Ventilation system, ventilation fan, learning device, and inference device

Abstract

To obtain a ventilation system that can perform ventilation even during rainfall by adjusting a ventilation air volume in accordance with a monitoring result of a rainfall situation.SOLUTION: A ventilation system 100 comprises: a fan 30 comprising an impeller for generating an air flow by rotation, and a motor 4 for rotating the impeller; a camera 15 for photographing a space through which the air flow passes; and an inverter uniform velocity regulator 19 that is a control device for controlling the fan 30. The inverter uniform velocity regulator 19 comprises: a sucked rainwater amount estimation unit 21 for estimating a sucked rainwater amount indicating the amount of rainwater sucked together with the air flow, from an image of the space through which the air flow passes, photographed by the camera 15; and a motor control unit 22 for rotating the motor 4 at a rotational velocity corresponding to the sucked rainwater amount.SELECTED DRAWING: Figure 1

Description

[Technical Field]

[0001] The present disclosure relates to a ventilation system that draws outdoor air into a room, a ventilation fan, a learning device, and an inference device. [Background technology]

[0002] When bringing outdoor air into a room for ventilation, it is necessary to take measures to prevent rain from being brought in along with the outdoor air when it rains.

[0003] Patent Document 1 discloses a method for controlling indoor humidity by using a rain sensor installed outdoors to detect when rain has started to fall and stopping the ventilation fan, and by using the rain sensor to detect when the rain has stopped and starting the ventilation fan. This method makes it possible to prevent the humidity inside a building where a ventilation fan is installed from rising during rainfall. [Prior art documents] [Patent documents]

[0004] [Patent Document 1] Patent No. 6383698 Summary of the Invention [Problem to be solved by the invention]

[0005] However, the ventilation fan control described in Patent Document 1 cannot ventilate while rain continues because the ventilation fan is turned off. In other words, the method disclosed in Patent Document 1 cannot adjust the ventilation air volume according to the monitoring results of the rainfall situation to ventilate during rainfall.

[0006] The present disclosure has been made in consideration of the above, and aims to provide a ventilation system that is capable of ventilating during rainfall by adjusting the ventilation air volume according to the results of monitoring the rainfall conditions. [Means for solving the problem]

[0007] To solve the above-mentioned problems and achieve the object, the present disclosure provides a ventilation system including a fan having an impeller that generates an airflow by rotation and a motor that rotates the impeller, a camera that captures an image of a space through which the airflow passes, and a control device that controls the fan. The control device includes an intake rainwater amount estimating unit that estimates an intake rainwater amount that indicates the amount of rainwater sucked in along with the airflow from an image of the space through which the airflow passes captured by the camera, and a motor control unit that rotates the motor at a rotational speed corresponding to the intake rainwater amount. [Effects of the Invention]

[0008] According to the present disclosure, it is possible to obtain an advantageous effect of a ventilation system that can adjust the ventilation air volume according to the monitoring results of the rainfall conditions and thus can provide ventilation during rainfall. [Brief explanation of the drawings]

[0009] [Figure 1] FIG. 1 is a diagram showing the configuration of a ventilation system according to a first embodiment. [Figure 2] FIG. 1 is a diagram showing an example of the installation position of a ventilation fan in a ventilation system according to a first embodiment. [Figure 3] FIG. 1 is a diagram showing a configuration of a ventilation fan according to a first embodiment. [Figure 4] Cross-sectional view of a ventilation fan according to the first embodiment [Figure 5] FIG. 1 is a diagram showing the configuration of an intake rainwater amount estimation unit of a ventilation system according to Embodiment 1. [Figure 6] Configuration diagram of a learning device for a ventilation system according to embodiment 1 [Figure 7] FIG. 1 is a diagram showing an example of a neural network of a learning device for a ventilation system according to a first embodiment. [Figure 8] Flowchart of learning process of the learning device of the ventilation system according to the first embodiment [Figure 9] Configuration diagram of an inference device for a ventilation system according to embodiment 1 [Figure 10] Flowchart for inference processing of the inference device of the ventilation system according to the first embodiment [Figure 11]FIG. 10 is a diagram showing the configuration of a ventilation fan according to a second embodiment. [Figure 12] Cross-sectional view of a ventilation fan according to a second embodiment [Figure 13] FIG. 10 is a diagram showing an installation state of a ventilation fan according to a second embodiment. [Figure 14] FIG. 1 is a diagram showing the hardware configuration of a ventilation system according to a first embodiment and a suction rainwater amount estimation unit of a ventilation fan according to a second embodiment. DETAILED DESCRIPTION OF THE INVENTION

[0010] The ventilation system, ventilation fan, learning device, and inference device according to the embodiments will be described in detail below with reference to the drawings. In each drawing, identical or corresponding parts are designated by the same reference numerals. Duplicate descriptions of these parts will be appropriately simplified or omitted. Furthermore, the size relationships between the components in each drawing may differ from those in reality.

[0011] Embodiment 1 FIG. 1 is a diagram showing the configuration of a ventilation system according to the first embodiment. FIG. 2 is a diagram showing an example of the installation position of a ventilation fan in the ventilation system according to the first embodiment. The ventilation system 100 includes a ventilation fan 1 having a fan 30, an inverter speed regulator 19 which is a control device that controls the fan 30, and a camera 15. The ventilation fan 1 and the camera 15 are each connected to the inverter speed regulator 19 by electric wires so as to be able to communicate with each other. Note that the ventilation fan 1 and the camera 15 may each be connected wirelessly as long as they can communicate with the inverter speed regulator 19.

[0012] FIG. 3 is a diagram showing the configuration of a ventilation fan according to embodiment 1. FIG. 4 is a cross-sectional view of the ventilation fan according to embodiment 1. Ventilation fan 1 includes a square main body frame 2 for fixing to a wall surface. Fan 30 includes an impeller 7 that forms an airflow by rotation, and a motor 4 that drives impeller 7. Impeller 7 has a plurality of blades 6. In the center of main body frame 2, an air tunnel 3 is formed, which serves as a ventilation path for the airflow formed by fan 30 using impeller 7. Air tunnel 3 is cylindrical.

[0013] The motor 4 is positioned so that the central axis of rotation of the motor shaft 5 is the same as the central axis of the wind tunnel 3. An impeller 7 is fixedly connected to the tip of the motor shaft 5. The motor 4 is fixed to a box-shaped motor fixing base 9 which has four motor fixing legs 8. The motor fixing legs 8 are fixed to the main body frame 2 at the four corners of the main body frame 2.

[0014] An installation hole 10 is drilled at the tip of each motor fixing leg 8. In addition, installation holes 11 are drilled at the four corners of the main body frame 2. The installation holes 10 and 11 are positioned so as to overlap each other.

[0015] The inverter speed regulator 19 has a built-in control board 12. The control board 12 is provided with a suction rainwater amount estimation unit 21 that estimates the amount of suction rainwater, which is the amount of rainwater sucked in together with the airflow generated by the fan 30, and a motor control unit 22 that drives the motor 4. The motor control unit 22 arbitrarily changes the frequency and voltage of the power source 20 supplied to the motor 4 to control the rotation speed of the motor 4.

[0016] A terminal storage box 14 containing a power connection terminal 13 is attached to the outside of the wind tunnel 3 of the main frame 2. A speed regulator 19 such as an inverter is connected to the power connection terminal 13. When power 20 is supplied to the motor 4 via the power connection terminal 13, the motor shaft 5 rotates, and the impeller 7 rotates. When the impeller 7 rotates, a pressure difference occurs between the front and back surfaces of the blades 6, forming an airflow in the wind tunnel 3.

[0017] 2, camera 15 is installed in any location where it can capture an image of the space through which the airflow generated by fan 30 passes, such as inside c1 of weather cover 16, inside c2 of hole 18 in wall 17, or on the side surface c3 of motor fixing leg 8. Image data of the image of the space through which the airflow generated by fan 30 passes, captured by camera 15, is sent to inverter speed regulator 19, which is connected to camera 15 by an electric wire.

[0018] The ventilation fan 1 is typically installed on a building wall 17 together with a vent member such as a weather cover 16. A rectangular wall hole 18 with sides slightly longer than the inner diameter of the cylindrical wind tunnel 3 of the ventilation fan 1 is drilled in the building wall 17. The motor 4 is fixed to the building's indoor wall 17 with bolts and nuts passing through mounting holes 11 formed in the main frame 2 and mounting holes 10 formed in the motor fixing legs 8, with the center of the wall hole 18 coinciding with the center of the wind tunnel 3 of the ventilation fan 1. The weather cover 16 is also installed on the outdoor side of the building wall 17 with bolts and nuts passing through mounting holes 161 at the four corners of the weather cover 16, with the center of the wall hole 18 coinciding with the center of the vent in the weather cover 16.

[0019] In the case of an intake-type ventilation fan 1, outdoor air is sent into the building through the weather cover 16, wall holes 18, and wind tunnel 3 of the ventilation fan 1, but if it is raining outside, raindrops are also sucked in along with the air. The diameter of the raindrops sucked in varies, and their sum total is the amount of suctioned rainwater, which varies greatly depending on external conditions such as the amount of precipitation outside and the outside wind. It is difficult to reduce the amount of suctioned rainwater to zero, but by adjusting the ventilation airflow from the ventilation fan 1, it is possible to reduce the amount of suctioned rainwater to a practically acceptable level while still providing ventilation.

[0020] 5 is a diagram showing the configuration of the suction rainwater volume estimation unit of the ventilation system according to Embodiment 1. The suction rainwater volume estimation unit 21 includes a learning device 212, a trained model storage unit 213, and an inference device 214. The learning device 212 learns the suction rainwater volume corresponding to an image of a space through which the airflow generated by the fan 30 passes, captured by the camera 15, and generates a trained model. The trained model storage unit 213 stores the trained model generated by the learning device 212. The inference device 214 uses the trained model stored in the trained model storage unit 213 to infer the suction rainwater volume corresponding to the image of the space through which the airflow generated by the fan 30 passes.

[0021] 6 is a configuration diagram of a learning device for a ventilation system according to Embodiment 1. The learning device 212 includes a learning data acquisition unit 2121 and a model generation unit 2122.

[0022] The learning data acquisition unit 2121 acquires, as learning data, an image of the space through which the airflow generated by the fan 30 passes, taken by the camera 15, and information on the amount of suctioned rainwater. The amount of suctioned rainwater can be obtained, for example, by providing a curved ventilation duct downstream of the fan 30, collecting all rainwater adhering to the ventilation duct, and measuring its volume. The amount of suctioned rainwater may also be calculated by extracting rainwater droplets captured in an image of the space through which the airflow generated by the fan 30 passes through using image analysis, and calculating the density of the rainwater droplets.

[0023] The model generation unit 2122 learns the amount of suctioned rainwater corresponding to the image of the space through which the airflow generated by the fan 30 passes, based on learning data created based on a combination of the image of the space through which the airflow generated by the fan 30 passes and information on the amount of suctioned rainwater output from the learning data acquisition unit 2121. In other words, the learning data is data that associates the image of the space through which the airflow generated by the fan 30 passes and the amount of suctioned rainwater with each other.

[0024] The learning algorithm used by the model generation unit 2122 can be a known algorithm such as supervised learning, reinforcement learning, etc. As an example, a case where a neural network is applied will be described.

[0025] The model generation unit 2122 uses so-called supervised learning, for example, in accordance with a neural network model, to learn the amount of rainwater to be suctioned that corresponds to an image of the space through which the airflow generated by the fan 30 passes. Here, supervised learning refers to a method of providing the learning device 212 with data pairs of inputs and labels as results, thereby learning the features of the learning data and inferring the results from the inputs.

[0026] A neural network consists of an input layer consisting of multiple neurons, an intermediate layer consisting of multiple neurons, and an output layer consisting of multiple neurons. The intermediate layer is also called a hidden layer. There may be one intermediate layer, or two or more intermediate layers.

[0027] FIG. 7 is a diagram illustrating an example of a neural network of the learning device for the ventilation system according to Embodiment 1. The neural network 70 shown in FIG. 7 has a three-layer structure including an input layer 71 composed of neurons X1, X2, and X3, a middle layer 72 composed of neurons Y1 and Y2, and an output layer 73 composed of neurons Z1, Z2, and Z3. When multiple inputs are input to the input layer 71, the neural network 70 multiplies the input values by weights w11, w12, w13, w14, w15, and w16 before inputting the result to the middle layer 72. The result is further multiplied by weights w21, w22, w23, w24, w25, and w26 before being output from the output layer 73. This output result varies depending on the values of the weights w11, w12, w13, w14, w15, and w16 and the values of the weights w21, w22, w23, w24, w25, and w26.

[0028] In this embodiment, the neural network 70 learns the amount of suctioned rainwater corresponding to the image of the space through which the airflow generated by the fan 30 passes, by so-called supervised learning, in accordance with learning data created based on a combination of an image of the space through which the airflow generated by the fan 30 passes and information on the amount of suctioned rainwater, acquired by the learning data acquisition unit 2121.

[0029] That is, the neural network 70 learns by inputting an image of the space through which the airflow generated by the fan 30 passes into the input layer 71 and adjusting the weights w11, w12, w13, w14, w15, w16 and the weights w21, w22, w23, w24, w25, w26 so that the result output from the output layer 73 approaches the amount of rainwater suction.

[0030] The model generation unit 2122 generates and outputs a trained model by performing the above-described learning.

[0031] The trained model storage unit 213 stores the trained model output from the model generation unit 2122.

[0032] Next, the learning process of the learning device 212 will be described with reference to Fig. 8. Fig. 8 is a flowchart showing the learning process of the learning device of the ventilation system according to the first embodiment.

[0033] In step S11, the learning data acquisition unit 2121 acquires learning data, which are an image of the space through which the airflow generated by the fan 30 passes and information on the amount of suctioned rainwater. Note that the image of the space through which the airflow generated by the fan 30 passes and the information on the amount of suctioned rainwater are acquired simultaneously, but it is sufficient if the image of the space through which the airflow generated by the fan 30 passes and the information on the amount of suctioned rainwater are input in association with each other, and the image of the space through which the airflow generated by the fan 30 passes and the information on the amount of suctioned rainwater may be acquired at different times.

[0034] In step S12, the model generation unit 2122 learns the amount of suctioned rainwater corresponding to the image of the space through which the airflow generated by the fan 30 passes, by so-called supervised learning, in accordance with the learning data created based on a combination of the image of the space through which the airflow generated by the fan 30 passes and information on the amount of suctioned rainwater acquired by the learning data acquisition unit 2121, and generates a learned model.

[0035] In step S13, the trained model storage unit 213 stores the trained model generated by the model generation unit 2122.

[0036] The learning device 212 repeats the above learning process by changing the combination of the image of the space through which the airflow generated by the fan 30 passes, which is the learning data, and the information on the amount of suctioned rainwater, to learn the amount of suctioned rainwater corresponding to the image of the space through which the airflow generated by the fan 30 passes. In other words, the learning device 212 learns the amount of suctioned rainwater corresponding to each of multiple images of the space through which the airflow generated by the fan 30 passes, taken at preset intervals.

[0037] 9 is a configuration diagram of the inference device of the ventilation system according to Embodiment 1. The inference device 214 includes an inference data acquisition unit 2141 and an inference unit 2142.

[0038] The inference data acquisition unit 2141 acquires the amount of suctioned rainwater, which is inference data.

[0039] The inference unit 2142 infers the amount of suctioned rainwater corresponding to the image of the space through which the airflow generated by the fan 30 passes, obtained by using the trained model. That is, by inputting the image of the space through which the airflow generated by the fan 30 passes, obtained by the inference data acquisition unit 2141, into this trained model, it is possible to output the amount of suctioned rainwater corresponding to the image of the space through which the airflow generated by the fan 30 passes.

[0040] In this embodiment, it has been described that the amount of suctioned rainwater corresponding to an image of the space through which the airflow generated by the fan 30 passes is output using a trained model trained by the model generation unit 2122 of the ventilation fan 1, but it is also possible to obtain a trained model from an external source such as another ventilation fan, and output the amount of suctioned rainwater corresponding to an image of the space through which the airflow generated by the fan 30 passes based on this trained model.

[0041] Next, a process for obtaining the amount of suctioned rainwater corresponding to an image of the space through which the airflow generated by the fan 30 passes using the inference device 214 will be described with reference to Fig. 10. Fig. 10 is a flowchart showing the inference process of the inference device of the ventilation system according to the first embodiment.

[0042] In step S21, the inference data acquisition unit 2141 acquires an image of the space through which the airflow generated by the fan 30 passes, which is inference data.

[0043] In step S22, the inference unit 2142 acquires a trained model from the trained model storage unit 213, inputs an image of the space through which the airflow generated by the fan 30 passes into the acquired trained model, and obtains the amount of suctioned rainwater corresponding to the image of the space through which the airflow generated by the fan 30 passes. In this embodiment, information is obtained indicating whether the amount of suctioned rainwater falls within a first range below a preset lower threshold, a second range equal to or greater than the lower threshold and equal to or less than a preset upper threshold, or a third range exceeding the upper threshold.

[0044] In step S23, the inference unit 2142 outputs information indicating whether the amount of suctioned rainwater obtained by the trained model falls within the first range, the second range, or the third range to the motor control unit 22 of the control board 12. That is, the inference unit 2142 outputs information indicating the range within which the amount of suctioned rainwater falls to the motor control unit 22 of the control board 12.

[0045] In step S24, based on the information output from the inference unit 2142, the motor control unit 22 controls the rotation speed of the motor so that the ventilation airflow rate corresponds to the monitoring result of the rainfall condition. For example, when the amount of suctioned rainwater falls within a first range, the motor control unit 22 increases the rotation speed of the motor 4 by a certain amount. When the amount of suctioned rainwater falls within a second range, the motor control unit 22 maintains the rotation speed of the motor 4. When the amount of suctioned rainwater falls within a third range, the motor control unit 22 decreases the rotation speed of the motor 4 by a certain amount.

[0046] By repeating the series of operations from step S21 to step S24 described above, if the ventilation fan 1 is operating during rainfall with the amount of suctioned rainwater exceeding the upper threshold, the airflow rate can be gradually reduced to reduce the amount of suctioned rainwater to below the upper threshold and above the lower threshold. If the amount of rainfall outdoors subsequently decreases or the rain stops, the amount of suctioned rainwater will be output below the lower threshold, and the airflow rate will gradually increase with each step from step S21 to step S24. However, the upper limit of the rotational speed of the motor 4 is set in advance, and the rotational speed will not exceed that limit.

[0047] Here, the inference unit 2142 is assumed to output information indicating whether the amount of suctioned rainwater obtained by the trained model falls within the first, second, or third range, but the inference unit 2142 may obtain information indicating the amount of suctioned rainwater itself by inputting an image of the space through which the airflow generated by the fan 30 passes into the trained model, and output this information to the motor control unit 22. In this case, the motor control unit 22 may increase the rotation speed of the motor 4 by a certain amount if the amount of suctioned rainwater is below a preset lower threshold, maintain the rotation speed of the motor 4 if the amount of suctioned rainwater is equal to or greater than the lower threshold and equal to or less than a preset upper threshold, and decrease the rotation speed of the motor 4 by a certain amount if the amount of suctioned rainwater exceeds the upper threshold.

[0048] The model generation unit 2122 may also learn the amount of suction rainwater corresponding to an image of a space through which the airflow generated by the fan 30 passes, based on learning data created for multiple ventilation fans 1. The model generation unit 2122 may acquire learning data from multiple ventilation fans 1 used in the same area, or may learn the amount of suction rainwater corresponding to an image of a space through which the airflow generated by the fan 30 passes, using learning data collected from multiple ventilation fans 1 operating independently in different areas. It is also possible to add or remove ventilation fans 1 from which learning data is collected during the process. Furthermore, the learning device 212 that has learned the amount of suction rainwater corresponding to an image of a space through which the airflow generated by the fan 30 passes for a certain ventilation fan 1 may be applied to another ventilation fan 1, and the amount of suction rainwater corresponding to an image of a space through which the airflow generated by the fan 30 passes for the other ventilation fan 1 may be re-learned and updated.

[0049] Furthermore, the learning algorithm used in the model generation unit 2122 can be deep learning, which learns to extract the features themselves, or machine learning can be performed according to other known methods, such as genetic programming, functional logic programming, or support vector machines.

[0050] The learning device 212 and the inference device 214 are used by the ventilation fan 1 to learn the amount of rainwater to be sucked in corresponding to an image of the space through which the airflow generated by the fan 30 passes, but may be separate devices connected to the ventilation fan 1 via a network, for example. The learning device 212 and the inference device 214 may also reside on a cloud server.

[0051] The ventilation system 100 according to embodiment 1 determines the rotation speed of the motor 4 using a trained model generated by learning the amount of rainwater suction corresponding to an image taken by the camera 15 of the space through which the airflow generated by the fan 30 passes, and therefore can adjust the ventilation air volume according to the monitoring results of the rainfall conditions.

[0052] The ventilation system 100 according to the first embodiment does not require an increase in the number of blades (not shown) installed on the vent member to prevent rainwater intrusion, and therefore can prevent rainwater intrusion without increasing pressure loss. While increasing the size of the vent member to expand the air passage area would increase the number of blades (not shown) installed on the vent member and suppress the increase in pressure loss, the vent member is installed on the exterior wall of the building, and therefore increasing the size of the vent member would result in a decrease in aesthetic appearance and ease of installation. The ventilation system 100 according to the first embodiment adjusts the ventilation airflow by controlling the rotation speed of the motor 4 according to the results of rainfall monitoring, and therefore does not require an increase in the size of the weather cover 16, which is the vent member, and therefore does not result in a decrease in aesthetic appearance or ease of installation.

[0053] Embodiment 2 Fig. 11 is a diagram showing the configuration of a ventilation fan according to embodiment 2. Fig. 12 is a cross-sectional view of the ventilation fan according to embodiment 2. Fig. 13 is a diagram showing the installation state of the ventilation fan according to embodiment 2. The ventilation fan 1 according to embodiment 2 is generally similar to the ventilation fan 1 according to embodiment 1, so only the differences will be described and a description of the common parts will be omitted.

[0054] Camera 15 is attached to the outside of wind tunnel 3 of main body frame 2. A hole is formed in main body frame 2 that penetrates cylindrical wind tunnel 3 in a direction perpendicular to the axial direction, and camera 15 is able to photograph the inside of wind tunnel 3 through the hole in main body frame 2. Camera 15 photographs, through the hole in main body frame 2, the space through which the airflow generated by fan 30 flows in wind tunnel 3. Image data of the image photographed by camera 15 is sent to control board 12, which is connected to camera 15 by an electric wire.

[0055] The terminal storage box 14 houses a control board 12 in addition to a power connection terminal 13. An external power supply (not shown) is connected to the power connection terminal 13. The control board 12 includes an suction rainwater amount estimation unit 21 that estimates the amount of suction rainwater, which is the amount of rainwater suctioned together with the airflow generated by the fan 30, and a motor control unit 22 that controls the rotation speed of the motor 4 by arbitrarily changing the frequency and voltage of the power supplied to the motor 4. The suction rainwater amount estimation unit 21 is similar to the suction rainwater amount estimation unit 21 of the ventilation system 100 according to the first embodiment shown in FIG. 5 , and includes a learning device 212, a trained model storage unit 213, and an inference device 214.

[0056] Control board 12 uses suction rainwater amount estimation unit 21 to estimate the amount of suction rainwater corresponding to the image of the space through which the airflow generated by fan 30 passes, taken by camera 15. Based on the amount of suction rainwater estimated by suction rainwater amount estimation unit 21, motor control unit 22 supplies power supply 20 with the frequency, voltage, etc. changed to motor 4 of ventilation fan 1, and controls the rotation speed of motor 4 to adjust the ventilation air volume generated by ventilation fan 1.

[0057] The ventilation fan 1 according to the second embodiment has the suction rainwater amount estimation unit 21, which includes the learning device 212, the learned model storage unit 213, and the inference device 214, provided on the control board 12, and therefore can infer the amount of suction rainwater using a learned model generated by learning the amount of suction rainwater corresponding to an image of the space through which the airflow generated by the fan 30 passes. Therefore, the ventilation fan 1 according to the second embodiment can adjust the ventilation airflow in accordance with the monitoring results of the rainfall conditions.

[0058] The hardware configuration of the ventilation system 100 according to the first embodiment and the suction rainwater amount estimating unit 21 of the ventilation fan 1 according to the second embodiment will be described.

[0059] Fig. 14 is a diagram showing the hardware configuration of the suction rainwater volume estimating unit of the ventilation system according to Embodiment 1 and the ventilation fan according to Embodiment 2. The suction rainwater volume estimating unit 21 is realized by a computer system including a processor 91 that executes various processes, a memory 92 that is a main memory, and a storage device 93 that stores information.

[0060] The processor 91 may be a computing means such as an arithmetic unit, a microprocessor, a microcomputer, a CPU (Central Processing Unit), or a DSP (Digital Signal Processor). The memory 92 may be a non-volatile or volatile semiconductor memory such as a RAM (Random Access Memory), a ROM (Read Only Memory), a flash memory, an EPROM (Erasable Programmable Read Only Memory), or an EEPROM (Electrically Erasable Programmable Read Only Memory). The storage device 93 stores programs for learning the set value of the rotation speed of the motor 4 and for inferring the set value of the rotation speed of the motor 4.

[0061] The computer system described above realizes the functions of the suction rainwater volume estimation unit 21 by having the processor 91 read into the memory 92 programs stored in the storage device 93 and execute the programs corresponding to the processing of each component. The memory 92 is also used as a temporary memory for each processing executed by the processor 91. The programs executed by the processor 91 may be provided in a state stored in a storage medium or provided via a network.

[0062] The configurations shown in the above embodiments are merely examples of the content, and may be combined with other known technologies, or parts of the configurations may be omitted or modified without departing from the spirit of the invention. [Explanation of symbols]

[0063] 1 ventilation fan, 2 main body frame, 3 wind tunnel, 4 motor, 5 motor shaft, 6 blades, 7 impeller, 8 motor fixing leg, 9 motor fixing base, 10, 11, 161 installation hole, 12 control board, 13 power connection terminal, 14 terminal storage box, 15 camera, 16 weather cover, 17 wall, 18 wall hole, 19 inverter etc. speed regulator, 20 power supply, 21 suction rainwater amount estimation unit, 22 motor control unit, 30 fan, 70 neural network, 71 input layer, 72 intermediate layer, 73 output layer, 91 processor, 92 memory, 93 storage device, 100 ventilation system, 212 learning device, 213 learned model storage unit, 214 inference device, 2121 learning data acquisition unit, 2122 model generation unit, 2141 inference data acquisition unit, 2142 inference unit.

Claims

1. a fan having an impeller that generates an air flow by rotation and a motor that rotates the impeller; a camera that photographs a space through which the airflow passes; A ventilation system comprising a control device that controls the fan, The control device an intake rainwater amount estimation unit that estimates an intake rainwater amount indicating an amount of rainwater that has been intaken together with the airflow from an image of a space through which the airflow passes, the image being taken by the camera; A ventilation system comprising: a motor control unit that rotates the motor at a rotational speed corresponding to the amount of suctioned rainwater.

2. the suction rainwater amount estimation unit outputs information indicating whether the suction rainwater amount is within a first range below a preset lower limit threshold, a second range equal to or greater than the lower limit threshold and equal to or less than a preset upper limit threshold, or a third range exceeding the upper limit threshold, to the motor control unit; The ventilation system described in claim 1, characterized in that the motor control unit increases the rotation speed of the motor by a certain amount when the amount of suctioned rainwater is within the first range, maintains the rotation speed of the motor when the amount of suctioned rainwater is within the second range, and decreases the rotation speed of the motor by a certain amount when the amount of suctioned rainwater is within the third range.

3. the suction rainwater amount estimation unit outputs information about the suction rainwater amount to the motor control unit; The ventilation system described in claim 1, characterized in that the motor control unit increases the rotation speed of the motor by a certain amount when the amount of suctioned rainwater is less than a predetermined lower threshold, maintains the rotation speed of the motor by a certain amount when the amount of suctioned rainwater is equal to or greater than the lower threshold and equal to or less than a predetermined upper threshold, and decreases the rotation speed of the motor by a certain amount when the amount of suctioned rainwater exceeds the upper threshold.

4. The ventilation system according to any one of claims 1 to 3, characterized in that the suction rainwater volume estimation unit estimates the suction rainwater volume based on a plurality of images of the space through which the air flow passes, taken at predetermined intervals.

5. a fan having an impeller that generates an air flow by rotation and a motor that rotates the impeller; a camera that photographs a space through which the airflow passes; an intake rainwater amount estimation unit that estimates an intake rainwater amount indicating an amount of rainwater that has been intaken together with the airflow from an image of a space through which the airflow passes, the image being taken by the camera; a motor control unit that rotates the motor at a rotation speed corresponding to the amount of suctioned rainwater.

6. a learning data acquisition unit that acquires learning data including an image of a space through which an airflow generated by the fan due to rotation of the impeller passes and information on the amount of suctioned rainwater; A learning device characterized by comprising a model generation unit that uses the learning data to generate a trained model for inferring the amount of suctioned rainwater corresponding to an image of a space through which the airflow generated by the fan passes.

7. The learning device according to claim 6, wherein the information on the amount of suctioned rainwater is a result of calculating the density of rainwater droplets captured on an image of the space through which the air flow passes.

8. An inference device for inferring an amount of suctioned rainwater, which is an amount of rainwater sucked together with an airflow generated by a fan, an inference data acquisition unit that acquires an image of a space through which the airflow generated by the fan passes; An inference device characterized by comprising an inference unit that outputs the amount of suctioned rainwater corresponding to an image of the space through which the airflow generated by the fan passes, acquired by the inference data acquisition unit.

9. The inference device according to claim 8, characterized in that the inference unit infers the amount of suctioned rainwater corresponding to the image of the space through which the airflow generated by the fan passes, using a trained model for inferring the amount of suctioned rainwater corresponding to the image of the space through which the airflow generated by the fan passes.

Images