperature estimation device 50 stores an estimation model 53 trained by supervised learning so that a temperature of a magnet installed in a rotor of a power generator is outputted when time-series data including wind velocity of wind which a blade receives and a pitch angle of the blade as explanatory variables are inputted. In the magnet temperature estimation device 50, the processing device 51 estimates the temperature of the magnet by acquiring time-series data and inputting data to the estimation model 53.SELECTED DRAWING: Figure 4

Description

[Technical field]

[0001] The present invention relates to a magnet temperature estimating device. [Background technology]

[0002] Patent Document 1 discloses a magnet temperature estimation device. This magnet temperature estimation device uses a trained model that has been trained with a supervised learning using teacher data including time-series sensor data related to the rotation of an electric motor of a vehicle and magnet temperature data. The magnet temperature estimation device uses the trained model to estimate the temperature of the magnet of the rotor of the electric motor using the time-series sensor data received from the electric motor as an input. [Prior art documents] [Patent documents]

[0003] [Patent Document 1] JP 2022-168480 A Summary of the Invention [Problem to be solved by the invention]

[0004] Wind turbines generate electricity by rotating a generator using the wind that hits the blades. A generator, like an electric motor, consists of a stator and a rotor equipped with magnets. As the temperature of a magnet increases, its magnetic force decreases. As the magnetic force decreases, the amount of electricity generated decreases. Therefore, in order to operate a wind turbine efficiently, it is necessary to know the temperature of the generator's magnets. [Means for solving the problem]

[0005] The means for solving the above problems and their effects will be described below. The magnet temperature estimation device for solving the above problem is a magnet temperature estimation device for a wind power generator in which wind received by the blades rotates the shaft and is transmitted to the generator. The magnet temperature estimation device includes a processing device and a storage device. In the magnet temperature estimation device, an estimation model trained by supervised learning is stored in the storage device so as to output the temperature of a magnet installed in the rotor of the generator when time series data including the wind speed of the wind received by the blades and the pitch angle of the blades as explanatory variables is input. In this magnet temperature estimation device, the processing device estimates the temperature of the magnet by acquiring the time series data and inputting it into the estimation model. Effect of the Invention

[0006] The magnet temperature estimation device can estimate the temperature of a magnet installed in the rotor of a generator of a wind turbine generator. [Brief description of the drawings]

[0007] [Figure 1] FIG. 1 is a schematic diagram showing the configuration of a wind power generator equipped with a magnet temperature estimation device according to one embodiment. [Diagram 2] FIG. 2 is a cross-sectional view of the blade taken along line 2-2 in FIG. [Diagram 3] FIG. 3 is a schematic diagram showing the configuration of the second motor generator of the wind power generator. [Figure 4] FIG. 4 is a schematic diagram showing the configuration of a magnet temperature estimation device according to an embodiment and connections of each part of a wind power generator. [Diagram 5] FIG. 5 is a schematic diagram showing the structure of a convolutional neural network. [Figure 6] FIG. 6 is a table showing time-series data of explanatory variables input to an estimation model used by the magnet temperature estimation device of the embodiment. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0008] Hereinafter, one embodiment of a magnet temperature estimation device will be described with reference to FIGS. <Configuration of wind power generator 10> Fig. 1 shows the configuration of a wind power generator 10 equipped with a magnet temperature estimation device 50. As shown in Fig. 1, the wind power generator 10 is composed of a nacelle 11 and components installed on the outside of the nacelle 11. The components installed on the outside of the nacelle 11 are blades 13, a hub 14, a pitch angle control device 15, a shaft 16, and a pitch angle control mechanism 19.

[0009] A plurality of blades 13 are installed on the wind power generator 10. Each blade 13 is connected to a hub 14. When the blades 13 receive wind blowing from the front of the wind power generator 10, the hub 14 rotates.

[0010] A pitch angle control device 15 is installed on the hub 14. In addition, a pitch angle control mechanism 19 is installed between the blades 13 and the hub 14. The pitch angle control device 15 and the pitch angle control mechanism 19 are connected so that they can communicate with each other. The pitch angle control device 15 controls the pitch angle of the blades 13 through the pitch angle control mechanism 19.

[0011] The shaft 16 connects the hub 14 to a gearbox 18 inside the nacelle 11. The shaft 16 rotates together with the hub 14, thereby transmitting the rotation of the hub 14 to the gearbox 18. In this manner, the hub 14 and the inside of the nacelle 11 are connected through the shaft 16.

[0012] The nacelle 11 houses therein the yaw control motor 12, a gearbox 18, and a hybrid system (described later). Inside the nacelle 11, power is generated using the rotation transmitted from the shaft 16.

[0013] The yaw control motor 12 controls the orientation of the nacelle 11 in accordance with the direction of the wind blowing around the wind power generator 10. This enables the wind power generator 10 to receive the wind from the front.

[0014] The speed increaser 18 is made up of a plurality of gears. The speed increaser 18 increases the speed of the rotation transmitted from the shaft 16. <Configuration of the hybrid system in which the wind turbine 10 is used> The wind power generator 10 is a repurposed hybrid system of a hybrid vehicle. The hybrid system is made up of a drive shaft 20, a transaxle, a first motor generator 33, an inverter 34, and a second motor generator 100.

[0015] The drive shaft 20 connects the speed increaser 18 and the transaxle. The drive shaft 20 transmits the rotation, the speed of which has been increased by the speed increaser 18, to the transaxle. The transaxle is made up of a differential gear 23, a first drive gear 24, a first driven gear 25, a second drive gear 26, a second driven gear 27, a speed-increasing gear 28, and a planetary gear mechanism 32.

[0016] In the transaxle, the differential gear 23 is connected to the drive shaft 20. The differential gear 23 is also connected to a first drive gear 24. The differential gear 23 transmits the rotation of the drive shaft 20 to the first drive gear 24.

[0017] A first driven gear 25 meshes with the first drive gear 24. A second drive gear 26 is connected to the first driven gear 25 so as to be able to rotate integrally with the first driven gear 25. The rotation transmitted from the differential gear 23 is transmitted in the order of the first drive gear 24, the first driven gear 25, and the second drive gear 26.

[0018] The second drive gear 26 is meshed with a second driven gear 27. The second driven gear 27 is connected to a planetary gear mechanism 32. The planetary gear mechanism 32 has three rotating elements: a ring gear 29, a planetary carrier 30, and a sun gear 31. The ring gear 29 is connected to the second driven gear 27 so as to be rotatable together with the second driven gear 27. The sun gear 31 is connected to a first rotating shaft 37 of the first motor generator 33. In this manner, the second drive gear 26 and the first motor generator 33 are connected in the order of the second driven gear 27, the ring gear 29, the planetary carrier 30, and the sun gear 31.

[0019] The rotation transmitted to the second driven gear 27 is normally finally transmitted to the first motor generator 33 via the sun gear 31. However, as will be described later, in this embodiment, the first motor generator 33 is not used as a generator in the wind power generator 10. In other words, the first motor generator 33 does not contribute to the power generation of the wind power generator 10. Therefore, from the viewpoint of energy efficiency, it is desirable that the rotation is not transmitted to the first motor generator 33. Therefore, a shaft lock is applied to the first rotating shaft 37 of the first motor generator 33 to which the sun gear 31 is connected.

[0020] The second drive gear 26 is also meshed with the speed-increasing gear 28. The speed-increasing gear 28 is connected to a second rotating shaft 104 of the second motor-generator 100. The speed-increasing gear 28 transmits the rotation transmitted from the second drive gear 26 to the second motor-generator 100 while increasing the speed.

[0021] In this manner, the rotation transmitted from the drive shaft 20 is finally transmitted to the second motor generator 100 via the transaxle. The first motor generator 33 and the second motor generator 100 have the same structure. Both the first motor generator 33 and the second motor generator 100 can generate electricity by receiving rotation from an external source. In this embodiment, the second motor generator 100 functions as a generator in the wind power generator 10. On the other hand, the first motor generator 33 is not used as a generator.

[0022] The inverter 34 is electrically connected to the first motor generator 33 and the second motor generator 100. The inverter 34 converts the AC current generated by the second motor generator 100 into a DC current. The inverter 34 also controls the second motor generator 100 and adjusts the amount of power generated by the second motor generator 100. On the other hand, although the inverter 34 is connected to the first motor generator 33, it does not convert the current or control the amount of power generated by the first motor generator 33.

[0023] In this way, the wind power generator 10 uses the wind received by the blades 13 to drive the second motor generator 100 and generate electricity. <Control of the Pitch Angle of the Blade 13 by the Pitch Angle Control Device 15> FIG. 2 shows a cross-sectional view of the blade 13 receiving wind from the direction of wind direction 40 in FIG. 1, taken along line 2-2 in FIG.

[0024] When blade 13 receives wind from wind direction 40, slow wind passes above blade 13 in Fig. 2, and fast wind passes below. Slow wind has high pressure, and fast wind has low pressure. Due to this pressure difference, blade 13 receiving wind from wind direction 40 generates lift 42 in the direction indicated by the dotted arrow.

[0025] Such lift force 42 causes blade 13 to move in the blade movement direction 44. Then, as hub 14 connected to blade 13 rotates, a rotation plane 45 including the vector of blade movement direction 44 is formed.

[0026] A relative wind direction 41 seen from the blade 13 is determined by the wind speed of the wind blowing from the direction of wind direction 40 and the speed at which the blade 13 moves in the blade movement direction 44, i.e., the rotational speed of the hub 14. At this time, a lift force 42 is generated in a direction perpendicular to the relative wind direction 41.

[0027] In addition to lift 42, drag 43 is also generated on blade 13 when it receives wind from wind direction 40. Drag 43 is a force generated in a direction horizontal to wind direction 41 relative to blade 13. Drag 43 is a factor in causing blade 13 to stall when moving in blade movement direction 44. For this reason, it is desirable for drag 43 to be as small as possible so that wind turbine generator 10 can generate power efficiently.

[0028] The pitch angle is the angle between the plane of rotation 45 and the blade center line 46. The pitch angle is indicated by θ in FIG. 2. The blade center line 46 is a straight line connecting the leading edge and trailing edge of the airfoil of the blade 13. The smaller the pitch angle, the greater the amount of wind received from the wind direction 40. At this time, the generated lift force 42 becomes larger, but the generated drag force 43 also becomes larger accordingly.

[0029] The pitch angle control device 15 controls the pitch angle so that the lift 42 is as large as possible and the drag 43 is as small as possible. This allows the wind power generator 10 to generate power more efficiently. On the other hand, if the wind blowing from the wind direction 40 is strong, the pitch angle control device 15 makes the blades 13 parallel to the wind direction 40. In other words, the pitch angle control device 15 sets the pitch angle to 90°. This allows the pitch angle control device 15 to prevent damage to the blades 13.

[0030] <Configuration of second motor generator 100> As shown in FIG. 3, the second motor generator 100 includes a rotor 101, a stator 102, a second rotating shaft 104, a bearing 105, an oil pump 108, and a housing 109.

[0031] A magnet is installed in the rotor 101. The magnet may be embedded in the rotor 101 or may be fixed onto the outer circumferential surface of the rotor 101. The magnet temperature estimation device 50 estimates the temperature of the magnet.

[0032] The rotor 101 is installed on the second rotating shaft 104. As described above, the second rotating shaft 104 is connected to the speed-increasing gear 28. The second rotating shaft 104 is supported by a bearing 105 and is rotatable. The second rotating shaft 104 rotates by receiving rotation from the speed-increasing gear 28. At this time, the rotor 101 also rotates together with the second rotating shaft 104.

[0033] The stator 102 is disposed so as to surround the rotor 101. A stator coil is disposed within the stator 102. The stator coil generates electricity by electromagnetic induction accompanying the rotation of the rotor 101.

[0034] Oil having lubricating and cooling properties flows inside the housing 109, i.e., inside the second motor generator 100. After passing through the inside of the second motor generator 100, the oil accumulates at the bottom of the housing 109. The oil that has accumulated at the bottom of the housing 109 is sucked up by the oil pump 108 and returned to the housing 109. Note that the oil pump 108 may use a radiator or the like to cool the oil on the way back to the housing 109.

[0035] A stator coil temperature sensor 103 is installed on the stator 102, a second motor-generator revolution speed sensor 106 is installed on the second rotating shaft 104, and an oil temperature sensor 107 is installed on the bottom of the housing 109. The function of each sensor will be described later.

[0036] <Configuration of magnet temperature estimation device 50 and connections of each part in wind power generator 10> As shown in Fig. 4, the magnet temperature estimation device 50 is composed of a processing device 51 and a storage device 52. Note that, as shown in Fig. 1, the magnet temperature estimation device 50 is installed inside the nacelle 11.

[0037] The storage device 52 stores a program and an estimation model 53. The processing device 51 executes various processes by executing the program stored in the storage device 52. The processing device 51 includes a processor.

[0038] The magnet temperature estimation device 50 is, for example, a device converted from a hybrid system, and is communicatively connected to the inverter 34. As described above, the inverter 34 is connected to the second motor generator 100. The processing device 51 controls the inverter 34 and adjusts the amount of power generated by the second motor generator 100.

[0039] The pitch angle control device 15 is connected to be able to communicate with the pitch angle control mechanism 19. The pitch angle control device 15 is connected to a shaft rotation speed sensor 17 and an anemometer 36. The pitch angle control device 15 obtains information on the rotation speed of the shaft 16 from the shaft rotation speed sensor 17. The pitch angle control device 15 obtains information on the wind speed of the wind receiving the blades 13 from the anemometer 36.

[0040] To control the pitch angle, the pitch angle control device 15 uses information on the speed of the wind that the blades 13 are receiving and information on the rotation speed of the shaft 16. For example, the pitch angle control device 15 grasps the moving speed of the blades 13 from the information on the rotation speed of the shaft 16, and estimates the relative wind direction 41 based on the wind speed information. The pitch angle control device 15 determines an optimal pitch angle from the estimated relative wind direction 41. Thereafter, the pitch angle of the blades 13 is changed to the determined pitch angle via the pitch angle control mechanism 19.

[0041] An estimation model 53 stored in a storage device 52 outputs the temperature of the magnet installed in the rotor 101 when time-series data of explanatory variables is input. 4, the magnet temperature estimation device 50 is connected to a plurality of devices in order to acquire time series data of explanatory variables. Specifically, the magnet temperature estimation device 50 is communicatively connected to the anemometer 36, the pitch angle control device 15, the outside air temperature sensor 35, and the shaft rotation speed sensor 17. Furthermore, the magnet temperature estimation device 50 is communicatively connected to the drive shaft torque sensor 21, the second motor generator rotation speed sensor 106, the oil temperature sensor 107, the stator coil temperature sensor 103, and the drive shaft rotation speed sensor 22.

[0042] The explanatory variables input to the estimation model 53 include the wind speed of the wind receiving the blades 13. The temperature of the magnets installed in the rotor 101 increases according to the rotation speed of the second motor generator 100. The wind speed of the wind receiving the blades 13 is one of the factors that determine the rotation speed of the second motor generator 100.

[0043] The anemometer 36 is installed near the wind power generator 10. The anemometer 36 measures the speed of the wind that the blades 13 receive. The processing device 51 of the magnet temperature estimation device 50 obtains the speed of the wind that the blades 13 receive from the anemometer 36.

[0044] The explanatory variables input to the estimation model 53 include the pitch angle of the blades 13. The pitch angle of the blades 13 is one of the factors that determine the rotation speed of the second motor generator 100. The processing device 51 obtains the pitch angle of the blade 13 from the pitch angle control device 15 .

[0045] The explanatory variables input to the estimation model 53 include the outside temperature, which affects the temperature of the magnet. The outside air temperature sensor 35 measures the outside air temperature. The processing device 51 obtains the outside air temperature from the outside air temperature sensor 35.

[0046] The explanatory variables input to the estimation model 53 include the rotation speed of the shaft 16. The rotation speed of the shaft 16 affects the rotation speed of the second motor generator 100. The shaft rotation speed sensor 17 is installed on the shaft 16. The shaft rotation speed sensor 17 measures the rotation speed of the shaft 16. The processing device 51 obtains the rotation speed of the shaft 16 from the shaft rotation speed sensor 17.

[0047] The explanatory variables input to the estimation model 53 include explanatory variables used in the magnet temperature estimation method in a hybrid vehicle. The torque of the second motor generator 100 is included in the explanatory variables used in the magnet temperature estimation method for a hybrid vehicle.

[0048] The drive shaft torque sensor 21 is installed on the drive shaft 20. The drive shaft torque sensor 21 measures the torque of the drive shaft 20. The processing device 51 acquires the torque of the drive shaft 20 from the drive shaft torque sensor 21. The processing device 51 acquires the torque of the second motor generator 100 by calculating the torque of the second motor generator 100 from the torque of the drive shaft 20.

[0049] The rotation speed of the second motor generator 100 is included in the explanatory variables used in the magnet temperature estimation method for a hybrid vehicle. The second motor generator rotation speed sensor 106 measures the rotation speed of the second motor generator 100. The processing device 51 obtains the rotation speed of the second motor generator 100 from the second motor generator rotation speed sensor 106.

[0050] The temperature of the oil in the second motor generator 100 is included in the explanatory variables used in the magnet temperature estimation method for a hybrid vehicle. The oil temperature sensor 107 measures the temperature of the oil. The processing device 51 obtains the oil temperature from the oil temperature sensor 107.

[0051] The temperature of the stator coil provided in the stator 102 is included in the explanatory variables used in the magnet temperature estimation method in a hybrid vehicle. The stator coil temperature sensor 103 measures the temperature of the stator coil. The processing device 51 acquires the temperature of the stator coil from the stator coil temperature sensor 103.

[0052] The rotation speed of the drive shaft 20 is included in the explanatory variables used in the magnet temperature estimation method in a hybrid vehicle. The drive shaft rotation speed sensor 22 is installed on the drive shaft 20 and measures the rotation speed of the drive shaft 20. The processing device 51 obtains the rotation speed of the drive shaft 20 from the drive shaft rotation speed sensor 22.

[0053] In this way, the processing device 51 acquires data related to explanatory variables from multiple devices. At this time, the devices providing information to the processing device 51 transmit data related to explanatory variables to the processing device 51, for example, every second. The processing device 51 acquires the received data as time-series data by storing the received data in the storage device 52 for several tens of seconds. The processing device 51 inputs the acquired time-series data into the estimation model 53 and estimates the temperature of the magnet installed in the rotor 101.

[0054] <Magnet temperature estimation method using magnet temperature estimation device 50> FIG. 5 shows the structure of a convolutional neural network, which is an estimation model 53 used by the magnet temperature estimation device 50 for magnet temperature estimation.

[0055] The magnet temperature estimation device 50 first inputs the time series data of explanatory variables acquired from multiple devices to the input layer 201. At this time, it is considered that some of the time series data of explanatory variables, such as the rotation speed of the second motor generator 100, has a large amount of change in value at each point in time when viewed from a microscopic perspective every second, and the trend can only be grasped by viewing it from a macroscopic perspective. Therefore, it is desirable for the processing device 51 to apply processing using a low-pass filter and a high-pass filter in advance to the time series data to be input to the input layer 201.

[0056] Fig. 6 is a table showing the time series data input by the processing device 51 to the input layer 201. The processing device 51 performs magnet temperature estimation using the time series data for 10 seconds from time t-9 to time t for each explanatory variable as shown in the table of Fig. 6, for example.

[0057] The processing device 51 inputs the time series data shown in the table of Fig. 6 output from the input layer 201 to the convolution layer 202 and performs convolution processing. In the convolution processing, the processing device 51 uses a filter having a filter size of 3 and a number of channels of 3, as shown by a solid-line rectangle in Fig. 6, for example.

[0058] The processing device 51 calculates the sum of products between the pitch angle, wind speed, and outside air temperature data from time t-9 to time t-7, which is the overlapping portion of the table and the filter in Fig. 6, and the filter elements corresponding to each data. Next, the processing device 51 calculates the sum of products between the pitch angle, wind speed, and outside air temperature data from time t-8 to time t-6, and the filter elements represented by the dashed rectangle in Fig. 6 corresponding to each data. The processing device 51 continues to move the filter laterally in this manner until processing is performed on the pitch angle, wind speed, and outside air temperature data from time t-2 to time t.

[0059] After the process of the pitch angle, wind speed, and outside air temperature data from time t-9 to time t by the horizontal movement of the filter is completed, the filter moves vertically. That is, the filter shown in FIG. 6 moves vertically, and the same process is performed on the data of the wind speed, outside air temperature, and the number of revolutions of the shaft 16 from time t-9 to time t. In this way, the filter moves and processes the time series data of all explanatory variables from time t-9 to time t. Note that, although the stride of the filter is 1 in this embodiment, it is also possible to set the stride to any value other than 1.

[0060] In this way, each value of the feature map obtained by performing the process of calculating the sum of products while the filter moves through the table in Fig. 6 is input to the activation function 203. Note that it is preferable to prepare multiple types of filters for use in the convolution process in order to improve the output accuracy of the estimation model 53.

[0061] Each value of the feature map output from the convolution layer 202 is weighted, added up, and then multiplied by an activation function 203 such as a sigmoid function. The values ​​thus obtained are input to each node of the fully connected layer 204. In this case, it is also possible to provide a pooling layer before the fully connected layer 204.

[0062] The output from each node in the fully connected layer 204 is input to a node in the output layer 205, which outputs an estimate of the current temperature of the magnet. Estimation model 53 learns weights for estimating the magnet temperature through supervised learning using the time-series data and the actual magnet temperature input to input layer 201. Storage device 52 stores trained estimation model 53 thus trained to be able to estimate the temperature of the magnet provided in rotor 101.

[0063] <Action of this embodiment> The temperature of the magnets installed in the rotor of the generator rises according to the rotation speed of the generator. The wind speed and the pitch angle of the blades 13 are among the factors that determine the rotation speed of the generator in the wind power generator 10. The magnet temperature estimation device 50 inputs time-series data including the wind speed and the pitch angle of the blades 13 as explanatory variables into an estimation model 53 to estimate the magnet temperature.

[0064] <Effects of this embodiment> (1) The magnet temperature estimation device 50 can estimate the temperature of the magnets installed in the rotor of the generator of the wind power generator 10.

[0065] (2) The magnet temperature estimation device 50 further includes the outside air temperature as an explanatory variable. The temperature of the magnet installed in the rotor of the generator is affected by the outside air temperature, which is the air temperature outside the generator. The magnet temperature estimation device 50 inputs time-series data including the outside air temperature as an explanatory variable into the estimation model 53. This enables the magnet temperature estimation device 50 to more accurately estimate the temperature of the magnet installed in the rotor of the generator.

[0066] (3) The magnet temperature estimation device 50 further includes the rotation speed of the shaft 16 as an explanatory variable. The rotation speed of the shaft 16 affects the rotation speed of the generator. The magnet temperature estimation device 50 inputs time series data including the rotation speed of the shaft 16 as an explanatory variable to the estimation model 53. This enables the magnet temperature estimation device 50 to more accurately estimate the temperature of the magnet installed in the rotor of the generator.

[0067] (4) The wind power generator 10 is a hybrid system for a hybrid vehicle having a drive shaft 20, a transaxle, and a second motor generator 100. In the wind power generator 10, the rotation of the shaft 16 is transmitted in the order of the drive shaft 20, the transaxle, and the second motor generator 100. In the wind power generator 10, the rotation transmitted to the second motor generator 100 rotates the rotor 101 inside the second motor generator 100 to generate electricity, so that the second motor generator 100 functions as a generator. In the magnet temperature estimation device 50, the processing device 51 estimates the temperature of the magnet installed in the rotor 101 by acquiring time series data of the explanatory variables and inputting it to the estimation model 53. The magnet temperature estimation device 50 estimates the temperature of the magnet using the estimation model 53 in the wind power generator 10 that is a hybrid system. In this way, the magnet temperature estimation device 50 can estimate the temperature of the magnet installed in the rotor 101.

[0068] (5) The magnet temperature estimation device 50 further includes, as explanatory variables to be input to the estimation model 53, the torque of the second motor generator 100, the rotation speed of the second motor generator 100, and the oil temperature in the second motor generator 100. In addition, the magnet temperature estimation device 50 further includes, as explanatory variables to be input to the estimation model 53, the temperature of the stator coil of the second motor generator 100, and the rotation speed of the drive shaft 20. In the hybrid system of the hybrid vehicle, the torque of the second motor generator 100, the rotation speed of the second motor generator 100, and the oil temperature in the second motor generator 100 are used to estimate the temperature of the magnet installed in the rotor 101. In addition, in the hybrid system of the hybrid vehicle, the temperature of the stator coil of the second motor generator 100 and the rotation speed of the drive shaft 20 are used to estimate the temperature of the magnet installed in the rotor 101. In the wind power generator 10 that is a converted hybrid system, the magnet temperature estimation device 50 inputs time series data including the same explanatory variables as those in magnet temperature estimation in a hybrid vehicle to the estimation model 53. This enables the magnet temperature estimation device 50 to estimate the temperature of the magnet installed in the rotor 101 more accurately.

[0069] <Example of change> This embodiment can be modified as follows: This embodiment and the following modifications can be combined with each other to the extent that there is no technical contradiction.

[0070] In this embodiment, the processing device 51 estimates the magnet temperature using a convolutional neural network. Alternatively, the processing device 51 may estimate the magnet temperature using a method that uses other neural networks that can handle time-series data, such as LSTM or Transformer.

[0071] In this embodiment, the wind power generator 10 is diverted from the hybrid system, including the first motor generator 33. However, as described above, the first motor generator 33 does not generate electricity, and is therefore not an essential component of the present invention. In this case, it can be said that the part of the transaxle connected to the first motor generator 33 is also not an essential component. In other words, in the transaxle diverted by the wind power generator 10, the second driven gear 27 and the planetary gear mechanism 32 are not essential components, just like the first motor generator 33. The wind power generator 10 may be configured not to include these non-essential components.

[0072] The explanatory variables do not have to include all of the pitch angle, wind speed, outside air temperature, rotation speed of the shaft 16, torque of the second motor generator 100, rotation speed of the second motor generator 100, oil temperature, temperature of the stator coil, and rotation speed of the drive shaft 20. The magnet temperature estimation device 50 only needs to be equipped with an estimation model 53 that estimates the magnet temperature by inputting time-series data that includes at least the wind speed of the wind received by the blades 13 and the pitch angle of the blades 13 as explanatory variables.

[0073] The wind power generator 10 does not have to be a hybrid system. In other words, the wind power generator 10 may be any generator. The magnet temperature estimation device 50 estimates the temperature of the magnets provided in the rotor of the generator. [Explanation of symbols]

[0074] 10...wind turbine generator, 11...nacelle, 12...yaw control motor, 13...blade, 14...hub, 15...pitch angle control device, 16...shaft, 17...shaft rotation speed sensor, 18...gearbox, 19...pitch angle control mechanism, 20...drive shaft, 21...drive shaft torque sensor, 22...drive shaft rotation speed sensor, 23...differential gear, 24...first drive gear, 25...first driven gear, 26...second drive gear, 27...second driven gear, 28...speed increasing gear, 29...ring gear, 30...planet carrier, 31...sun gear, 32...planetary gear mechanism, 33...first motor generator, 34...inverter, 35...outside air temperature sensor , 36... anemometer, 37... first rotating shaft, 40... wind direction, 41... relative wind direction, 42... lift, 43... drag, 44... blade movement direction, 45... rotation surface, 46... blade centerline, 50... magnet temperature estimation device, 51... processing device, 52... storage device, 53... estimation model, 100... second motor generator, 101... rotor, 102... stator, 103... stator coil temperature sensor, 104... second rotating shaft, 105... bearing, 106... second motor generator rotation speed sensor, 107... oil temperature sensor, 108... oil pump, 109... housing, 201... input layer, 202... convolution layer, 203... activation function, 204... fully connected layer, 205... output layer

Claims

1. A magnet temperature estimation device for a wind power generator in which the shaft rotates due to the wind received by the blades and the rotation is transmitted to the generator, A processing device and a storage device, an estimation model trained by supervised learning to output a temperature of a magnet installed in a rotor of the generator when time-series data including a wind speed of the wind receiving the blades and a pitch angle of the blades as explanatory variables is input, and is stored in the storage device; The processing device comprises: The time series data is acquired and input to the estimation model to estimate the temperature of the magnet. Magnet temperature estimation device.

2. The explanatory variables further include outside temperature. The magnet temperature estimation device according to claim 1 .

3. The explanatory variables further include the rotation speed of the shaft. The magnet temperature estimation device according to claim 1 or 2.

4. The wind power generator is a hybrid system for a hybrid vehicle that includes a drive shaft, a transaxle, and a motor generator, The rotation of the shaft is transmitted in the order of the drive shaft, the transaxle, and the motor generator to rotate a rotor inside the motor generator to generate electricity, so that the motor generator functions as the generator; The processing device comprises: The temperature of a magnet installed in the rotor of the motor generator is estimated by acquiring time series data of the explanatory variables and inputting the data to the estimation model. The magnet temperature estimation device according to claim 1 or 2.

5. the explanatory variables to be input to the estimation model further include a torque of the motor generator, a rotation speed of the motor generator, an oil temperature in the motor generator, a temperature of a stator coil of the motor generator, and a rotation speed of the drive shaft; The magnet temperature estimation device according to claim 4 .

Images