JP7059557B2 - Wind turbine control program, wind turbine control method, and wind turbine control device - Google Patents

Description

The present invention relates to a wind turbine control program, a wind turbine control method, and a wind turbine control device.

Conventionally, there is a wind power generation system as a power generation system using natural energy. The wind power generation system controls the rotation speed of the wind turbine and the amount of power generated by the generator by controlling the pitch of the wind turbine or the load torque of the generator. Further, the wind power generation system has a brake for stopping the rotation of the wind turbine to prevent the wind turbine from being damaged so that the rotation speed of the wind turbine does not reach the upper limit of the rotation speed at which the wind turbine may be damaged.

As the prior art, for example, there is a technique of shifting from the torque command pattern to the protection torque command pattern when the rotation speed of the generator rises and the torque command value rises to the protection detection level or higher. Further, for example, every time the operating point of the generator is searched, it is determined whether or not it is a stall region determined by the relationship between the output power of the generator and the time differential of the output power, and it is determined that the region is a stall region. There is a wind power generation system that releases or reduces the load on the generator.

Japanese Unexamined Patent Publication No. 2006-257985 Japanese Unexamined Patent Publication No. 2011-60290

However, with the prior art, it is difficult to properly control the wind power generation system in response to changes in the environment. For example, it is conceivable to control the pitch of the wind turbine by reinforcement learning, but it is not possible to respond to changes in the wind speed between the time the pitch of the wind turbine is changed and the next time the pitch of the wind turbine is changed again. The brake may stop the rotation of the wind turbine and stop the power generation by the generator.

In one aspect, the present invention aims to perform appropriate load control in accordance with changes in the environment.

According to one embodiment, the load torque of the generator using the wind turbine is controlled based on the wind speed and the rotation speed of the wind turbine, and the wind speed and the rotation speed of the wind turbine are used as observed values to generate the power generation. Whether to move the operating point on the characteristic curve showing the relationship between the rotation speed of the wind turbine and the wind turbine torque of the wind turbine to the region on the side where the rotation speed of the wind turbine is smaller than the maximum point, using the amount of power generated by the machine as a reward. When the agent is made to execute the reinforcement learning with the action of whether or not, and the agent to which the wind speed and the rotation speed of the wind turbine are input outputs to move the operating point to the region, the generator of the generator. A wind turbine control program, a wind turbine control method, and a wind turbine control device for moving the operating point to the region by controlling the load torque are proposed.

According to one aspect, it becomes possible to perform appropriate load control according to changes in the environment.

FIG. 1 is an explanatory diagram showing an embodiment of a wind turbine control method according to an embodiment. FIG. 2 is an explanatory diagram showing an example of the wind power generation system 101. FIG. 3 is a block diagram showing a hardware configuration example of the wind turbine control device 100. FIG. 4 is an explanatory diagram showing an example of the stored contents of the latest measured value information 400. FIG. 5 is an explanatory diagram showing an example of the stored contents of the wind turbine torque characteristic information 500. FIG. 6 is an explanatory diagram showing an example of the stored contents of the ratio information 600. FIG. 7 is an explanatory diagram showing an example of the stored contents of the movement flag information 700. FIG. 8 is an explanatory diagram showing an example of the stored contents of the observation history 800. FIG. 9 is an explanatory diagram showing an example of the stored contents of the action value table 900. FIG. 10 is an explanatory diagram showing an example of the stored contents of the action history 1000. FIG. 11 is a block diagram showing a functional configuration example of the wind turbine control device 100. FIG. 12 is a block diagram showing a specific functional configuration example of the wind power generation system 101. FIG. 13 is an explanatory diagram (No. 1) showing a control index of the load torque of the generator 120 based on the wind turbine torque characteristic. FIG. 14 is an explanatory diagram (No. 2) showing a control index of the load torque of the generator 120 based on the wind turbine torque characteristic. FIG. 15 is an explanatory diagram (No. 3) showing a control index of the load torque of the generator 120 based on the wind turbine torque characteristic. FIG. 16 is an explanatory diagram (No. 4) showing a control index of the load torque of the generator 120 based on the wind turbine torque characteristic. FIG. 17 is an explanatory diagram (No. 5) showing a control index of the load torque of the generator 120 based on the wind turbine torque characteristic. FIG. 18 is an explanatory diagram (No. 6) showing a control index of the load torque of the generator 120 based on the wind turbine torque characteristic. FIG. 19 is an explanatory diagram (No. 7) showing a control index of the load torque of the generator 120 based on the wind turbine torque characteristic. FIG. 20 is an explanatory diagram showing a flow for controlling the load torque of the generator 120. FIG. 21 is an explanatory diagram (No. 1) showing an operation example of controlling the load torque of the generator 120. FIG. 22 is an explanatory diagram (No. 2) showing an operation example of controlling the load torque of the generator 120. FIG. 23 is an explanatory diagram (No. 3) showing an operation example of controlling the load torque of the generator 120. FIG. 24 is an explanatory diagram (No. 4) showing an operation example of controlling the load torque of the generator 120. FIG. 25 is an explanatory diagram (No. 5) showing an operation example of controlling the load torque of the generator 120. FIG. 26 is an explanatory diagram (No. 6) showing an operation example of controlling the load torque of the generator 120. FIG. 27 is an explanatory diagram (No. 7) showing an operation example of controlling the load torque of the generator 120. FIG. 28 is an explanatory diagram (No. 8) showing an operation example of controlling the load torque of the generator 120. FIG. 29 is an explanatory diagram (No. 9) showing an operation example of controlling the load torque of the generator 120. FIG. 30 is an explanatory diagram (No. 10) showing an operation example of controlling the load torque of the generator 120. FIG. 31 is a flowchart showing an example of the overall processing procedure. FIG. 32 is a flowchart showing an example of the operation / stop processing procedure. FIG. 33 is a flowchart showing an example of the reinforcement learning processing procedure. FIG. 34 is a flowchart showing an example of the load torque control processing procedure. FIG. 35 is a flowchart showing an example of the left side control processing procedure. FIG. 36 is a flowchart showing an example of the right side control processing procedure.

Hereinafter, embodiments of a wind turbine control program, a wind turbine control method, and a wind turbine control device according to the present invention will be described in detail with reference to the drawings.

(An example of a wind turbine control method according to an embodiment)
FIG. 1 is an explanatory diagram showing an embodiment of a wind turbine control method according to an embodiment. In FIG. 1, the wind turbine control device 100 is a computer that controls a wind power generation system 101.

The wind power generation system 101 has a wind turbine 110 and a generator 120. The wind turbine 110 receives the wind, converts the wind power into the wind turbine torque, and transmits the wind power to the shaft of the generator 120. The wind speed can fluctuate. Wind is lost when converted to wind turbine torque. The wind turbine 110 has a brake.

The generator 120 uses the wind turbine 110 to generate electricity. The generator 120 generates electricity by using, for example, the wind turbine torque transmitted from the wind turbine 110 to the shaft. The generator 120 can apply a load torque in the opposite direction of the wind turbine torque transmitted to the shaft. The load torque can be generated, for example, by making the generator 120 also function as an electric motor. The load torque has an upper limit for the load torque. When the energy supplied to the generator 120 is surplus, the rotation speed of the wind turbine 110 increases. When the energy supplied to the generator 120 is insufficient for the energy consumed by the generator 120, the rotation speed of the wind turbine 110 decreases.

Here, the wind power generation system 101 has a property that when the environment such as the wind speed changes, the rotation speed and the wind turbine torque of the wind turbine 110 that can maximize the power generation efficiency of the generator 120 also change. Therefore, it is desired to appropriately control the wind power generation system 101 in response to changes in the environment such as wind speed.

However, it is difficult to appropriately control the wind power generation system 101 in accordance with changes in the environment such as wind speed. For example, the first technique, the second technique, the third technique, the fourth technique, and the like shown below can be considered, but they are appropriate for the wind power generation system 101 according to changes in the environment such as wind speed. It is difficult to perform such control.

As the first technique, for example, a technique of controlling the wind turbine torque of the wind turbine 110 and the rotation speed of the wind turbine 110 by controlling the pitch of the wind turbine 110 can be considered. The pitch is an element that controls the magnitude of the force that the wind turbine 110 receives from the wind. The pitch is, for example, the angle of the blades of the wind turbine 110. Pitch control is performed using, for example, reinforcement learning.

However, if a structure for controlling the pitch is provided, the manufacturing cost of the wind power generation system 101 will increase. Further, in the small wind power generation system 101, the smaller the size of the wind turbine 110, the shorter the response time in which the rotation speed of the wind turbine 110 changes according to the change in the wind speed. Also, when controlling the pitch using reinforcement learning, the interval for changing the pitch tends to be long. Therefore, between the time when the pitch is changed and the time when the pitch is changed again, it is not possible to respond to the change in the rotation speed of the wind turbine 110 according to the change in the wind speed, and the power generation efficiency of the generator 120 is lowered. Can be invited.

As a second technique, for example, a technique of controlling the wind turbine torque of the wind turbine 110 and the rotation speed of the wind turbine 110 by controlling the load torque of the generator 120 can be considered. Specifically, by controlling the load torque of the generator 120, the operating point of the wind turbine 110 on the characteristic curve showing the relationship between the wind turbine torque of the wind turbine 110 and the rotation speed of the wind turbine 110 is set to the power generation efficiency of the generator 120. A technique for moving to the operating point that maximizes the torque can be considered.

The operating point that maximizes the power generation efficiency of the generator 120 is in the region on the characteristic curve where the rotation speed is higher than the maximum point of the wind turbine torque. In the following description, the region on the characteristic curve where the rotation speed is higher than the maximum point of the wind turbine torque may be referred to as "the right side of the mountain". Further, in the following description, the region on the characteristic curve where the rotation speed is smaller than the maximum point of the wind turbine torque may be referred to as "the left side of the mountain".

However, on the right side of the mountain of the characteristic curve, there is a property that the rotation speed increases as the wind speed increases, and the rotation speed tends to reach the upper limit of the rotation speed. Further, when the wind speed increases, it may be difficult to reduce the rotation speed even if the load torque is increased to the upper limit of the load torque. Therefore, in order to prevent the wind turbine 110 from being damaged, the rotation of the wind turbine 110 is stopped by the brake, the power generation of the generator 120 is also stopped, and the long-term power generation efficiency of the generator 120 may be lowered.

As a third technique, for example, by controlling the load torque of the generator 120 so that the rotation of the wind turbine 110 is not stopped by the brake when the wind speed exceeds a certain level, the wind turbine 110 on the characteristic curve can be controlled. A technique to move the operating point to the left side of the peak of the characteristic curve can be considered.

However, even if the wind speed exceeds a certain level, the rotation of the wind turbine 110 is not always stopped by the brake. Therefore, if the operating point of the wind turbine 110 is set to the right side of the mountain of the characteristic curve, the operating point of the wind turbine 110 is moved to the left side of the mountain of the characteristic curve even if the long-term power generation efficiency of the generator 120 is good. This can lead to a long-term decrease in power generation efficiency of the generator 120.

As a fourth technique, for example, a technique of controlling the wind turbine torque of the wind turbine 110 and the rotation speed of the wind turbine 110 by controlling the load torque of the generator 120 by using reinforcement learning can be considered.

However, as in the first technique, when the load torque of the generator 120 is controlled by using reinforcement learning, the interval for changing the load torque of the generator 120 tends to be long. Therefore, it is not possible to respond to the change in the rotation speed of the wind turbine 110 according to the change in the wind speed between the time when the load torque of the generator 120 is changed and the time when the load torque of the generator 120 is changed again. ..

As a result, when the operating point of the wind turbine 110 is on the left side of the mountain of the characteristic curve, the rotation of the wind turbine 110 may stop when the wind speed suddenly decreases. Further, in this case, if the wind speed suddenly increases, the operating point of the wind turbine 110 may unintentionally move to the right side of the mountain of the characteristic curve. Further, when the operating point of the wind turbine 110 is on the right side of the mountain of the characteristic curve, if the wind speed suddenly increases, the rotation speed may reach the upper limit of the rotation speed.

Therefore, in the present embodiment, a wind turbine control method for determining the timing of moving the operating point of the wind turbine 110 to the left side of the mountain of the characteristic curve by the agent performing reinforcement learning while controlling the load torque of the generator 120 will be described. .. Thereby, the wind turbine control method controls the load torque of the generator 120 according to the change of the environment.

(1-1) The wind turbine control device 100 controls the load torque of the generator 120 based on the wind speed and the rotation speed of the wind turbine 110. As a result, the wind turbine control device 100 can suppress a decrease in the speed of controlling the load torque of the generator 120 in response to a change in the wind speed regardless of the execution interval of the action determination of the agent, and the operating point of the wind turbine 110. Can be prevented from moving unintentionally.

(1-2) The wind turbine control device 100 uses the wind speed and the rotation speed of the wind turbine 110 as observation values, the amount of power generated by the generator 120 as a reward, and the operating point of the wind turbine 110 on the characteristic curve as a mountain of the characteristic curve. Let the agent perform reinforcement learning with the action of whether or not to move it to the left side of. Thereby, the wind turbine control device 100 can make the agent learn a control model for determining the behavior for improving the long-term power generation efficiency of the generator 120 in consideration of the environment such as the wind speed.

The control model is, for example, a model that can output an action when an observed value is input. The control model is, for example, a table in which the condition for the observed value is associated with what kind of action is output when the condition for the observed value is satisfied. The control model may be, for example, a mathematical model. The control model may be, for example, a decision tree model.

(1-3) When the agent to which the wind speed and the rotation speed of the wind turbine 110 are input outputs an output to move the operating point to the left side of the mountain of the characteristic curve, the wind turbine control device 100 loads the generator 120. By controlling the torque, the operating point is moved to the left side of the peak of the characteristic curve. Thereby, the wind turbine control device 100 can move the operating point of the wind turbine 110 on the characteristic curve to the left side of the mountain of the characteristic curve when the long-term power generation efficiency of the generator 120 can be improved.

In this way, the wind turbine control device 100 can appropriately control the wind power generation system 101 in accordance with changes in the environment such as wind speed, which causes a long-term decrease in power generation efficiency of the generator 120. Can be avoided.

The wind turbine control device 100 can suppress a decrease in the speed for controlling the load torque of the generator 120 in response to a change in the wind speed, and prevent the operating point of the wind turbine 110 from unintentionally moving. be able to. Specifically, when the operating point of the wind turbine 110 is on the right side of the mountain of the characteristic curve, the wind turbine control device 100 prevents the rotation speed from reaching the upper limit of the rotation speed even if the wind speed suddenly increases. can do. Therefore, the wind turbine control device 100 can prevent a decrease in the power generation efficiency of the generator 120.

On the other hand, in the wind turbine control device 100, specifically, when the operating point of the wind turbine 110 is on the left side of the mountain of the characteristic curve, the rotation of the wind turbine 110 stops even if the wind speed suddenly decreases. Can be prevented. Further, in the wind turbine control device 100, specifically, when the operating point of the wind turbine 110 is on the left side of the mountain of the characteristic curve, the operating point of the wind turbine 110 is unintentionally characteristic even if the wind speed suddenly increases. It is possible to prevent the vehicle from moving to the right side of the mountain of the curve. Therefore, the wind turbine control device 100 can prevent a decrease in the power generation efficiency of the generator 120.

Further, the wind turbine control device 100 maximizes the power generation efficiency of the generator 120 at the operating point of the wind turbine 110, for example, in an environment where the rotation of the wind turbine 110 is not stopped by the brake even if the wind speed is above a certain level. It can be moved to the operating point to be changed. Therefore, the wind turbine control device 100 can improve the long-term power generation efficiency of the generator 120.

On the other hand, the wind turbine control device 100 can move the operating point of the wind turbine 110 to the left side of the mountain of the characteristic curve, for example, in an environment where the rotation of the wind turbine 110 can be stopped by the brake. Therefore, the wind turbine control device 100 can prevent the rotation of the wind turbine 110 from being stopped by the brake and the power generation of the generator 120 from being stopped, thereby improving the long-term power generation efficiency of the generator 120. be able to.

Further, the wind turbine control device 100 has a structure for controlling the pitch of the wind turbine 110 because, for example, the wind turbine torque of the wind turbine 110 and the rotation speed of the wind turbine 110 can be controlled without controlling the pitch of the wind turbine 110. It can also be applied to the wind power generation system 101 that does not. Further, the wind turbine control device 100 does not need to be provided with a structure for controlling the pitch of the wind turbine 110 when manufacturing the wind power generation system 101, thereby suppressing an increase in the manufacturing cost of the wind power generation system 101. Can be done.

(Example of wind power generation system 101)
Next, an example of the wind power generation system 101 to which the wind turbine control device 100 shown in FIG. 1 is applied will be described with reference to FIG.

FIG. 2 is an explanatory diagram showing an example of the wind power generation system 101. In FIG. 2, the wind power generation system 101 includes a wind turbine control device 100, a wind turbine 110, and a generator 120.

In the wind power generation system 101, the wind turbine control device 100, the wind turbine 110, and the generator 120 are directly connected. The wind turbine control device 100, the wind turbine 110, and the generator 120 may be connected via a wired or wireless network. The network is, for example, a LAN (Local Area Network), a WAN (Wide Area Network), the Internet, or the like.

The wind turbine 110 is an object that receives wind, converts the wind power into wind turbine torque, and transmits the wind power to the shaft of the generator 120. The wind turbine 110 is provided with a brake for stopping the rotation of the wind turbine 110. The wind turbine 110 is provided with a measuring device for measuring the wind speed in the vicinity of the wind turbine 110, the wind turbine torque of the wind turbine 110, the rotation speed of the wind turbine 110, and the like. The measuring instrument provided in the wind turbine 110 has, for example, a sensor device. The sensor device may include at least one of an acceleration sensor, a geomagnetic sensor, an optical sensor, a vibration sensor, and the like.

The generator 120 is a machine that generates electricity using the wind turbine 110. The generator 120 generates electricity by using, for example, the wind turbine torque transmitted from the wind turbine 110 to the shaft. The generator 120 can apply a load torque in the opposite direction of the wind turbine torque transmitted to the shaft. The generator 120 is provided with a measuring device for measuring the load torque of the generator 120, the cumulative power generation amount of the generator 120, and the like. The measuring instrument provided in the generator 120 has, for example, a sensor device. The sensor device may include at least one of an acceleration sensor, a geomagnetic sensor, an optical sensor, a vibration sensor, a power sensor, a voltage sensor, a current sensor, and the like.

The wind turbine control device 100 controls the wind power generation system 101. The wind turbine control device 100 acquires, for example, measured values such as a wind speed in the vicinity of the wind turbine 110, a wind turbine torque of the wind turbine 110, and a rotation speed of the wind turbine 110 from a measuring device provided in the wind turbine 110. The wind turbine control device 100 acquires, for example, measured values such as the load torque of the generator 120 and the cumulative power generation amount of the generator 120 from the measuring device provided in the generator 120.

The wind turbine control device 100 controls the rotation speed of the wind turbine 110, for example, by controlling the load torque of the generator 120. The wind turbine control device 100 determines, for example, whether or not to move the operating point of the wind turbine 110 to the left side of the mountain of the characteristic curve. The wind turbine control device 100 controls, for example, a brake provided on the wind turbine 110 to control the rotation of the wind turbine 110. The wind turbine control device 100 is, for example, a server, a PC (Personal Computer), a microcomputer, a PLC (Programmable Logic Controller), or the like.

Here, a case where the wind turbine control device 100 controls the load torque of the generator 120 and determines whether or not to move the operating point of the wind turbine 110 to the left side of the mountain of the characteristic curve has been described. Not exclusively. For example, a device that controls the load torque of the generator 120 and a device that determines whether or not to move the operating point of the wind turbine 110 to the left side of the mountain of the characteristic curve are different devices, and each device cooperates. It may work.

Here, the case where the measuring instrument provided in the wind turbine 110 and the measuring instrument provided in the generator 120 are different devices has been described, but the present invention is not limited to this. For example, the measuring instrument provided in the wind turbine 110 and the measuring instrument provided in the generator 120 may be the same device.

(Hardware configuration example of wind turbine control device 100)
Next, a hardware configuration example of the wind turbine control device 100 will be described with reference to FIG.

FIG. 3 is a block diagram showing a hardware configuration example of the wind turbine control device 100. In FIG. 3, the wind turbine control device 100 includes a CPU (Central Processing Unit) 301, a memory 302, a communication I / F (Interface) 303, a recording medium I / F 304, and a recording medium 305. Further, each component is connected by a bus 300.

Here, the CPU 301 controls the entire wind turbine control device 100. The memory 302 includes, for example, a ROM (Read Only Memory), a RAM (Random Access Memory), a flash ROM, and the like. Specifically, for example, a flash ROM or ROM stores various programs, and RAM is used as a work area of CPU 301. The program stored in the memory 302 is loaded into the CPU 301 to cause the CPU 301 to execute the coded process. The memory 302 may store various tables 400 to 1000 described later in FIGS. 4 to 10.

The communication I / F 303 is connected to the network through the communication line, and is connected to another computer via the network. The communication I / F 303 controls the network and the internal interface, and controls the input / output of data from another computer. For the communication I / F 303, for example, a modem, a LAN adapter, or the like can be adopted.

The recording medium I / F 304 controls read / write of data to the recording medium 305 according to the control of the CPU 301. The recording medium I / F 304 is, for example, a disk drive, an SSD (Solid State Drive), a USB (Universal Serial Bus) port, or the like. The recording medium 305 is a non-volatile memory that stores data written under the control of the recording medium I / F 304. The recording medium 305 is, for example, a disk, a semiconductor memory, a USB memory, or the like. The recording medium 305 may be detachable from the wind turbine control device 100. The recording medium 305 may store various tables 400 to 1000, which will be described later in FIGS. 4 to 10, instead of the memory 302.

The wind turbine control device 100 may include, for example, a keyboard, a mouse, a display, a printer, a scanner, a microphone, a speaker, and the like, in addition to the above-mentioned components. Further, the wind turbine control device 100 may have a plurality of recording media I / F 304 and recording media 305. Further, the wind turbine control device 100 does not have to have the recording medium I / F 304 or the recording medium 305.

(Memory content of the latest measured value information 400)
Next, the stored contents of the latest measured value information 400 will be described with reference to FIG. The latest measurement value information 400 is realized by, for example, a storage area such as a memory 302 or a recording medium 305 of the wind turbine control device 100 shown in FIG.

FIG. 4 is an explanatory diagram showing an example of the stored contents of the latest measured value information 400. As shown in FIG. 4, the latest measured value information 400 has fields of measurement time, wind speed, rotation speed, and power generation amount. The latest measurement value information 400 is stored as a record by setting information in each field in association with the measurement time.

In the field of the measurement time, the wind speed of the wind turbine 110, the rotation speed of the wind turbine 110, and the measurement time in which the power generation amount of the generator 120 is measured are set. The wind speed of the wind turbine 110, the rotation speed of the wind turbine 110, and the sampling interval of the power generation amount of the generator 120 may be different. In this case, in the field of the measurement time, the wind speed of the wind turbine 110, the rotation speed of the wind turbine 110, and the acquisition time at which the power generation amount of the generator 120 is acquired may be set instead of the measurement time.

In the wind speed field, the measured value of the wind speed of the wind turbine 110 measured at the measurement time is set. In the rotation speed field, the measured value of the rotation speed of the wind turbine 110 measured at the measurement time is set. In the field of power generation amount, the measured value of the power generation amount of the generator 120 measured at the measurement time is set. The amount of power generation is the amount of power generation from the previous measurement time to the latest measurement time.

(Memory content of wind turbine torque characteristic information 500)
Next, the stored contents of the wind turbine torque characteristic information 500 will be described with reference to FIG. The wind turbine torque characteristic information 500 is realized, for example, by a storage area such as a memory 302 or a recording medium 305 of the wind turbine control device 100 shown in FIG.

FIG. 5 is an explanatory diagram showing an example of the stored contents of the wind turbine torque characteristic information 500. As shown in FIG. 5, the wind turbine torque characteristic information 500 has a field of rotation speed and wind turbine torque. The wind turbine torque characteristic information 500 is stored as a record by setting information in each field for each rotation speed.

The wind turbine torque characteristic information 500 shows a characteristic curve of the wind turbine torque at a reference wind speed. The reference wind speed is, for example, 5 m / s. The rotation speed of the wind turbine 110 is set in the rotation speed field. In the field of the wind turbine torque, the wind turbine torque corresponding to the rotation speed of the wind turbine 110 is set on the characteristic curve of the wind turbine torque at the reference wind speed.

The characteristic curve of the wind turbine torque at a wind speed different from the reference wind speed can be specified from the characteristic curve of the wind turbine torque at the reference wind speed. Therefore, the wind turbine torque characteristic information 500 does not have to show the characteristic curve of the wind turbine torque at a wind speed different from the reference wind speed.

(Memory content of ratio information 600)
Next, the stored contents of the ratio information 600 will be described with reference to FIG. The ratio information 600 is realized, for example, by a storage area such as a memory 302 or a recording medium 305 of the wind turbine control device 100 shown in FIG.

FIG. 6 is an explanatory diagram showing an example of the stored contents of the ratio information 600. As shown in FIG. 6, the ratio information 600 has a field of the ratio of the maximum generated torque to the maximum torque. The ratio information 600 is stored as a record by setting information in the field.

In the field of the ratio of the maximum power generation torque to the maximum torque, the ratio of the maximum power generation torque, which is the wind turbine torque of the wind turbine 110 that maximizes the power generation efficiency of the generator 120, to the maximum torque, which is the maximum value of the wind turbine torque of the wind turbine 110, is Set.

(Memory content of movement flag information 700)
Next, the stored contents of the movement flag information 700 will be described with reference to FIG. 7. The movement flag information 700 is realized, for example, by a storage area such as a memory 302 or a recording medium 305 of the wind turbine control device 100 shown in FIG.

FIG. 7 is an explanatory diagram showing an example of the stored contents of the movement flag information 700. As shown in FIG. 7, the movement flag information 700 has a field of a flag for moving to the left or not. The movement flag information 700 is stored as a record by setting information in the field.

In the field of the flag for whether or not to move to the left, a flag indicating whether or not to move the operating point of the wind turbine 110 to the left side of the mountain of the characteristic curve is set. If the flag is 1, it indicates that the operating point of the wind turbine 110 is moved to the left side of the mountain of the characteristic curve, and if it is 0, the operating point of the wind turbine 110 is not moved to the left side of the mountain of the characteristic curve. Is shown.

(Memory content of observation history 800)
Next, the stored contents of the observation history 800 will be described with reference to FIG. The observation history 800 is realized, for example, by a storage area such as a memory 302 or a recording medium 305 of the wind turbine control device 100 shown in FIG.

FIG. 8 is an explanatory diagram showing an example of the stored contents of the observation history 800. As shown in FIG. 8, the observation history 800 has fields of measurement time, wind speed, rotation speed, and power generation amount. The observation history 800 stores the history information as a record by setting information in each field for each measurement time.

In the field of the measurement time, the wind speed of the wind turbine 110, the rotation speed of the wind turbine 110, and the measurement time in which the power generation amount of the generator 120 is measured are set. The wind speed of the wind turbine 110, the rotation speed of the wind turbine 110, and the sampling interval of the power generation amount of the generator 120 may be different. In this case, in the field of the measurement time, the wind speed of the wind turbine 110, the rotation speed of the wind turbine 110, and the acquisition time at which the power generation amount of the generator 120 is acquired may be set instead of the measurement time.

In the wind speed field, the measured value of the wind speed of the wind turbine 110 measured at the measurement time is set. In the rotation speed field, the measured value of the rotation speed of the wind turbine 110 measured at the measurement time is set. In the field of power generation amount, the measured value of the power generation amount of the generator 120 measured at the measurement time is set. The amount of power generation is the amount of power generation from the previous measurement time to the latest measurement time.

(Memory content of action value table 900)
Next, the stored contents of the action value table 900 will be described with reference to FIG. The action value table 900 is realized, for example, by a storage area such as a memory 302 or a recording medium 305 of the wind turbine control device 100 shown in FIG.

FIG. 9 is an explanatory diagram showing an example of the stored contents of the action value table 900. As shown in FIG. 9, the action value table 900 has fields of one or more wind speeds, one or more rotation speeds, actions, and Q values. In the action value table 900, the action value information is stored as a record by setting information in each field for each reinforcement learning.

Conditions for wind speed are set in the wind speed field. The condition is a range. In the example of FIG. 9, a condition for the current wind speed is set in the field of wind speed 1. Further, in the field of wind speed 2, conditions for past wind speeds are set. In the rotation speed field, a condition for the rotation speed of the wind turbine 110 is set. In the example of FIG. 9, a condition for the current rotation speed of the wind turbine 110 is set in the field of the rotation speed 1. Further, in the field of the rotation speed 2, a condition for the rotation speed of the wind turbine 110 in the past is set.

In the field of action, a flag for moving to the left, which is determined as an action by the agent, is set. In the Q value field, if the current and past wind speeds and speeds meet the conditions indicated by the wind speed and speed fields, the determined action will result in an increase in the amount of power generated by the generator 120, which is a reward. A Q value is set to indicate how much it contributes to.

(Memory content of action history 1000)
Next, the stored contents of the action history 1000 will be described with reference to FIG. The action history 1000 is realized by, for example, a storage area such as a memory 302 or a recording medium 305 of the wind turbine control device 100 shown in FIG.

FIG. 10 is an explanatory diagram showing an example of the stored contents of the action history 1000. As shown in FIG. 10, the action history 1000 has a field of a measurement time and a flag of whether or not to move to the left. In the action history 1000, the history information is stored as a record by setting information in each field for each measurement time.

In the field of the measurement time, the wind speed of the wind turbine 110, the rotation speed of the wind turbine 110, and the measurement time in which the power generation amount of the generator 120 is measured are set. The wind speed of the wind turbine 110, the rotation speed of the wind turbine 110, and the sampling interval of the power generation amount of the generator 120 may be different. In this case, in the field of the measurement time, the wind speed of the wind turbine 110, the rotation speed of the wind turbine 110, and the acquisition time at which the power generation amount of the generator 120 is acquired may be set instead of the measurement time. In the field of the flag for whether or not to move to the left, whether or not to move to the left is determined by the agent based on the wind speed of the wind turbine 110 measured at the measurement time, the rotation speed of the wind turbine 110, and the amount of power generated by the generator 120. Flag is set.

(Example of functional configuration of wind turbine control device 100)
Next, a functional configuration example of the wind turbine control device 100 will be described with reference to FIG.

FIG. 11 is a block diagram showing a functional configuration example of the wind turbine control device 100. The wind turbine control device 100 includes a storage unit 1100, an acquisition unit 1101, a load torque control unit 1102, an agent 1103, and an output unit 1104.

The storage unit 1100 is realized by, for example, a storage area such as the memory 302 or the recording medium 305 shown in FIG. Hereinafter, the case where the storage unit 1100 is included in the wind turbine control device 100 will be described, but the present invention is not limited to this. For example, the storage unit 1100 may be included in a device different from the wind turbine control device 100, and the stored contents of the storage unit 1100 may be visible from the wind turbine control device 100.

The acquisition unit 1101 to the output unit 1104 are functions that serve as control units. Specifically, the acquisition unit 1101 to the output unit 1104 may cause the CPU 301 to execute a program stored in a storage area such as the memory 302 or the recording medium 305 shown in FIG. 3, or the communication I / F 303. To realize the function. The processing result of each functional unit is stored in a storage area such as the memory 302 or the recording medium 305 shown in FIG. 3, for example.

The storage unit 1100 stores various information used for processing of each functional unit. The storage unit 1100 may store, for example, the wind speed, the wind turbine torque of the wind turbine 110, the rotation speed of the wind turbine 110, the power generation amount of the generator 120, and the like. The storage unit 1100 may store, for example, information showing a characteristic curve showing the relationship between the wind turbine torque of the wind turbine 110 and the rotation speed of the wind turbine 110.

The storage unit 1100 stores, for example, a reinforcement learning algorithm and an action selection algorithm. The reinforcement learning algorithm is, for example, a Q-learning algorithm. The storage unit 1100 may store, for example, a control model learned by reinforcement learning or an action determined by the control model.

The control model is, for example, a model that can output an action when an observed value is input. The control model is, for example, a table in which the condition for the observed value is associated with what kind of action is output when the condition for the observed value is satisfied. The control model may be, for example, a mathematical model or a decision tree model. Specifically, the storage unit 1100 stores various tables 400 to 1000 shown in FIGS. 4 to 10.

The acquisition unit 1101 acquires various information used for processing of each functional unit from the storage unit 1100 and outputs the information to each functional unit. The acquisition unit 1101 may acquire, for example, the wind speed, the wind turbine torque of the wind turbine 110, the rotation speed of the wind turbine 110, the power generation amount of the generator 120, and the like.

Specifically, the acquisition unit 1101 may acquire measured values such as the wind speed in the vicinity of the wind turbine 110, the wind turbine torque of the wind turbine 110, and the rotation speed of the wind turbine 110 from the measuring instrument provided in the wind turbine 110. good. Specifically, the acquisition unit 1101 may acquire measured values such as the load torque of the generator 120 and the cumulative power generation amount of the generator 120 from the measuring device provided in the generator 120. Specifically, the acquisition unit 1101 may acquire the power generation amount of the generator 120 in a certain time from the difference in the cumulative power generation amount of the generator 120.

The load torque control unit 1102 controls the load torque of the generator 120 based on the wind speed and the rotation speed of the wind turbine 110. The wind turbine 110 does not have to have a function of controlling the pitch of the wind turbine 110. As a result, the load torque control unit 1102 can suppress a decrease in the speed of controlling the load torque of the generator 120 in response to a change in the wind speed regardless of the execution interval of the action determination of the agent 1103, and the operating point is set. It is possible to prevent unintentional movement.

The load torque control unit 1102 controls the load torque of the generator 120 based on, for example, the wind speed, the rotation speed of the wind turbine 110, and the current operating point on the characteristic curve. As a result, the load torque control unit 1102 can set the load torque according to the current operating point, and can easily maintain the current operating point.

Specifically, the current operating point on the characteristic curve may be in the region on the side where the rotation speed of the wind turbine 110 is smaller than the maximum point. The maximum point is that the load torque of the wind turbine 110 becomes maximum. The region on the side where the rotation speed of the wind turbine 110 is smaller than the maximum point is the region on the left side of the mountain. In this case, the load torque control unit 1102 alternately sets the load torque of the generator 120 to a value larger than the wind turbine torque at the operating point and a value smaller than the wind turbine torque at the operating point, thereby increasing the operating point. Maintain in the area to the left of.

Here, a value larger than the wind turbine torque at the operating point is, for example, the upper limit of the load torque. Further, a value larger than the wind turbine torque at the operating point is smaller than the load torque upper limit and close to the wind turbine torque at the operating point, and the speed at which the operating point moves as compared with the case where the load torque upper limit is set for the load torque. It may be a value that can suppress.

A value smaller than the wind turbine torque at the operating point is, for example, zero. Further, a value smaller than the wind turbine torque at the operating point is larger than zero and close to the wind turbine torque at the operating point, and the speed at which the operating point moves is suppressed as compared with the case where zero is set for the load torque. It may be a value that can be used. As a result, the load torque control unit 1102 can maintain the operating point within a certain range, suppress fluctuations in the rotation speed, and suppress the load applied to the wind turbine 110.

Further, specifically, the current operating point on the characteristic curve may be in the region on the side where the rotation speed of the wind turbine 110 is higher than the maximum point. The region on the side where the rotation speed of the wind turbine 110 is higher than the maximum point is the region on the right side of the mountain. In this case, the load torque control unit 1102 maintains the operating point in the region on the right side of the mountain by setting the load torque of the generator 120 to the maximum power generation torque that maximizes the power generation efficiency of the generator 120. Further, if the maximum power generation torque is larger than the load torque upper limit, the load torque control unit 1102 may set the load torque of the generator 120 to the load torque upper limit. As a result, the load torque control unit 1102 can maximize the power generation efficiency of the generator 120.

The load torque control unit 1102 may control the load torque of the generator 120 based on the wind speed, the rotation speed of the wind turbine 110, and the amount of power generated by the generator 120. As a result, the load torque control unit 1102 can control the operating point by controlling the load torque so that the amount of power generated by the generator 120 increases even if the ratio of the maximum power generation torque to the maximum torque is not set. ..

When the action output from the agent 1103 is the action of moving the operating point to the area on the left side of the mountain, the load torque control unit 1102 controls the load torque of the generator 120 to move the operating point to the left side of the mountain. Move to the area of. As a result, the load torque control unit 1102 can prevent the wind turbine 110 from stopping due to the brake, and can improve the long-term power generation efficiency.

The load torque control unit 1102 moves the operating point to the region on the left side of the mountain by setting the load torque of the generator 120 to a value larger than the wind turbine torque at the maximum point, for example. As a result, the load torque control unit 1102 can improve the speed at which the operating point is moved to the left side of the mountain, and can easily cope with the case where the speed at which the wind speed changes is large.

At the time of learning, the agent 1103 used the wind speed and the rotation speed of the wind turbine 110 as observation values, the amount of power generated by the generator 120 as a reward, and whether or not to move the operating point to the area on the left side of the mountain. Perform reinforcement learning. Further, the agent 1103 outputs whether or not to move the operating point to the region on the left side of the mountain by using the wind speed and the rotation speed of the wind turbine 110 as observation values at the time of the action determination operation. The agent 1103 reinforces learning whether or not to move the operating point to the area on the left side of the mountain by, for example, the Q-Learning algorithm, and outputs it as an action at the time of the action determination.

As a result, the agent 1103 can learn and update the control model that outputs the behavior when the observed value is input so as to match the nature of the environment such as the change tendency of the wind speed at the time of learning. In addition, the agent 1103 can determine an action that can maximize the amount of power generation, which is a reward, according to the actual tendency of the wind speed to change during the action determination operation. Here, the interval at which the agent 1103 determines the action may be longer than the interval at which the load torque control unit 1102 controls the load torque of the generator 120.

The output unit 1104 may output the processing result of each functional unit. The output format is, for example, display on a display, print output to a printer, transmission to an external device by communication I / F 303, or storage in a storage area such as a memory 302 or a recording medium 305. As a result, the output unit 1104 can notify the user of the processing result of each functional unit, and can support the management and operation of the wind turbine control device 100, for example, updating the set value of the wind turbine control device 100. The convenience of the wind turbine control device 100 can be improved.

(Specific functional configuration example of wind power generation system 101)
Next, a specific functional configuration example of the wind power generation system 101 will be described with reference to FIG. 12.

FIG. 12 is a block diagram showing a specific functional configuration example of the wind power generation system 101. The wind power generation system 101 includes a wind turbine 110, a generator 120, a measurement unit 1201, a brake 1202, an start / stop unit 1203, a power consumption destination 1204, and a wind turbine control device 100.

The wind turbine 110 receives the wind, converts the wind power into the wind turbine torque, and transmits it to the shaft of the generator 120 as rotational energy. The start / stop unit 1203 controls the brake 1202 as the wind speed increases, and determines whether to stop the wind turbine 110 or operate the wind turbine 110. The brake 1202 stops the wind turbine 110 according to the control of the operation / stop unit 1203. The wind turbine control device 100 may have the start / stop unit 1203.

The generator 120 generates electricity using the wind turbine torque transmitted from the wind turbine 110 to the shaft as rotational energy, and supplies the power to the power consumption destination 1204. The power consumption destination 1204 consumes the generated power. The measuring unit 1201 measures the wind speed, the wind turbine torque of the wind turbine 110, the rotation speed of the wind turbine 110, and the cumulative power generation amount of the generator 120, and outputs the measured values.

The wind turbine control device 100 includes a measurement value acquisition unit 1211, a left / right designation unit 1212, a left side maintenance unit 1213, a follow-up control unit 1214, and a load torque setting unit 1215. The load torque control unit 1102 shown in FIG. 11 is realized by, for example, a measurement value acquisition unit 1211, a left / right designation unit 1212, a left side maintenance unit 1213, a follow-up control unit 1214, and a load torque setting unit 1215.

The measured value acquisition unit 1211 acquires the measured values of the wind speed, the wind turbine torque of the wind turbine 110, the rotation speed of the wind turbine 110, and the cumulative power generation amount of the generator 120 from the measurement unit 1201 at regular intervals. The measured value acquisition unit 1211 updates the latest measured value information 400 shown in FIG. 4 based on the acquired measured values.

The left / right designation unit 1212 outputs a control instruction to the left side maintenance unit 1213 or the follow-up control unit 1214 at regular intervals. The control instruction is generated based on, for example, the latest measured value information 400 shown in FIG. 4, the wind turbine torque characteristic information 500 shown in FIG. 5, and the ratio information 600 shown in FIG. The left / right designation unit 1212 outputs a control instruction to the left side maintenance unit 1213, for example, if the movement flag information 700 shown in FIG. 7 indicates that the operating point is moved to the left side of the mountain. On the other hand, if the left / right designation unit 1212 indicates that the movement flag information 700 shown in FIG. 7 does not move the operating point to the left side of the mountain, the left / right designation unit 1212 outputs a control instruction to the follow-up control unit 1214.

The left side maintenance unit 1213 maintains the operating point of the wind turbine 110 on the left side of the peak of the characteristic curve by controlling the load torque of the generator 120 via the load torque setting unit 1215. The left side maintenance unit 1213 causes the load torque setting unit 1215 to alternately set the load torque of the generator 120 to a value larger and smaller than the wind turbine torque at the operating point of the current wind turbine 110, thereby operating the wind turbine 110. Keep the point to the left of the peak of the characteristic curve.

The follow-up control unit 1214 maintains the operating point of the wind turbine 110 on the right side of the peak of the characteristic curve by controlling the load torque of the generator 120 via the load torque setting unit 1215. The follow-up control unit 1214 maintains the operating point in the region on the right side of the mountain by causing the load torque setting unit 1215 to set the load torque of the generator 120 to the maximum power generation torque that maximizes the power generation efficiency of the generator 120. However, the power generation efficiency of the generator 120 will be maximized.

The load torque setting unit 1215 sets the load torque of the generator 120 according to the control of the left side maintenance unit 1213 or the follow-up control unit 1214.

Further, the wind turbine control device 100 includes an observation unit 1221, a reward function unit 1222, a state update unit 1223, and an action determination unit 1224. The agent 1103 shown in FIG. 11 is realized by, for example, an observation unit 1221, a reward function unit 1222, a state update unit 1223, and an action determination unit 1224.

The observation unit 1221 acquires the measured values of the wind speed, the wind turbine torque of the wind turbine 110, the rotation speed of the wind turbine 110, and the cumulative power generation amount of the generator 120 from the measurement unit 1201 at regular intervals. The observation unit 1221 stores the acquired measured values in the observation history 800 shown in FIG.

The reward function unit 1222 acquires the power generation amount of the generator 120 from the observation history 800 shown in FIG. 8, calculates the reward value corresponding to the power generation amount of the generator 120, and outputs it to the state update unit 1223.

The state update unit 1223 executes reinforcement learning, acquires the wind speed and the rotation speed of the wind turbine 110 from the observation history 800 shown in FIG. 8, acquires the reward value from the reward function unit 1222, and performs the action history shown in FIG. 1000 is acquired and the action value table 900 shown in FIG. 9 is updated.

Based on the action value table 900 shown in FIG. 9, the action determination unit 1224 determines as an action whether or not to move the operating point of the wind turbine 110 to the left side of the mountain of the characteristic curve. The action determination unit 1224 updates the movement flag information 700 shown in FIG. 7 and the action history 1000 shown in FIG. 10 based on the determined action.

(Control index of load torque of generator 120 based on wind turbine torque characteristics)
A control index of the load torque of the generator 120 based on the wind turbine torque characteristics will be described with reference to FIGS. 13 to 19.

13 to 19 are explanatory views showing a control index of the load torque of the generator 120 based on the wind turbine torque characteristic. In the example of FIG. 13, Table 1300 shows the wind turbine torque characteristics for each wind speed and the power generation amount characteristics for each wind speed. The wind turbine torque characteristics for each wind speed are curves 1321 to 1323. The wind turbine torque characteristic is a mountain characteristic. The power generation characteristics for each wind speed are curves 1311 to 1313. The power generation characteristic is a mountainous characteristic. The maximum power generation point indicating the combination of the rotation speed of the wind turbine 110 capable of maximizing the power generation amount of the generator 120 and the wind turbine torque of the wind turbine 110 with respect to the change of the wind speed is on the curve 1301.

Therefore, the operating point of the wind turbine 110 is preferably set to the maximum power generation points a ₀ , a ₁ , and a ₂ at the intersection of the curves 1301 and the curves 1321 to 1323, and is preferably set to the right side of the mountain. An index is obtained. Next, the description proceeds to FIG.

In the example of FIG. 14, Table 1400 shows the wind turbine torque characteristics for each wind speed and the power generation amount characteristics for each wind speed. The wind turbine torque characteristics of the wind speeds V1 and V2 are curves 1421,1422. Wind speed V1 <wind speed V2. The power generation characteristics of the wind speeds V1 and V2 are curves 1411, 1412. The maximum power generation point indicating the combination of the rotation speed of the wind turbine 110 capable of maximizing the power generation amount of the generator 120 and the wind turbine torque of the wind turbine 110 with respect to the change of the wind speed is on the curve 1401.

When the operating point is a1 at the wind speed V1, when the wind speed increases from V1 to V2, if the load torque is constant, the operating point moves from a ₁ to a _2B due to the increase in the wind speed. Therefore, when the wind speed increases from V1 to V2, a load torque having a magnitude commensurate with the wind turbine torque at the maximum power generation point a ₂ of the wind speed V2 is set, and the operating point of the wind turbine 110 is set to the maximum power generation point a ₂ . A control index is obtained that it is preferable to move. Next, the process proceeds to the description of FIG.

In the example of FIG. 15, Table 1500 shows the wind turbine torque characteristics for each wind speed and the power generation amount characteristics for each wind speed, and shows the upper limit of the number of revolutions at which the wind turbine 110 may be damaged. In addition, there is a load torque upper limit that can be set for the load torque. The wind turbine torque characteristics of the wind speeds V2, V3, and V4 are curves 1511, 1512, and 1513. Wind speed V2 <wind speed V3 <wind speed V4. The power generation characteristics of the wind speeds V2, V3, and V4 are curves 1511, 1512, and 1513. The maximum power generation point indicating the combination of the rotation speed of the wind turbine 110 capable of maximizing the power generation amount of the generator 120 and the wind turbine torque of the wind turbine 110 with respect to the change of the wind speed is on the curve 1501.

Here, at the wind speed V3, even if the load torque of the generator 120 is set to the upper limit of the load torque, the operating point of the wind turbine 110 cannot be set to the maximum power generation amount point _a3 . When the rotation speed of the wind turbine 110 exceeds the upper limit of the rotation speed, the brake 1202 is activated, the rotation of the wind turbine 110 is stopped, and the power generation by the generator 120 is stopped. Therefore, when the wind speed V3 is increased to the wind speed V4, the operating point of the wind turbine 110 moves to a ₄ , the rotation speed of the wind turbine 110 exceeds the upper limit of the rotation speed, the brake 1202 is activated, and the rotation of the wind turbine 110 is stopped. Will be done. Once the rotation of the wind turbine 110 is stopped, it takes time to restart the wind turbine 110, which causes a long-term decrease in power generation efficiency. Next, the description proceeds to FIG.

In the example of FIG. 16, Table 1600 is shown. Table 1600 is the same as Table 1500. As shown in FIG. 15, when the wind speed becomes V4, the operating point of the wind turbine 110 moves to a ₄ , the rotation speed of the wind turbine 110 exceeds the upper limit of the rotation speed, and the brake 1202 operates.

Therefore, in the case of the wind speed V4, it is possible to obtain a control index that it is preferable to set the operating point of the wind turbine 110 to the point a _4L on the left side of the peak of the characteristic curve. However, when the wind speed V2 is exceeded, even if the load torque is set to the upper limit of the load torque, the operating point of the wind turbine 110 cannot be moved from the right side to the left side of the mountain. A control index of preference can be obtained. On the other hand, when the wind speed does not reach V4 and the wind speed fluctuates in the vicinity of the wind speed V3, it is possible to obtain a control index that the operating point of the wind turbine 110 is preferably set to the right side of the mountain. Next, the description shifts to FIG.

In the example of FIG. 17, Table 1700 is shown, the wind turbine torque characteristics for each wind speed are shown, and the case where the operating point of the wind turbine 110 is maintained on the right side of the mountain is shown. The wind turbine torque characteristics of the wind speeds V1 and V2 are curves 1701,1702. Wind speed V1 <wind speed V2.

Here, on the right side of the mountain, the operating point of the wind turbine 110 has a property of moving in a direction in which the wind turbine torque and the load torque are balanced when there is a difference between the wind turbine torque and the load torque. Further, when the wind speed increases, the rotation speed increases, and when the wind speed decreases, the rotation speed decreases.

Therefore, even if time elapses after setting the load torque, the operating point of the wind turbine 110 stops at the point where the wind turbine torque and the load torque are balanced, so that the time interval for controlling the load torque is relatively long. A control index that is good is obtained. Next, the description proceeds to FIG.

In the example of FIG. 18, Table 1800 is shown, the wind turbine torque characteristics for each wind speed are shown, and the case where the operating point of the wind turbine 110 is maintained on the left side of the mountain is shown. The wind turbine torque characteristics of the wind speeds V1 and V2 are curves 1801, 1802. Wind speed V1 <wind speed V2.

Here, the left side of the mountain is an unstable region, and on the left side of the mountain, the operating point of the wind turbine 110 moves in the direction in which the wind turbine torque and the load torque are separated from each other when there is a difference between the wind turbine torque and the load torque. There is a property that. Further, when the wind speed increases, the rotation speed increases, and when the wind speed decreases, the rotation speed decreases.

Therefore, when time elapses after the load torque is set, the operating point of the wind turbine 110 is separated from the wind turbine torque and the load torque, and the rotation speed tends to change at an accelerating rate, so that the load torque is controlled. A control index is obtained that the time interval is preferably relatively short. The speed at which the rotation speed changes is specified, for example, based on the following equation (1).

I is the moment of inertia. The moment of inertia is the magnitude of the inertia of rotation in which the rotating body tries to maintain the same rotational motion. The smaller the moment of inertia, the faster the speed of change of the number of revolutions to the steady state. The smaller the weight and length, the smaller the moment of inertia. Therefore, the moment of inertia tends to be small in the small wind power generation system 101.

_TU is the wind turbine torque. λ is the peripheral speed ratio. U is the wind speed. The higher the wind speed, the larger the absolute value of the wind turbine torque. _TL is the load torque. The larger the right side of the above equation (1), the larger the change amount dw / dt of the rotation speed. Therefore, as the wind speed increases, the speed at which the rotation speed changes tends to increase. Next, the description proceeds to FIG.

In the example of FIG. 19, Table 1900 is shown, the wind turbine torque characteristics for each wind speed are shown, and the case where the operating point of the wind turbine 110 is maintained on the left side of the mountain is further shown. The wind turbine torque characteristic is a curve 1901. On the left side of the mountain, the operating point of the wind turbine 110 has the property that when there is a difference between the wind turbine torque and the load torque, the wind turbine torque and the load torque move in a direction away from each other.

Therefore, by alternately setting the load torque for moving the rotation speed in the decreasing direction and the load torque for moving the rotation speed in the increasing direction, the operating point of the wind turbine 110 is not set to zero and is within a certain range. A control index is obtained that it is preferable to store in. The load torque for moving the number of revolutions in the decreasing direction is larger than the wind turbine torque at the current operating point on the left side of the mountain. The load torque for moving the number of revolutions in the increasing direction is smaller than the wind turbine torque at the current operating point on the left side of the mountain. Here, it is preferable to gradually narrow the range in which the operating point of the wind turbine 110 is accommodated.

(Flow to control the load torque of the generator 120)
Next, the flow in which the wind turbine control device 100 controls the load torque of the generator 120 will be described with reference to FIG. 20 in consideration of the control indexes shown in FIGS. 13 to 19.

FIG. 20 is an explanatory diagram showing a flow for controlling the load torque of the generator 120. In FIG. 20, the wind turbine control device 100 controls the load torque so that the operating point of the wind turbine 110 is on the right side of the mountain. When the agent 1103 determines the timing for moving the operating point of the wind turbine 110 to the left side of the mountain, the wind turbine control device 100 moves the operating point of the wind turbine 110 to the left side of the mountain, and the operating point of the wind turbine 110 is set. Control the load torque so that it is maintained on the left side of the mountain. It is preferable that the interval for controlling the load torque is set to be shorter than the time required for the action determination by the agent 1103.

Table 2010 corresponds to the case where the load torque is controlled so that the operating point of the wind turbine 110 is on the right side of the mountain. The wind turbine torque characteristic is a curve 2011. The maximum power generation point indicating the combination of the rotation speed of the wind turbine 110 capable of maximizing the power generation amount of the generator 120 and the wind turbine torque of the wind turbine 110 is on the curve 2012. For example, the wind turbine control device 100 acquires the wind speed from the measuring device, calculates the load torque of the maximum power generation point, and sets it in the generator 120.

Further, the wind turbine control device 100 controls, for example, to lower the load torque when the rotation speed drops to the rotation speed of the maximum point or the rotation speed of the maximum point + the margin. Specifically, the wind turbine control device 100 sets the load torque to 0, 80% of the immediately preceding set value, or the like, as a value that brings the operating point closer to the maximum power generation point.

Further, the wind turbine control device 100 controls, for example, to increase the load torque when the rotation speed rises to the rotation speed upper limit or the rotation speed upper limit-margin. Specifically, the wind turbine control device 100 sets the load torque to the upper limit of the load torque, 120% of the immediately preceding set value, or the like as a value for absorbing the rotational energy due to the increase in the wind speed.

Table 2020 corresponds to the case where the load torque is controlled so that the operating point of the wind turbine 110 is on the left side of the mountain. The wind turbine torque characteristic is curve 2021. The wind turbine control device 100 acquires, for example, a wind speed from a measuring device and calculates the current wind turbine torque.

For example, if the calculated wind turbine torque is larger than the target torque, the wind turbine control device 100 sets the wind turbine torque at the same rotation speed and the wind speed of 120% as the load torque. The target torque is, for example, the load torque upper limit-margin. Then, the wind turbine control device 100 allows the wind speed up to 120% even if the wind speed increases, and changes the rotation speed and the wind turbine torque in the decreasing direction.

For example, if the calculated wind turbine torque is smaller than the target torque, the wind turbine control device 100 sets the wind turbine torque at the same rotation speed and the wind speed of 80% as the load torque. Then, the wind turbine control device 100 allows the wind speed up to 80% even if the wind speed decreases, and changes the rotation speed and the wind turbine torque in the increasing direction.

For example, when the operating point is on the right side of the mountain, the wind turbine control device 100 sets the load torque upper limit to the load torque. Then, the wind turbine control device 100 changes the rotation speed in the decreasing direction. As a result, the wind turbine control device 100 can stably control the wind turbine torque and the rotation speed by controlling the load torque regardless of the time required for the action determination of the agent 1103, and can maximize the power generation efficiency. can.

(Operation example for controlling the load torque of the generator 120)
Next, an operation example in which the wind turbine control device 100 controls the load torque of the generator 120 will be described with reference to FIGS. 21 to 30.

21 to 30 are explanatory views showing an operation example of controlling the load torque of the generator 120. In the examples of FIGS. 21 to 30, the wind speed is measured at 1-second intervals. Further, the rotation speed is measured according to the control interval of the load torque. The control interval is, for example, 0.01 to 0.1 seconds. For the amount of power generation, the difference between the integrated values is measured at 1-second intervals. In the following description, the time obtained by collectively acquiring the wind speed, the number of revolutions, and the amount of power generation may be referred to as "measured time".

Further, when the wind speed exceeds 6 m / s, the wind turbine torque exceeds the load torque upper limit. Further, when the wind speed is 8 m / s or more, even if the load torque is set to the upper limit of the load torque, the rotation speed exceeds the upper limit of the rotation speed of 500 rpm. The generated power is about 20 W at a wind speed of 6 m / s, and the integrated power value, which is a measured value, is about 5.5 mWh. Reinforcement learning is performed by Q-Learning by discretizing the wind speed and the number of revolutions with a width of 1 m / s and 50 rpm, respectively. Next, the description proceeds to FIGS. 21 and 22.

The examples of FIGS. 21 and 22 show changes in action value due to reinforcement learning when the operating point of the wind turbine 110 is set to the right side of the mountain in the area A where the wind speed is 5 m / s to 7 m / s. First, the description shifts to FIG.

The first state of FIG. 21 is a state in which the flag for moving to the left = 0 at a wind speed of 6 m / s, and the wind turbine control device 100 controls the operating point of the wind turbine 110 on the right side of the mountain.

(21-1) The wind turbine control device 100 acquires the measured value from the measuring unit and adds it to the observation history 800. The wind turbine control device 100 newly acquires a measured value such as a wind speed of 7 m / s, and adds a record 2101 to the observation history 800. The amount of power generation is the result of the action at the previous measurement time.

(21-2) The wind turbine control device 100 calculates the reward. The wind turbine control device 100 uses, for example, the latest power generation amount of 5.03 as a reward. The wind turbine control device 100 calculates a Q value of 2.51 as the value of the action at the previous measurement time.

From the action value table 900, the wind turbine control device 100 meets the conditions of the wind speed and the rotation speed at the previous measurement time and the wind speed and the rotation speed at the previous measurement time, and at the previous measurement time. The record 2102 showing the behavior of is specified. The wind turbine control device 100 updates the Q value of the specified record 2102 with the calculated Q value.

As a result, the wind turbine control device 100 has the operating point of the wind turbine 110 on the right side of the mountain in response to various changes in the environment, which is a preferable behavior from the viewpoint of maximizing the power generation efficiency of the generator 120. Information indicating the presence can be stored. Next, the description proceeds to FIG. 22.

The example of FIG. 22 is a continuation of the example of FIG. The first state of FIG. 22 is a state in which the wind speed is 7 m / s even at the measurement time next to the last state of FIG. 21, and the wind turbine control device 100 controls the operating point of the wind turbine 110 on the right side of the mountain.

(22-1) The wind turbine control device 100 acquires the measured value from the measuring unit and adds it to the observation history 800. The wind turbine control device 100 newly acquires a measured value such as a wind speed of 7 m / s, and adds a record 2201 to the observation history 800. The amount of power generation is the result of the action at the previous measurement time.

(22-2) The wind turbine control device 100 calculates the reward. The wind turbine control device 100 uses, for example, the latest power generation amount of 8.56 as a reward. The wind turbine control device 100 calculates a Q value of 4.28 as the value of the action at the previous measurement time.

From the action value table 900, the wind turbine control device 100 meets the conditions of the wind speed and the rotation speed at the previous measurement time and the wind speed and the rotation speed at the previous measurement time, and at the previous measurement time. Identify record 2202 showing the behavior of. The wind turbine control device 100 updates the Q value of the specified record 2202 with the calculated Q value.

As a result, the wind turbine control device 100 has the operating point of the wind turbine 110 on the right side of the mountain in response to various changes in the environment, which is a preferable behavior from the viewpoint of maximizing the power generation efficiency of the generator 120. Information indicating the presence can be stored. Next, the description proceeds to FIGS. 23 and 24.

In the example of FIGS. 23 and 24, in the area A where the wind speed is 5 m / s to 7 m / s as in FIGS. 21 and 22, unlike the case of FIGS. 21 and 22, the operating point of the wind turbine 110 is set to the left side of the mountain. The change in behavioral value due to reinforcement learning is shown. First, the description shifts to FIG. 23.

The first state of FIG. 23 is a state in which the flag for moving to the left = 1 at a wind speed of 6 m / s, and the wind turbine control device 100 controls the operating point of the wind turbine 110 on the left side of the mountain.

(23-1) The wind turbine control device 100 acquires the measured value from the measuring unit and adds it to the observation history 800. The wind turbine control device 100 newly acquires a measured value such as a wind speed of 7 m / s, and adds a record 2301 to the observation history 800. The amount of power generation is the result of the action at the previous measurement time.

(23-2) The wind turbine control device 100 calculates the reward. The wind turbine control device 100 uses, for example, the latest power generation amount of 4.75 as a reward. The wind turbine control device 100 calculates a Q value of 2.37 as the value of the action at the previous measurement time.

From the action value table 900, the wind turbine control device 100 meets the conditions of the wind speed and the rotation speed at the previous measurement time and the wind speed and the rotation speed at the previous measurement time, and at the previous measurement time. Identify record 2302 showing the behavior of. The wind turbine control device 100 updates the Q value of the specified record 2302 with the calculated Q value.

As a result, the wind turbine control device 100 has the operating point of the wind turbine 110 on the left side of the mountain in response to various changes in the environment, which is a preferable behavior from the viewpoint of maximizing the power generation efficiency of the generator 120. Information indicating the presence can be stored. The wind turbine control device 100 can store records having the same wind speed and rotation speed conditions but different behaviors, such as record 2102 and record 2302.

Therefore, the wind turbine control device 100 can specify which action has a larger Q value when the current and past wind speeds and rotation speeds meet the wind speed and rotation speed conditions. can. An action with a large Q value is a preferable action from the viewpoint of maximizing the power generation efficiency of the generator 120. Next, the description proceeds to FIG. 24.

The example of FIG. 24 is a continuation of the example of FIG. The first state of FIG. 24 is a state in which the wind speed is 7 m / s even at the measurement time following the last state of FIG. 23, and the wind turbine control device 100 controls the operating point of the wind turbine 110 on the left side of the mountain.

(24-1) The wind turbine control device 100 acquires the measured value from the measuring unit and adds it to the observation history 800. The wind turbine control device 100 newly acquires a measured value such as a wind speed of 7 m / s, and adds a record 2401 to the observation history 800. The amount of power generation is the result of the action at the previous measurement time.

(24-2) The wind turbine control device 100 calculates the reward. The wind turbine control device 100 uses, for example, the latest power generation amount 3.31 as a reward. The wind turbine control device 100 calculates a Q value of 1.65 as the value of the action at the previous measurement time.

From the action value table 900, the wind turbine control device 100 meets the conditions of the wind speed and the rotation speed at the current measurement time and the wind speed and the rotation speed at the previous measurement time, and the action at the previous measurement time. The record 2402 indicating the above is specified. The wind turbine control device 100 updates the Q value of the specified record 2402 with the calculated Q value.

As a result, the wind turbine control device 100 has the operating point of the wind turbine 110 on the left side of the mountain in response to various changes in the environment, which is a preferable behavior from the viewpoint of maximizing the power generation efficiency of the generator 120. Information indicating the presence can be stored. The wind turbine control device 100 can store, for example, a record 2202 and a record 2402, which have the same wind speed and rotation speed conditions but different behaviors.

Therefore, the wind turbine control device 100 can specify which action has a larger Q value when the current and past wind speeds and rotation speeds meet the wind speed and rotation speed conditions. can. An action with a large Q value is a preferable action from the viewpoint of maximizing the power generation efficiency of the generator 120. Next, the process proceeds to the description of FIGS. 25 and 26.

The examples of FIGS. 25 and 26 show changes in action value due to reinforcement learning when the operating point of the wind turbine 110 is set to the right side of the mountain in the area B where the wind speed is 5 m / s to 8 m / s. First, the description shifts to FIG. 25.

The first state of FIG. 25 is a state in which the flag for moving to the left = 0 at a wind speed of 6 m / s, and the wind turbine control device 100 controls the operating point of the wind turbine 110 on the right side of the mountain.

(25-1) The wind turbine control device 100 acquires the measured value from the measuring unit and adds it to the observation history 800. The wind turbine control device 100 newly acquires a measured value such as a wind speed of 7 m / s, and adds a record 2501 to the observation history 800. The amount of power generation is the result of the action at the previous measurement time.

(25-2) The wind turbine control device 100 calculates the reward. The wind turbine control device 100 uses, for example, the latest power generation amount of 5.03 as a reward. The wind turbine control device 100 calculates a Q value of 2.51 as the value of the action at the previous measurement time.

From the action value table 900, the wind turbine control device 100 meets the conditions of the wind speed and the rotation speed at the previous measurement time and the wind speed and the rotation speed at the previous measurement time, and at the previous measurement time. Identify record 2502 showing the behavior of. The wind turbine control device 100 updates the Q value of the specified record 2502 with the calculated Q value.

As a result, the wind turbine control device 100 has the operating point of the wind turbine 110 on the right side of the mountain in response to various changes in the environment, which is a preferable behavior from the viewpoint of maximizing the power generation efficiency of the generator 120. Information indicating the presence can be stored. Next, the description shifts to FIG. 26.

The example of FIG. 26 is a continuation of the example of FIG. The first state of FIG. 26 is a wind speed of 7 m / s even at the measurement time following the last state of FIG. 25, and is a state after the wind turbine control device 100 controls the operating point of the wind turbine 110 on the right side of the mountain. The first state of FIG. 26 is a state in which the wind speed is 8 m / s at the next measurement time and the operating point of the wind turbine 110 is controlled on the right side of the mountain.

(26-1) The wind turbine control device 100 acquires the measured value from the measuring unit and adds it to the observation history 800. The wind turbine control device 100 newly acquires a measured value such as a wind speed of 8 m / s, and adds a record 2601 to the observation history 800. The amount of power generation is the result of the action at the previous measurement time. Here, since the wind speed is 8 m / s, the rotation of the wind turbine 110 is stopped by the brake, and the amount of power generation becomes small.

(26-2) The wind turbine control device 100 calculates the reward. The wind turbine control device 100 uses, for example, the latest power generation amount of 0.85 as a reward. The wind turbine control device 100 calculates a Q value of 0.42 as the value of the action at the previous measurement time.

From the action value table 900, the wind turbine control device 100 meets the conditions of the wind speed and the rotation speed at the previous measurement time and the wind speed and the rotation speed at the previous measurement time, and at the previous measurement time. Identify record 2602 showing the behavior of. The wind turbine control device 100 updates the Q value of the specified record 2602 with the calculated Q value.

As a result, the wind turbine control device 100 has the operating point of the wind turbine 110 on the right side of the mountain in response to various changes in the environment, which is a preferable behavior from the viewpoint of maximizing the power generation efficiency of the generator 120. Information indicating the presence can be stored. Further, the wind turbine control device 100 can generate different action value tables 900 between the wind power generation system in region A and the wind power generation system in region B. Next, the description proceeds to FIGS. 27 and 28.

In the example of FIGS. 27 and 28, in the area B where the wind speed is 5 m / s to 8 m / s as in FIGS. 25 and 26, the operating point of the wind turbine 110 is set to the left side of the mountain unlike FIGS. 25 and 26. The change in behavioral value due to reinforcement learning is shown. First, the description shifts to FIG. 27.

The first state of FIG. 27 is a state in which the flag for moving to the left = 1 at a wind speed of 6 m / s, and the wind turbine control device 100 controls the operating point of the wind turbine 110 on the left side of the mountain.

(27-1) The wind turbine control device 100 acquires the measured value from the measuring unit and adds it to the observation history 800. The wind turbine control device 100 newly acquires a measured value such as a wind speed of 7 m / s, and adds a record 2701 to the observation history 800. The amount of power generation is the result of the action at the previous measurement time.

(27-2) The wind turbine control device 100 calculates the reward. The wind turbine control device 100 uses, for example, the latest power generation amount of 4.75 as a reward. The wind turbine control device 100 calculates a Q value of 2.37 as the value of the action at the previous measurement time.

From the action value table 900, the wind turbine control device 100 meets the conditions of the wind speed and the rotation speed at the previous measurement time and the wind speed and the rotation speed at the previous measurement time, and at the previous measurement time. Identify record 2702 showing the behavior of. The wind turbine control device 100 updates the Q value of the specified record 2702 with the calculated Q value.

As a result, the wind turbine control device 100 has the operating point of the wind turbine 110 on the left side of the mountain in response to various changes in the environment, which is a preferable behavior from the viewpoint of maximizing the power generation efficiency of the generator 120. Information indicating the presence can be stored. The wind turbine control device 100 can store records such as record 2502 and record 2702, which have the same wind speed and rotation speed conditions but different behaviors.

Therefore, the wind turbine control device 100 can specify which action has a larger Q value when the current and past wind speeds and rotation speeds meet the wind speed and rotation speed conditions. can. An action with a large Q value is a preferable action from the viewpoint of maximizing the power generation efficiency of the generator 120. Next, the description shifts to FIG. 28.

The example of FIG. 28 is a continuation of the example of FIG. 27. The first state of FIG. 28 is a state after the wind speed is 7 m / s even at the measurement time next to the last state of FIG. 27, and the wind turbine control device 100 controls the operating point of the wind turbine 110 on the left side of the mountain. The first state of FIG. 28 is a state in which the wind speed is 8 m / s at the next measurement time, and the wind turbine control device 100 controls the operating point of the wind turbine 110 on the left side of the mountain.

(28-1) The wind turbine control device 100 acquires the measured value from the measuring unit and adds it to the observation history 800. The wind turbine control device 100 newly acquires a measured value such as a wind speed of 7 m / s, and adds a record 2801 to the observation history 800. The amount of power generation is the result of the action at the previous measurement time. Here, since the operating point of the wind turbine 110 is controlled on the left side of the mountain, the rotation of the wind turbine 110 is not stopped by the brake, and the amount of power generation is larger than that in the case where the rotation is stopped.

(28-2) The wind turbine control device 100 calculates the reward. The wind turbine control device 100 uses, for example, the latest power generation amount of 3.27 as a reward. The wind turbine control device 100 calculates a Q value of 1.63 as the value of the action at the previous measurement time.

From the action value table 900, the wind turbine control device 100 meets the conditions of the wind speed and the rotation speed at the current measurement time and the wind speed and the rotation speed at the previous measurement time, and the action at the previous measurement time. The record 2802 indicating the above is specified. The wind turbine control device 100 updates the Q value of the specified record 2802 with the calculated Q value.

As a result, the wind turbine control device 100 has the operating point of the wind turbine 110 on the left side of the mountain in response to various changes in the environment, which is a preferable behavior from the viewpoint of maximizing the power generation efficiency of the generator 120. Information indicating the presence can be stored. The wind turbine control device 100 can store, for example, a record 2602 and a record 2802, which have the same wind speed and rotation speed conditions but different behaviors.

Therefore, the wind turbine control device 100 can specify which action has a larger Q value when the current and past wind speeds and rotation speeds meet the wind speed and rotation speed conditions. can. An action with a large Q value is a preferable action from the viewpoint of maximizing the power generation efficiency of the generator 120. Specifically, the wind turbine control device 100 can specify that the action of controlling the operating point of the wind turbine 110 on the left side of the mountain is preferable because the rotation of the wind turbine 110 is not stopped by the brake. Further, the wind turbine control device 100 can generate different action value tables 900 between the wind power generation system in region A and the wind power generation system in region B. Next, the description proceeds to FIG. 29.

The example of FIG. 29 shows an example in which the wind turbine control device 100 determines the action by the agent 1103 based on the action value table 900 in the area A where the wind speed is 5 m / s to 7 m / s. The first state of FIG. 29 is a state of a wind speed of 6 m / s.

(29-1) The wind turbine control device 100 acquires the measured value from the measuring unit and adds it to the observation history 800. The wind turbine control device 100 newly acquires a measured value such as a wind speed of 6 m / s, and adds a record 2901 to the observation history 800. The amount of power generation is the result of the action at the previous measurement time.

(29-2) Based on the action value table 900, the wind turbine control device 100 has a record 2902 in which the wind speed and the rotation speed at the current measurement time and the wind speed and the rotation speed at the previous measurement time meet the conditions. Identify 2903. Then, the wind turbine control device 100 determines the action indicated by the record 2902 having the larger Q value among the records 2902 and 2903 as the next action.

Thereby, the wind turbine control device 100 can determine a preferable action from the viewpoint of maximizing the power generation efficiency of the generator 120 in response to various changes in the environment in the region A. The wind turbine control device 100 determines whether or not to move the operating point of the wind turbine 110 to the left side of the mountain so as to maximize the power generation efficiency of the generator 120 with respect to various environmental changes in the region A, for example. be able to. Next, the process proceeds to the description of FIG.

The example of FIG. 30 shows an example in which the wind turbine control device 100 determines the action by the agent 1103 based on the action value table 900 in the area B where the wind speed is 5 m / s to 8 m / s. The first state of FIG. 30 is a state of a wind speed of 6 m / s.

(30-1) The wind turbine control device 100 acquires the measured value from the measuring unit and adds it to the observation history 800. The wind turbine control device 100 newly acquires a measured value such as a wind speed of 6 m / s, and adds a record 3001 to the observation history 800. The amount of power generation is the result of the action at the previous measurement time.

(30-2) Based on the action value table 900, the wind turbine control device 100 has a record 3002 in which the wind speed and the rotation speed at the current measurement time and the wind speed and the rotation speed at the previous measurement time meet the conditions. Identify 3003. Then, the wind turbine control device 100 determines the action indicated by the record 3003 having the larger Q value among the records 3002 and 3003 as the next action.

Thereby, the wind turbine control device 100 can determine a preferable action from the viewpoint of maximizing the power generation efficiency of the generator 120 in response to various changes in the environment in the region B. The wind turbine control device 100 determines whether or not to move the operating point of the wind turbine 110 to the left side of the mountain so as to maximize the power generation efficiency of the generator 120 with respect to various environmental changes in the region B , for example. be able to.

(Overall processing procedure)
Next, an example of the overall processing procedure executed by the wind turbine control device 100 will be described with reference to FIG. 31. The entire processing is realized, for example, by the CPU 301 shown in FIG. 3, a storage area such as a memory 302 or a recording medium 305, and a communication I / F 303.

FIG. 31 is a flowchart showing an example of the overall processing procedure. In FIG. 31, first, the wind turbine control device 100 measures the wind speed, the rotation speed, and the amount of power generation at the sampling intervals corresponding to each (step S3101). Next, the wind turbine control device 100 acquires the measured value and updates the latest measured value (step S3102). Then, the wind turbine control device 100 acquires the measured value and updates the observation history 800 (step S3103).

Next, the wind turbine control device 100 executes the operation / stop process described later in FIG. 32 based on the measured values of the wind speed and the rotation speed (step S3104). Then, the wind turbine control device 100 executes the reinforcement learning process described later in FIG. 33, and updates the left movement presence / absence flag (step S3105). After that, the wind turbine control device 100 executes the load torque control process described later in FIG. 34 to set the load torque of the generator 120 (step S3106).

Next, the wind turbine control device 100 determines whether or not to end the entire process (step S3107). Here, if the entire process is not completed (step S3107: No), the wind turbine control device 100 shifts to the process of step S3101. On the other hand, when the whole process is finished (step S3107: Yes), the wind turbine control device 100 ends the whole process.

(Operation / stop processing procedure)
Next, an example of the operation / stop processing procedure executed by the wind turbine control device 100 will be described with reference to FIG. 32. The operation / stop processing procedure is realized, for example, by the CPU 301 shown in FIG. 3, a storage area such as a memory 302 or a recording medium 305, and a communication I / F 303.

FIG. 32 is a flowchart showing an example of the operation / stop processing procedure. In FIG. 32, the wind turbine control device 100 acquires the latest wind speed and rotation speed (step S3201).

Next, the wind turbine control device 100 determines whether or not the acquired wind speed is equal to or higher than the cutoff wind speed (step S3202). Here, when the wind speed is equal to or higher than the cut-off wind speed (step S3202: Yes), the wind turbine control device 100 shifts to the process of step S3205. On the other hand, when the wind speed is less than the cut-off wind speed (step S3202: No), the wind turbine control device 100 shifts to the process of step S3203.

In step S3203, the wind turbine control device 100 determines whether or not the acquired wind speed is less than the cut-in wind speed (step S3203). Here, when the wind speed is less than the cut-in wind speed (step S3203: Yes), the wind turbine control device 100 shifts to the process of step S3205. On the other hand, when the wind speed is equal to or higher than the cut-in wind speed (step S3203: No), the wind turbine control device 100 shifts to the process of step S3204.

In step S3204, the wind turbine control device 100 determines whether or not the acquired rotation speed is equal to or greater than the threshold value for the rotation speed upper limit (step S3204). Here, when the value is equal to or higher than the threshold value for the upper limit of the rotation speed (step S3204: Yes), the wind turbine control device 100 shifts to the process of step S3205. On the other hand, when it is less than the threshold value regarding the upper limit of the rotation speed (step S3204: No), the wind turbine control device 100 shifts to the process of step S3206.

In step S3205, the wind turbine control device 100 stops the wind turbine 110 by the brake (step S3205). Then, the wind turbine control device 100 ends the operation / stop process.

In step S3206, the wind turbine control device 100 releases the stop of the wind turbine 110 by the brake (step S3206). Then, the wind turbine control device 100 ends the operation / stop process.

(Reinforcement learning processing procedure)
Next, an example of the reinforcement learning processing procedure executed by the wind turbine control device 100 will be described with reference to FIG. 33. The reinforcement learning processing procedure is realized, for example, by the CPU 301 shown in FIG. 3, a storage area such as a memory 302 or a recording medium 305, and a communication I / F 303.

FIG. 33 is a flowchart showing an example of the reinforcement learning processing procedure. In FIG. 33, the wind turbine control device 100 acquires the latest K records of the observation history 800 (step S3301).

Next, the wind turbine control device 100 acquires the previous action from the action history (step S3302). Then, the wind turbine control device 100 calculates the latest power generation amount, which is the difference between the cumulative power generation amount up to the previous reinforcement learning process and the cumulative power generation amount up to the current reinforcement learning process, as a reward (step S3303).

Next, the wind turbine control device 100 sets the combination of K wind speeds and rotation speeds as a state, and updates the action value table 900 by the reinforcement learning algorithm based on the acquired previous action and the calculated reward. (Step S3304).

Then, the wind turbine control device 100 sets the combination of K wind speeds and rotation speeds as a state, refers to the action value table 900, determines whether or not to move to the left as an action by the action selection algorithm, and whether or not to move to the left. Update the flag (step S3305). After that, the wind turbine control device 100 ends the reinforcement learning process.

(Load torque control processing procedure)
Next, an example of the load torque control processing procedure executed by the wind turbine control device 100 will be described with reference to FIG. 34. The load torque control processing procedure is realized, for example, by the CPU 301 shown in FIG. 3, a storage area such as a memory 302 or a recording medium 305, and a communication I / F 303.

FIG. 34 is a flowchart showing an example of the load torque control processing procedure. In FIG. 34, the wind turbine control device 100 acquires the latest wind speed, rotation speed, and power generation amount from the latest measured value (step S3401).

Next, the wind turbine control device 100 acquires a flag for whether or not to move to the left (step S3402). Then, the wind turbine control device 100 acquires the wind turbine torque characteristic (step S3403).

Next, the wind turbine control device 100 determines whether or not to move the operating point of the wind turbine 110 to the left side of the mountain (step S3404). Here, when moving to the left side of the mountain (step S3404: Yes), the wind turbine control device 100 shifts to the process of step S3405. On the other hand, when it is not moved to the left side of the mountain (step S3404: No), the wind turbine control device 100 shifts to the process of step S3406.

In step S3405, the wind turbine control device 100 executes the left side control process described later in FIG. 35 to determine the set value of the load torque of the generator 120 (step S3405). Then, the wind turbine control device 100 shifts to the process of step S3407.

In step S3406, the wind turbine control device 100 executes the right side control process described later in FIG. 36 to determine the set value of the load torque of the generator 120 (step S3406). Then, the wind turbine control device 100 shifts to the process of step S3407.

In step S3407, the wind turbine control device 100 sets the determined load torque set value of the generator 120 to the generator 120 (step S3407). Then, the wind turbine control device 100 ends the load torque control process.

(Left side control processing procedure)
Next, an example of the left side control processing procedure executed by the wind turbine control device 100 will be described with reference to FIG. 35. The left side control processing procedure is realized, for example, by the CPU 301 shown in FIG. 3, a storage area such as a memory 302 or a recording medium 305, and a communication I / F 303.

FIG. 35 is a flowchart showing an example of the left side control processing procedure. In FIG. 35, the wind turbine control device 100 calculates the current wind turbine torque based on the latest wind speed and rotation speed and the wind turbine torque characteristics (step S3501).

Next, the wind turbine control device 100 determines whether or not the calculated wind turbine torque is equal to or greater than the reference torque on the left side of the mountain (step S3502). Here, when the torque is equal to or higher than the reference torque (step S3502: Yes), the wind turbine control device 100 shifts to the process of step S3503. On the other hand, when the torque is less than the reference torque (step S3502: No), the wind turbine control device 100 shifts to the process of step S3504.

In step S3503, the wind turbine control device 100 calculates the wind turbine torque corresponding to the wind speed adjusted in the increasing direction from the latest wind speed at the same rotation speed as the latest rotation speed, and determines the set value of the load torque of the generator 120. (Step S3503). Then, the wind turbine control device 100 shifts to the process of step S3505.

In step S3504, the wind turbine control device 100 calculates the wind turbine torque corresponding to the wind speed adjusted in a direction decreasing from the latest wind speed at the same rotation speed as the latest rotation speed, and determines the set value of the load torque of the generator 120. (Step S3504). Then, the wind turbine control device 100 shifts to the process of step S3505.

In step S3505, the wind turbine control device 100 determines whether or not the latest rotation speed is equal to or higher than the threshold value for the rotation speed upper limit on the left side of the mountain (step S3505). Here, when the value is equal to or higher than the threshold value for the upper limit of the rotation speed (step S3505: Yes), the wind turbine control device 100 shifts to the process of step S3506. On the other hand, when it is less than the threshold value regarding the upper limit of the rotation speed (step S3505: No), the wind turbine control device 100 ends the left side control process.

In step S3506, the wind turbine control device 100 changes the set value of the load torque of the generator 120 to the set value at the time of reaching the threshold value regarding the upper limit of the rotation speed on the left side of the mountain (step S3506). Then, the wind turbine control device 100 ends the left side control process.

(Right side control processing procedure)
Next, an example of the right side control processing procedure executed by the wind turbine control device 100 will be described with reference to FIG. 36. The right-hand control processing procedure is realized, for example, by the CPU 301 shown in FIG. 3, a storage area such as a memory 302 or a recording medium 305, and a communication I / F 303.

FIG. 36 is a flowchart showing an example of the right side control processing procedure. In FIG. 36, the wind turbine control device 100 calculates the load torque at the maximum power generation point based on the latest wind speed and rotation speed, the wind turbine torque characteristic, and the ratio of the maximum power generation torque to the maximum torque, and calculates the load torque of the generator 120. The set value of the load torque is determined (step S3601).

Next, the wind turbine control device 100 determines whether or not the latest rotation speed is equal to or higher than the threshold value for the rotation speed upper limit (step S3602). Here, when the value is equal to or higher than the threshold value for the upper limit of the rotation speed (step S3602: Yes), the wind turbine control device 100 shifts to the process of step S3603. On the other hand, if it is less than the threshold value regarding the upper limit of the rotation speed (step S3602: No), the wind turbine control device 100 shifts to the process of step S3604.

In step S3603, the wind turbine control device 100 changes the set value of the load torque of the generator 120 to the set value when the threshold value for the upper limit of the rotation speed is reached (step S3603). Then, the wind turbine control device 100 ends the right side control process.

In step S3604, the wind turbine control device 100 determines whether or not the latest rotation speed is equal to or less than the threshold value for the rotation speed lower limit (step S3604). Here, when it is equal to or less than the threshold value for the lower limit of the rotation speed (step S3604: Yes), the wind turbine control device 100 shifts to the process of step S3605. On the other hand, when it is larger than the threshold value regarding the lower limit of the rotation speed (step S3604: No), the wind turbine control device 100 ends the right side control process.

In step S3605, the wind turbine control device 100 changes the set value of the load torque of the generator 120 to the set value when the threshold value for the lower limit of the rotation speed is reached (step S3605). Then, the wind turbine control device 100 ends the right side control process.

As described above, according to the wind turbine control device 100, the load torque of the generator 120 using the wind turbine 110 can be controlled based on the wind speed and the rotation speed of the wind turbine 110. According to the wind turbine control device 100, the wind speed and the rotation speed of the wind turbine 110 are used as observation values, the amount of power generated by the generator 120 is used as a reward, and whether or not the operating point of the wind turbine 110 is moved to the left side of the mountain is an action. The enhanced learning can be executed by the agent 1103. According to the wind turbine control device 100, when the agent 1103 to which the wind speed and the rotation speed of the wind turbine are input outputs to move the operating point to the left side of the mountain, the load torque of the generator 120 is controlled. , The operating point can be moved to the left side of the mountain. As a result, the wind turbine control device 100 can appropriately control the wind power generation system 101 in accordance with changes in the environment such as wind speed, which causes a long-term decrease in power generation efficiency of the generator 120. It can be avoided.

According to the wind turbine control device 100, the wind power generation system 101 can be controlled even if the wind turbine 110 does not have the function of controlling the pitch of the wind turbine 110. As a result, the wind turbine control device 100 can increase the types of the wind power generation system 101 to which the wind turbine control device 100 can be applied. The wind turbine control device 100 can reduce the manufacturing cost of the wind power generation system 101.

According to the wind turbine control device 100, the load torque of the generator 120 can be controlled based on the wind speed, the rotation speed of the wind turbine 110, and the current operating point on the characteristic curve. As a result, the wind turbine control device 100 can set the load torque according to the current operating point, and can easily maintain the current operating point.

According to the wind turbine control device 100, the wind power generation system 101 can be controlled even if the interval for controlling the load torque of the generator 120 is shorter than the interval for outputting the action by the agent 1103. As a result, the wind turbine control device 100 can suppress a decrease in the speed of controlling the load torque of the generator 120 in response to a change in the wind speed, regardless of the execution interval of the action determination of the agent 1103.

According to the wind turbine control device 100, the operating point can be moved to the left side of the mountain by setting the load torque of the generator 120 to a value larger than the wind turbine torque at the maximum point. As a result, the wind turbine control device 100 can improve the speed at which the operating point is moved to the left side of the mountain, and can easily cope with the case where the speed at which the wind speed changes is large.

According to the wind turbine control device 100, when the current operating point on the characteristic curve is on the left side of the mountain, the load torque of the generator 120 is set to a value larger than the wind turbine torque at the operating point and a value smaller than the wind turbine torque at the operating point. And, it can be set alternately. As a result, the wind turbine control device 100 can maintain the operating point within a certain range, suppress fluctuations in the rotation speed, and suppress the load applied to the wind turbine 110.

According to the wind turbine control device 100, the load torque of the generator 120 can be controlled based on the wind speed, the rotation speed of the wind turbine 110, and the amount of power generated by the generator 120. As a result, the wind turbine control device 100 can control the operating point by controlling the load torque without setting the ratio of the maximum power generation torque to the maximum torque.

The wind turbine control method described in the present embodiment can be realized by executing a program prepared in advance on a computer such as a personal computer or a workstation. The wind turbine control program described in this embodiment is recorded on a computer-readable recording medium such as a hard disk, flexible disk, CD-ROM, MO, or DVD, and is executed by being read from the recording medium by the computer. Further, the wind turbine control program described in the present embodiment may be distributed via a network such as the Internet.

The following additional notes are further disclosed with respect to the above-described embodiment.

(Appendix 1) To the computer
The load torque of the generator using the wind turbine is controlled based on the wind speed and the rotation speed of the wind turbine.
The maximum point is the operating point on the characteristic curve showing the relationship between the rotation speed of the wind turbine and the wind turbine torque of the wind turbine, with the wind speed and the rotation speed of the wind turbine as the observed values and the amount of power generated by the generator as the reward. The agent is made to execute reinforcement learning with the action of whether or not to move to the area on the side where the rotation speed of the wind turbine is smaller than that of the wind turbine.
When the agent to which the wind speed and the rotation speed of the wind turbine are input outputs to move the operating point to the region, the operating point is moved to the region by controlling the load torque of the generator. Move to,
A wind turbine control program characterized by executing processing.

(Appendix 2) The wind turbine control program according to Appendix 1, wherein the wind turbine does not have a function of controlling the pitch of the wind turbine.

(Appendix 3) The process to be controlled is characterized in that the load torque of the generator is controlled based on the wind speed, the rotation speed of the wind turbine, and the current operating point on the characteristic curve. Or the wind turbine control program described in 2.

(Supplementary Note 4) The wind turbine control program according to any one of Supplementary note 1 to 3, wherein the interval for controlling the load torque of the generator is shorter than the interval for outputting the action by the agent.

(Appendix 5) The moving process is characterized in that the operating point is moved to the region by setting the load torque of the generator to a value larger than the wind turbine torque at the maximum point. The wind turbine control program according to any one of 4 to 4.

(Appendix 6) In the process to be controlled, when the current operating point on the characteristic curve is in the region, the load torque of the generator is set to a value larger than the wind turbine torque at the operating point, and the wind turbine at the operating point. The wind turbine control program according to Appendix 3, wherein the operating point is maintained in the region by alternately setting a value smaller than the torque.

(Appendix 7) The process to be controlled is characterized in that the load torque of the generator is controlled based on the wind speed, the rotation speed of the wind turbine, and the amount of power generated by the generator. The wind turbine control program described in any one.

(Appendix 8) The computer
The load torque of the generator using the wind turbine is controlled based on the wind speed and the rotation speed of the wind turbine.
The maximum point is the operating point on the characteristic curve showing the relationship between the rotation speed of the wind turbine and the wind turbine torque of the wind turbine, with the wind speed and the rotation speed of the wind turbine as the observed values and the amount of power generated by the generator as the reward. The agent is made to execute reinforcement learning with the action of whether or not to move to the area on the side where the rotation speed of the wind turbine is smaller than that of the wind turbine.
When the agent to which the wind speed and the rotation speed of the wind turbine are input outputs to move the operating point to the region, the operating point is moved to the region by controlling the load torque of the generator. Move to,
A wind turbine control method characterized by performing processing.

(Appendix 9) A control unit that controls the load torque of the generator using the wind turbine based on the wind speed and the rotation speed of the wind turbine.
The maximum point is the operating point on the characteristic curve showing the relationship between the rotation speed of the wind turbine and the wind turbine torque of the wind turbine, with the wind speed and the rotation speed of the wind turbine as the observed values and the amount of power generated by the generator as a reward. It has an agent that executes reinforcement learning with the action of whether or not to move the wind turbine to a region on the side where the rotation speed is smaller than that of the wind turbine.
The control unit controls the load torque of the generator when the agent to which the wind speed and the rotation speed of the wind turbine are input outputs to move the operating point to the region. A wind turbine control device characterized in that an operating point is moved to the region.

100 Windmill control device 101 Wind power generation system 110 Windmill 120 Generator 300 Bus 301 CPU
302 Memory 303 Communication I / F
304 Recording medium I / F
305 Recording medium 400 Most recently measured value information 500 Wind turbine torque characteristic information 600 Ratio information 700 Movement flag information 800 Observation history 900 Action value table 1000 Action history 1100 Storage unit 1101 Acquisition unit 1102 Load torque control unit 1103 Agent 1104 Output unit 1201 Measurement unit 1202 Brake 1203 Start / stop unit 1204 Power consumption destination 1211 Measurement value acquisition unit 1212 Left / right designation unit 1213 Left side maintenance unit 1214 Follow-up control unit 1215 Load torque setting unit 1221 Observation unit 1222 Reward function unit 1223 State update unit 1224 Action determination unit 1300, 1400 , 1500, 1600, 1700, 1800, 1900, 2010, 2020 Table 1301, 1311-1131, 1321 to 1323, 1411, 1412, 1421, 142, 1511-1513, 1701, 1702, 1801, 1802, 1901,2011, 2021 Curves 2101,102, 2201,202, 2301,302, 2401,402, 2501,2502, 2601,2602,2701,2702,2801,2802,2901 to 2903,3001 to 3003 records

Claims (9)

On the computer
The load torque of the generator using the wind turbine is controlled based on the wind speed and the rotation speed of the wind turbine.
The maximum point is the operating point on the characteristic curve showing the relationship between the rotation speed of the wind turbine and the wind turbine torque of the wind turbine, with the wind speed and the rotation speed of the wind turbine as the observed values and the amount of power generated by the generator as the reward. The agent is made to execute reinforcement learning with the action of whether or not to move to the area on the side where the rotation speed of the wind turbine is smaller than that of the wind turbine.
When the agent to which the wind speed and the rotation speed of the wind turbine are input outputs to move the operating point to the region, the operating point is moved to the region by controlling the load torque of the generator. Move to,
Execute the process,
The moving process is a wind turbine control program characterized in that the operating point is moved to the region by setting the load torque of the generator to a value larger than the wind turbine torque at the maximum point. On the computer
The load torque of the generator using the wind turbine is controlled based on the wind speed and the rotation speed of the wind turbine.
The maximum point is the operating point on the characteristic curve showing the relationship between the rotation speed of the wind turbine and the wind turbine torque of the wind turbine, with the wind speed and the rotation speed of the wind turbine as the observed values and the amount of power generated by the generator as the reward. The agent is made to execute reinforcement learning with the action of whether or not to move to the area on the side where the rotation speed of the wind turbine is smaller than that of the wind turbine.
When the agent to which the wind speed and the rotation speed of the wind turbine are input outputs to move the operating point to the region, the operating point is moved to the region by controlling the load torque of the generator. Move to,
Execute the process,
In the process to be controlled, when the current operating point on the characteristic curve is in the region, the load torque of the generator is set to a value larger than the wind turbine torque at the operating point and a value smaller than the wind turbine torque at the operating point. A wind turbine control program characterized in that the operating point is maintained in the region by alternately setting the operating points. The wind turbine control program according to claim 1 or 2, wherein the wind turbine does not have a function of controlling the pitch of the wind turbine. The process to be controlled according to claim 1 to 3 is characterized in that the load torque of the generator is controlled based on the wind speed, the rotation speed of the wind turbine, and the current operating point on the characteristic curve. The wind turbine control program described in any one . The wind turbine control program according to any one of claims 1 to 4, wherein the interval for controlling the load torque of the generator is shorter than the interval for outputting the action by the agent . The computer
The load torque of the generator using the wind turbine is controlled based on the wind speed and the rotation speed of the wind turbine.
The maximum point is the operating point on the characteristic curve showing the relationship between the rotation speed of the wind turbine and the wind turbine torque of the wind turbine, with the wind speed and the rotation speed of the wind turbine as the observed values and the amount of power generated by the generator as the reward. The agent is made to execute reinforcement learning with the action of whether or not to move to the area on the side where the rotation speed of the wind turbine is smaller than that of the wind turbine.
When the agent to which the wind speed and the rotation speed of the wind turbine are input outputs to move the operating point to the region, the operating point is moved to the region by controlling the load torque of the generator. Move to,
Execute the process,
The process of moving the wind turbine is characterized in that the operating point is moved to the region by setting the load torque of the generator to a value larger than the torque of the wind turbine at the maximum point. The computer
The load torque of the generator using the wind turbine is controlled based on the wind speed and the rotation speed of the wind turbine.
The maximum point is the operating point on the characteristic curve showing the relationship between the rotation speed of the wind turbine and the wind turbine torque of the wind turbine, with the wind speed and the rotation speed of the wind turbine as the observed values and the amount of power generated by the generator as the reward. The agent is made to execute reinforcement learning with the action of whether or not to move to the area on the side where the rotation speed of the wind turbine is smaller than that of the wind turbine.
When the agent to which the wind speed and the rotation speed of the wind turbine are input outputs to move the operating point to the region, the operating point is moved to the region by controlling the load torque of the generator. Move to,
Execute the process,
In the process to be controlled, when the current operating point on the characteristic curve is in the region, the load torque of the generator is set to a value larger than the wind turbine torque at the operating point and a value smaller than the wind turbine torque at the operating point. A wind turbine control method characterized in that the operating point is maintained in the region by alternately setting the operating points . A control unit that controls the load torque of the generator using the wind turbine based on the wind speed and the rotation speed of the wind turbine.
The maximum point is the operating point on the characteristic curve showing the relationship between the rotation speed of the wind turbine and the wind turbine torque of the wind turbine, with the wind speed and the rotation speed of the wind turbine as the observed values and the amount of power generated by the generator as a reward. It has an agent that executes reinforcement learning with the action of whether or not to move the wind turbine to a region on the side where the rotation speed is smaller than that of the wind turbine.
The control unit controls the load torque of the generator when the agent to which the wind speed and the rotation speed of the wind turbine are input outputs to move the operating point to the region. The operating point is moved to the region, and the movement is characterized in that the operating point is moved to the region by setting the load torque of the generator to a value larger than the wind turbine torque at the maximum point. Windmill control device. A control unit that controls the load torque of the generator using the wind turbine based on the wind speed and the rotation speed of the wind turbine.
The maximum point is the operating point on the characteristic curve showing the relationship between the rotation speed of the wind turbine and the wind turbine torque of the wind turbine, with the wind speed and the rotation speed of the wind turbine as the observed values and the amount of power generated by the generator as a reward. It has an agent that executes reinforcement learning with the action of whether or not to move the wind turbine to a region on the side where the rotation speed is smaller than that of the wind turbine.
The control unit controls the load torque of the generator when the agent to which the wind speed and the rotation speed of the wind turbine are input outputs to move the operating point to the region. When the operating point is moved to the region and the current operating point on the characteristic curve is in the region, the load torque of the generator is set to a value larger than the wind turbine torque of the operating point and the wind turbine torque of the operating point. A wind turbine control device, characterized in that the operating point is maintained in the region by alternately setting smaller values.

Images