Performance evaluation system, pollutant removal device, information processing device

Abstract

Predict the appropriate time to replace the adsorbent. [Solution] A performance evaluation system 100 in which a pollutant removal device 40 that removes pollutants from the air and an information processing device 60 can communicate via a network, wherein the pollutant removal device includes an adsorbent that adsorbs the pollutants, an imaging means 50 that images the adsorbent, and a communication device 51 that transmits image data of the adsorbent captured by the imaging means to the information processing device, and the control unit inputs the feature amounts into correspondence information that associates the feature amounts with the performance of the adsorbent to infer the performance of the adsorbent, and predicts when to replace the adsorbent based on the relationship between the elapsed time since use of the adsorbent began and the inferred performance of the adsorbent.

Description

[Technical Field]

[0001] The present invention relates to a performance evaluation system, a pollutant removal device, and an information processing device. [Background technology]

[0002] There are spaces where airborne pollutants (chemical substances, airborne particles, airborne microorganisms, NOx, etc.) must be kept below a specified cleanliness level. For example, in clean rooms, art galleries, museums, etc., the air is supplied to the room after various pollutants that may be present in the air are removed. Pollutant removal devices that remove pollutants from the air in this way are known. Pollutant removal devices remove pollutants by adsorbing them onto a built-in adsorbent. The performance of the adsorbent decreases as more pollutants are adsorbed.

[0003] A technique for diagnosing a decline in the performance of a pollutant removal device is known (see, for example, Patent Document 1). Patent Document 1 discloses a technique for evaluating the performance of an ion exchange resin under test from an image of the ion exchange resin under test, using a trained model that has been machine-learned based on training images of the ion exchange resin taken for training purposes and an evaluation result of the appearance of the ion exchange resin. [Prior art documents] [Patent documents]

[0004] [Patent Document 1] International Publication No. 2022 / 080416 Summary of the Invention [Problem to be solved by the invention]

[0005] However, with conventional technology, it has been difficult to accurately predict when to replace the adsorbent. For example, if the timing of adsorbent replacement is delayed, air containing a large amount of pollutants may be supplied to the room. On the other hand, if the adsorbent is replaced too early, excessive replacement may occur.

[0006] In view of the above-mentioned problems, the present disclosure provides a technique for predicting an appropriate time to replace an adsorbent. [Means for solving the problem]

[0007] A first aspect of the present disclosure is A performance evaluation system in which a pollutant removal device that removes pollutants from the air and an information processing device can communicate with each other via a network, The pollutant removal device includes: an adsorbent that adsorbs the contaminants; an imaging means for imaging the adsorbent; a communication device that transmits image data of the adsorbent captured by the imaging means to the information processing device, the information processing device has a control unit, the control unit analyzes the image data to acquire a feature amount related to the performance of the adsorbent; Inferring the performance of the adsorbent by inputting the feature into correspondence information that associates the feature with the performance of the adsorbent; The timing for replacing the adsorbent is predicted based on the relationship between the time elapsed since the start of use of the adsorbent and the inferred performance of the adsorbent.

[0008] According to the first aspect of the present disclosure, it is possible to predict the appropriate time to replace the adsorbent.

[0009] A second aspect of the present disclosure is a performance evaluation system according to the first aspect, The feature quantity related to the performance of the adsorbent is a feature quantity that changes depending on the amount of salt precipitated on the adsorbent.

[0010] A third aspect of the present disclosure is a performance evaluation system according to the first or second aspect, the characteristic amount is a change in color of the adsorbent, The control unit inputs the amount of change in color into the correspondence information to infer the performance of the adsorbent.

[0011] A fourth aspect of the present disclosure is a performance evaluation system according to the first or second aspect, the characteristic amount is a change in mass or volume of the adsorbent, The control unit inputs the amount of change in mass or volume of the adsorbent into the correspondence information to infer the performance of the adsorbent.

[0012] A fifth aspect of the present disclosure is a performance evaluation system according to the first or second aspect, the feature amount is a change amount of particle size information of the adsorbent, The control unit inputs the amount of change in the particle size information into the correspondence information to infer the performance of the adsorbent.

[0013] A sixth aspect of the present disclosure is a performance evaluation system according to any one of the first to fifth aspects, The control unit notifies the terminal device of the predicted replacement time.

[0014] A seventh aspect of the present disclosure is a performance evaluation system according to any one of the first to sixth aspects, The correspondence information is a first model obtained by machine learning the correspondence between the feature amount and the performance of the adsorbent.

[0015] An eighth aspect of the present disclosure is a performance evaluation system according to the seventh aspect, the contaminant removal device has the adsorbent disposed at a first location and a second location; the communication device transmits the environmental data surrounding the pollutant removal device to the information processing device; the control unit calculates a second model that has learned a correspondence between the performance of the adsorbent at the first placement location and the surrounding environment data, and the performance of the adsorbent at the second placement location, The performance of the adsorbent at the first location estimated by the first model and ambient environmental data of the pollutant removal device are input to estimate the performance of the adsorbent at the second location.

[0016] A ninth aspect of the present disclosure is a performance evaluation system according to the eighth aspect, the control unit weights the performance of the adsorbent at the first placement location estimated by the first model by an air flow rate or a deterioration rate of the adsorbent; weighting the performance of the adsorbent at the second location by the air flow rate or the degradation rate of the adsorbent; The weighted performance of the adsorbent at the first location and the performance of the adsorbent at the second location are combined to calculate the overall performance of the adsorbent for the pollutant removal device.

[0017] A tenth aspect of the present disclosure is a performance evaluation system according to any one of the first to ninth aspects, The imaging means has sensitivity to a wavelength of light that captures an image of the salt precipitated on the adsorbent.

[0018] An eleventh aspect of the present disclosure is A pollutant removal device for removing pollutants from air, comprising: an adsorbent that adsorbs the contaminants; imaging means for imaging the adsorbent; a control unit; the control unit analyzes image data of the adsorbent captured by the imaging means to acquire feature quantities related to the performance of the adsorbent; Inferring the performance of the adsorbent by inputting the feature into correspondence information that associates the feature with the performance of the adsorbent; The timing for replacing the adsorbent is predicted based on the relationship between the time elapsed since the start of use of the adsorbent and the inferred performance of the adsorbent.

[0019] According to the eleventh aspect of the present disclosure, it is possible to predict the appropriate time to replace the adsorbent.

[0020] A twelfth aspect of the present disclosure is an adsorbent that adsorbs contaminants; imaging means for imaging the adsorbent; and a communication device that transmits image data of the adsorbent captured by the imaging means to an information processing device, the information processing device being capable of communicating with a pollutant removal device that removes pollutants in the air via a network, A control unit is provided. the control unit analyzes the image data to acquire a feature amount related to the performance of the adsorbent; Inferring the performance of the adsorbent by inputting the feature into correspondence information that associates the feature with the performance of the adsorbent; The timing for replacing the adsorbent is predicted based on the relationship between the time elapsed since the start of use of the adsorbent and the inferred performance of the adsorbent.

[0021] According to the twelfth aspect of the present disclosure, it is possible to predict the appropriate time to replace the adsorbent. [Brief explanation of the drawings]

[0022] [Figure 1] FIG. 1 illustrates an example of a system configuration of a performance evaluation system. [Figure 2] 1 is an external view of an example of a pollutant removal device. [Figure 3] FIG. 1 is a side view of an example of a contaminant removal device. [Figure 4] FIG. 2 is a diagram illustrating a hardware configuration of an example of an edge device. [Figure 5] FIG. 2 is a diagram illustrating a hardware configuration of an example of a server device. [Figure 6] FIG. 2 is a functional block diagram of an example illustrating functions of a server device in a performance evaluation system, divided into blocks. [Figure 7] FIG. 2 is a functional block diagram of a learning device according to an embodiment. [Figure 8] FIG. 2 is a functional block diagram of an analysis unit according to an embodiment. [Figure 9] FIG. 10 is a diagram showing an example of the relationship between elapsed time, image data, and contaminant removal efficiency, which is experimentally determined. [Figure 10] FIG. 10 is a scatter plot illustrating an example of the relationship between the number of white pixels and the efficiency of removing contaminants. [Figure 11]FIG. 10 is a diagram schematically illustrating the relationship between the change in the number of white pixels and the ammonia removal efficiency. [Figure 12] FIG. 10 is a flowchart illustrating an example of a process in which a feature amount detection unit digitizes the number of white pixels. [Figure 13] FIG. 1 is an image diagram illustrating the increase in mass and volume due to the adsorption of pollutants. [Figure 14] 10A and 10B are diagrams illustrating an example of image data and particle size information of an adsorbent. [Figure 15] FIG. 10 is a flowchart illustrating an example of a process for detecting granularity information. [Figure 16] FIG. 10 is a diagram schematically showing the relationship between the amount of change in mass or volume and the removal efficiency. [Figure 17] FIG. 1 is a diagram illustrating an example of a performance evaluation model configured using a neural network. [Figure 18] FIG. 1 is a diagram illustrating an example of the configuration of a performance evaluation model using CNN. [Figure 19] 10 is a graph showing an example of the relationship between elapsed time and removal efficiency. [Figure 20] FIG. 10 is a sequence diagram illustrating an example of a process flow in which the performance evaluation system predicts the replacement time of the adsorption filter. [Figure 21] 10A and 10B are diagrams illustrating examples of the arrangement of a temperature sensor, a humidity sensor, and an air flow sensor. [Figure 22] FIG. 2 is a functional block diagram of an example illustrating functions of a performance evaluation system, divided into blocks. [Figure 23] FIG. 2 is a functional block diagram of a learning device according to an embodiment. [Figure 24] FIG. 2 is a functional block diagram of an analysis unit according to an embodiment. [Figure 25] FIG. 10 is a diagram showing an example of a second-stage model configured by a neural network. [Figure 26] FIG. 10 is a graph showing the removal efficiency curve of an example adsorptive filter versus time. [Figure 27] FIG. 10 is a sequence diagram illustrating an example of a process flow in which the performance evaluation system predicts the replacement time of the adsorption filter. DETAILED DESCRIPTION OF THE INVENTION

[0023] A performance evaluation system and an evaluation method performed by the performance evaluation system will be described below as an example of an embodiment of the present disclosure.

[0024] <General adsorbent performance evaluation> Pollutant removal devices remove pollutants by adsorbing them onto a built-in adsorbent. The performance of the adsorbent is evaluated by a service technician collecting samples of the adsorbent on-site and requesting a performance evaluation from the manufacturer. Performance, for example, refers to the efficiency of pollutant removal. This method requires costs for collecting samples on-site and for the manufacturer to evaluate the performance of the adsorbent.

[0025] Furthermore, since there is a relatively long time between the previous and current performance evaluations, it is difficult to constantly monitor the performance of the adsorbent. If the performance deteriorates between the previous and current performance evaluations, there is a risk that air containing the contaminants to be removed will be supplied into the room.

[0026] Furthermore, even if the evaluation result shows that the performance has not deteriorated enough to warrant replacement, replacing the adsorbent is likely to result in excessive replacement because it is expected that the performance will further deteriorate until the next performance evaluation. Conversely, if it is expected that the performance will not deteriorate until the next performance evaluation, there is a risk that replacement will be delayed.

[0027] [First embodiment] Therefore, in the performance evaluation system of the present disclosure, a camera (imaging means) is installed in the pollutant removal device, and the camera constantly transmits image data to a server device. The server device evaluates the performance of the adsorbent by analyzing the image data. The following is a flow of work or operations related to the performance evaluation system. Note that the following explanation uses ammonia as an example of a pollutant. Ammonia exists in the air as a gas (vapor). The adsorbent may also be referred to as a gas adsorbent.

[0028] (1) When adsorbents adsorb ammonia, they precipitate ammonium salts (hereinafter simply referred to as salts). When salts precipitate, the surface of the adsorbent turns white (changes color) and its mass increases. It is known that the more ammonia adsorbed, the worse the adsorbent's performance becomes.

[0029] (2) A camera is installed inside the contaminant removal device to capture images of the adsorbent. The camera repeatedly captures images of the adsorbent at a preset frequency. The image data of the adsorbent is sent to a server device.

[0030] (3) The server device analyzes the image data and quantifies the color and particle size of the adsorbent as feature quantities. When ammonia is adsorbed onto the adsorbent, salt precipitates, turning white over time. The amount of white correlates with the amount of salt precipitated. Similarly, when ammonia is adsorbed onto the adsorbent, the mass and volume of the adsorbent increase. The increase in mass or volume is reflected in an increase in particle size information (diameter, surface area, etc.). Therefore, the server device can estimate the performance of the adsorbent based on the color, mass, or volume of the adsorbent, or particle size information.

[0031] (4) The server device uses a performance evaluation model (an example of a first model) that has learned the correspondence between features and performance to evaluate the performance of the adsorbent each time image data is transmitted. Since the performance evaluation is obtained in chronological order, the server device can estimate the replacement time when the performance is expected to fall below a certain level.

[0032] (5) The server device predicts when it is time to replace the adsorbent and automatically notifies the person in charge via email or other means.

[0033] In this way, the performance evaluation system of the present disclosure (i) Since the server device can evaluate the performance of the adsorbent from a remote location, it is possible to predict the time to replace the adsorbent without having a serviceman visit the site, thereby reducing costs. (ii) The server device can constantly evaluate the performance of the adsorbent, making it easier to determine the appropriate time to replace it. (iii) The server device can predict the replacement time, which can prevent excessive replacement and delays in replacement.

[0034] <Terminology> Air pollutants are substances not found in pure air. Pollutants include chemicals, airborne particles, airborne microorganisms, and NOx. Specific pollutants include nitrogen monoxide, sulfur oxides, nitrogen dioxide, sulfur compounds, methyl mercaptan, alkaline gases such as ammonia or amines, acidic gases such as hydrochloric acid, nitric acid, sulfuric acid, or acetic acid, ozone, chlorine, organic solvents, and organic gases.

[0035] Adsorbents are substances that have the ability to attract and store pollutants. The adsorbent may vary depending on the pollutant to be adsorbed. Examples of adsorbents include potassium permanganate, phosphoric acid, calcium hydroxide, potassium carbonate, manganese dioxide, activated alumina, activated carbon, zeolite, and activated carbon.

[0036] The feature quantity related to the performance of the adsorbent is a feature quantity obtained by analyzing image data and serves as an index of the performance of the adsorbent. In the present disclosure, the feature quantity related to the performance of the adsorbent is, for example, a feature quantity correlated with the amount of precipitated salt. For example, the feature quantity related to the performance of the adsorbent may be a change in color, a change in particle size, or a change in volume. Depending on the combination of the adsorbent and the contaminant, substances other than salt may precipitate, and a feature quantity corresponding to the precipitated substance may be adopted.

[0037] <System configuration of performance evaluation system> The system configuration of the performance evaluation system 100 will be described with reference to Fig. 1. Fig. 1 is a diagram showing an example of the system configuration of the performance evaluation system 100.

[0038] The performance evaluation system 100 provides a variety of services utilizing IoT to everyone from administrators to general users by having various devices 30 and a cloud-side server device 60 communicate via a network N. The edge devices 10, devices 30, sensor switches 20, and user terminals 5 are mainly located on the customer side, and the server device 60 is located in a cloud such as a data center or the Internet.

[0039] The device 30 refers to any device that consumes power, such as a pollutant removal device 40, an air conditioner, security equipment, heat source equipment, fire alarms, AHUs (air handling units), watt-hour meters, and lighting. The device 30 may also be other devices. The sensor switches 20 are various sensors, lamps, relays, and the like. The sensor switches 20 may include a camera 50, or the device 30 may have a built-in camera 50. The device 30 and the sensor switches 20 are communicatively connected to the edge device 10 via a dedicated cable or a network such as a LAN. The device 30 and the sensor switches 20 may also be communicatively connected to the edge device 10 via wireless communication.

[0040] The pollutant removal device 40 is a device that removes pollutants from the air. What constitutes a pollutant may vary depending on the intended use of the space. The pollutant removal device 40 may be installed in, for example, buildings, commercial facilities, art galleries, museums, food factories, pharmaceutical factories, electronic factories, power generation facilities, nuclear disaster response facilities, etc., but is not limited to these.

[0041] The device 30 and the sensor switches 20 are controlled by the edge device 10. In other words, the edge device 10 applies the required operations to the device 30 and the sensor switches 20 to suit the purposes of the device 30 and the sensor switches 20. The content of the control varies depending on the type of device 30 and the sensor switches 20. For example, if the device 30 is an air conditioner, the control may include all control of the functions of the air conditioner, such as the cooling / heating mode, set temperature, air volume, humidity, and air direction that can generally be set on an air conditioner. The control may also include operating modes such as a pre-season inspection mode, microcomputer reset, operation shutdown, and function substitution. Furthermore, if the device 30 is a pollutant removal device 40, the control may include control of the timing of image capture on the camera 50.

[0042] The device 30 collects operating data appropriate for the device 30 and transmits it mainly periodically to the edge device 10. Periodically means, for example, once per minute, once per 10 minutes, once per 60 minutes, etc., but this can be set by the user or the server device 60. Furthermore, the device 30 can transmit operating data to the edge device 10 in response to a request from the edge device 10 or the user terminal 5. The operating data varies depending on the device 30, but in the case of an air conditioner, for example, the operating data may include the high pressure and low pressure of the refrigerant, the refrigerant temperature, the fan rotation speed, and the CPU temperature of the microcomputer.

[0043] Furthermore, if the device 30 detects an abnormality, it transmits an abnormality code to the edge device 10. The device 30 that detects an abnormality stops operation. The edge device 10 transmits the abnormality code to the server device 60. The processing of the edge device 10 may be similar for the sensor switches 20. The sensor switches 20 mainly periodically transmit their own information and abnormality codes to the edge device 10. The camera 50 transmits image data to the edge device 10. The timing of transmission may be periodically, at a set date and time, during a time period with low load, when the difference from the last image data is equal to or greater than a certain value, etc.

[0044] The edge device 10 is a controller that controls the device 30 and the sensor switches 20. The edge device 10 functions as a control device that controls the device 30 and the sensor switches 20, an information processing device that processes operation data, etc., and a communication means that communicates with the server device 60. For example, the edge device 10 transmits various information from the device 30 to the server device 60 and receives instructions from the server device 60 according to the information. Alternatively, the edge device 10 receives instructions from the server device 60 without transmitting any information to the server device 60 (for example, when an instruction is sent from the user terminal 5 to the server device 60). The edge device 10 converts the instructions into appropriate instructions according to the models of the device 30 and the sensor switches 20 and transmits them to the device 30 and the sensor switches 20.

[0045] The server device 60 is one or more information processing devices. Although one server device 60 is shown in FIG. 1, the server device 60 may be divided into several devices each with a different function. The functions of the server device 60 may be consolidated into one information processing device. Alternatively, a plurality of server devices 60 with the same functions may be provided, and the plurality of server devices 60 may communicate with each other and perform processing like a server cluster.

[0046] The server device 60 receives various information transmitted from the edge device 10 via the network N and generates necessary instructions. For example, the server device 60 evaluates the performance of image data of an adsorbent using a performance evaluation model. The server device 60 predicts the replacement time based on the evaluated performance and transmits the prediction to the user terminal 5. In addition, in response to an abnormality code from the edge device 10, the server device 60 instructs the edge device 10 to perform emergency operation regardless of the model of the device 30. The server device 60 can also transmit instructions to the edge device 10 for the device 30 according to a schedule or operation set by the user terminal 5.

[0047] The contaminant removal device 40 may evaluate the performance of the adsorbent, rather than the server device 60. The contaminant removal device 40 predicts the replacement time based on the evaluated performance and transmits the prediction to the user terminal 5. The edge device 10 may also have a performance evaluation model.

[0048] The server device 60 also has the functionality of a web server. In response to requests from client software (web clients) such as a web browser operated by a user, the web server provides the clients with screen information written in HTML files, XML, CSS files, JavaScript (registered trademark), etc. Applications that use the web mechanism in this way are called web applications.

[0049] It is preferable that the server device 60 is compatible with cloud computing. Cloud computing refers to a form of usage in which resources on a network are used without being aware of specific hardware resources. Cloud computing provides users with data and software that they previously used on their local computers as services via a network. Users can use various services from any device by providing a web browser running on a personal computer or mobile information terminal, an internet connection, and other necessary equipment.

[0050] The user terminal 5 is a terminal device that displays various screens provided by the server device 60. The user terminal 5 may be used by an administrator or a general user. There are administrators on the customer side and administrators on the management system side, but this disclosure does not distinguish between them. In addition, an administrator is a person who performs maintenance and management that are not performed by general users who use the device 30 on a daily basis.

[0051] The user terminal 5 displays a wide variety of screens, one example of which is a screen that displays information identifying the pollutant removal device 40 and notifies the user when it is time to replace it. Other screens include a list screen of the devices 30 and sensor switches 20 connected to the customer's edge device 10, an in-house map showing the locations of the devices 30 and sensor switches 20, and an operation screen for operating the devices 30 and sensors.

[0052] The user terminal 5 may be, for example, a PC (Personal Computer), a smartphone, a tablet terminal, a PDA (Personal Digital Assistant), or a wearable PC (such as a sunglasses type or a wristwatch type). However, it is sufficient that the terminal has a communication function and can run a web browser. Furthermore, instead of a web browser, the user terminal 5 may run a native app dedicated to the performance evaluation system.

[0053] <Appearance of the pollutant removal device> The contaminant removal device 40 will be described using Figures 2 and 3. Figure 2 shows an external view of an example of the contaminant removal device 40. Figure 3 shows a side view of an example of the contaminant removal device 40. The contaminant removal device 40 has a cubic appearance, and one or more drawers 25-27 can be attached and detached from the front. In the figure, there are three drawers 25-27, which is an example. Each drawer 25-27 is equipped with an adsorption filter 31A, 31B, or 31C, each of which is covered with an adsorbent 32. The adsorption filter 31A is referred to as the first-stage adsorption filter (an example of a first placement location), the adsorption filter 31B is referred to as the second-stage adsorption filter (an example of a second placement location), and the adsorption filter 31C is referred to as the third-stage adsorption filter. The adsorption filters 31A-31C are mounted in the drawers 25-27 and are therefore replaceable. Hereinafter, any of the adsorption filters 31A, 31B, and 31C will be referred to as an adsorption filter 31.

[0054] In FIG. 3, drawers 25 to 27 are arranged in a single vertical row, but contaminant removal device 40 may have multiple rows of drawers.

[0055] In Figure 3, arrows indicate the flow of air. Air is introduced from the front surface of the pollutant removal device 40, passes through adsorption filters 31A, 31B, and 31C, and flows out through vent 33 on the rear surface. As the air passes through the adsorption filters 31, contaminants (e.g., ammonia) are adsorbed onto the adsorbents 32. Although three adsorption filters 31 are shown in Figure 3, this number is an example. The greater the number of adsorption filters 31, the greater the amount of contaminants adsorbed per short period of time can be expected. Furthermore, by changing the type of adsorbent 32 for each adsorption filter 31, different types of contaminants can be adsorbed. In this disclosure, it is assumed that the type of adsorbent 32 in the adsorption filters 31A, 31B, and 31C is the same.

[0056] The air flow may flow in from the rear surface and out from the front surface, or may flow in from one side surface and out from the other side surface, or the air flow may be a combination of these. Also, the shape of the pollutant removal device 40 is an example.

[0057] The pollutant removal device 40 has a camera 50 that captures images of the adsorbent 32 and a communication device 51. For example, as shown in FIG. 3, the camera 50 is installed on the ceiling of the pollutant removal device 40 with its optical axis facing directly downward. Alternatively, the camera 50 may be installed with its optical axis facing so that the adsorption filter 31 is included in the angle of view. At least one camera 50 is installed. A camera 50 may be installed for each of the adsorption filters 31A to 31C, or if it is known that the performance of each of the adsorption filters 31A to 31C decreases to the same extent, the camera 50 may be installed only for one of the three adsorption filters 31A to 31C, that is, for the adsorption filter 31A. In FIG. 3, the camera 50 is installed on the first adsorption filter, but it may also be installed on the second or third adsorption filter.

[0058] One of the features of the present disclosure is the ability to predict when to replace the adsorption filter 31, but it is preferable to replace the adsorption filters 31A to 31C at the same time. This is because replacement requires the customer or a service technician to work in the room where the contaminant removal device 40 is installed, and work in this room requires preparations such as removing contaminants from the body.

[0059] When a camera 50 is installed on only one adsorption filter 31A, the replacement times of the adsorption filters 31B and 31C, which do not have a camera 50, are estimated by the model. The server device 60 evaluates the performance of the adsorption filters 31B and 31C by inputting image data and surrounding environment data (e.g., air temperature, humidity, air volume, etc.) into the model. The server device 60 integrates the performance of the adsorption filters 31A to 31C to calculate the overall performance of the adsorption filters 31A to 31C and predicts the replacement times for the adsorption filters 31A to 31C to be replaced at the same time.

[0060] The camera 50 is sensitive to wavelengths of light that can capture images of salts precipitated on the adsorbent 32. The camera 50 may have a light that emits one or more of visible light, near-infrared light, or ultraviolet light. Near-infrared light is light on the longer wavelength side of the wavelength range outside the visible light range and may capture discoloration that cannot be detected with visible light. Ultraviolet light is light on the shorter wavelength side of the wavelength range outside the visible light range and may capture discoloration that cannot be detected with visible light.

[0061] The camera 50 is integrated with or connected to a communication device 51. The communication device 51 transmits image data of the adsorbent 32 captured by the camera 50 to the edge device 10 in real time. The edge device 10 transmits the image data to the server device 60 in real time. Alternatively, the communication device 51 transmits the image data directly to the server device 60. The image data may be a moving image.

[0062] <Hardware configuration of edge devices and server devices> Next, the hardware configuration of the edge device 10 will be described with reference to Fig. 4. Fig. 4 is a diagram showing an example of the hardware configuration of the edge device 10. As shown in Fig. 4, the edge device 10 has a processor 201, a memory 202, an auxiliary storage device 203, an I / F (Interface) device 204, a communication module 205, and a drive device 206. Note that the hardware components of the edge device 10 are connected to each other via a bus 207.

[0063] The processor 201 has various arithmetic devices such as a CPU (Central Processing Unit), etc. The processor 201 reads and executes various programs onto the memory 202. The processor 201 corresponds to the control unit 110 that controls the entire edge device 10.

[0064] The memory 202 has a main storage device such as a read only memory (ROM) and a random access memory (RAM). The processor 201 and the memory 202 form a so-called computer, and the processor 201 executes various programs read onto the memory 202.

[0065] The auxiliary storage device 203 stores various programs and various data used when the processor 201 executes the various programs.

[0066] The I / F device 204 is a connection device that connects the edge device 10 with the equipment 30, which is an example of an external device, and the sensor switches 20.

[0067] The communication module 205 is a communication device for communicating with the server device 60 via the network N.

[0068] The drive device 206 is a device for loading a recording medium 208. The recording medium 208 here includes media that record information optically, electrically, or magnetically, such as a CD-ROM, a flexible disk, or a magneto-optical disk. The recording medium 208 may also include semiconductor memory that records information electrically, such as a ROM or flash memory.

[0069] The various programs to be installed in the auxiliary storage device 203 are installed, for example, by setting the distributed recording medium 208 in the drive device 206 and reading the various programs recorded on the recording medium 208 by the drive device 206. Alternatively, the various programs to be installed in the auxiliary storage device 203 may be installed by being downloaded from the network N via the communication module 205.

[0070] 5 is a diagram showing an example of the hardware configuration of the server device 60. Note that the hardware configuration of the server device 60 is generally the same as that of the edge device 10, and therefore the following description will focus on the differences from the hardware configuration of the edge device 10. The hardware configuration of the learning device 500, which will be described later, is assumed to be the same as that of the server device 60.

[0071] The processor 221 reads and executes various programs onto the memory 222. The processor 221 corresponds to the control unit 210 that controls the entire server device 60.

[0072] The I / F device 224 is a connection device that connects the display device 230 and the operation device 240, which are examples of external devices, with the server device 60. The display device 230 displays the internal state of the server device 60. The operation device 240 is used when an administrator of the server device 60 inputs various instructions to the server device 60.

[0073] The communication module 225 is a communication device for communicating with the edge device 10 and the user terminal 5 via the network N.

[0074] <About the function> Next, the functional configuration of each device in the performance evaluation system 100 will be described in detail with reference to Fig. 6. Fig. 6 is an example of a functional block diagram that explains the functions of the server device 60 in the performance evaluation system 100 by dividing them into blocks. The server device 60, the pollutant removal device 40, and the learning device 500 are communicably connected via a network N. Note that the edge device 10 is omitted from Fig. 6.

[0075] <<Pollutant removal equipment>> The camera 50 installed in the pollutant removal device 40 captures image data of the adsorbent 32 and transmits it to the communication device 51. The communication device 51 transmits the image data of the adsorbent 32 captured by the camera 50 to the server device 60. The communication device 51 may also instruct the camera 50 on the timing of capturing images. This timing may be set by the customer using the user terminal 5 for the server device 60, and the server device 60 may instruct the communication device 51 to capture the image, or it may be set directly and manually in the communication device 51.

[0076] In the figure, the camera 50 and the communication device 51 are separate entities, but the camera 50 and the communication device 51 may be configured as an integrated module. Also, at least one of the communication device 51 and the camera 50 may be located outside the pollutant removal device 40.

[0077] In addition, when the pollutant removal device 40, rather than the server device 60, predicts the replacement time of the adsorption filter 31, the pollutant removal device 40 has the analysis unit 62, prediction unit 63, and notification unit 64 that the server device 60 has.

[0078] <<Learning device>> The learning device 500 is an information processing device that learns the correspondence between the feature amount of image data and performance in the learning phase of machine learning. The learning device 500 is any information processing device. The learning device 500 generates a performance evaluation model. The server device 60 may have the functions of the learning device 500. The functions of the learning device 500 will be described in detail below.

[0079] <<Server device>> The server device 60 has a communication unit 61, an analysis unit 62, a prediction unit 63, and a notification unit 64. Each of these units of the server device 60 is a function or means realized when any of the components shown in Fig. 5 operates in response to an instruction from the processor 221 in accordance with a program loaded from the auxiliary storage device 223 into the memory 222.

[0080] The communication unit 61 communicates with the edge device 10 (or may directly communicate with the communication device 51) via the network N. In the present disclosure, the communication unit 61 receives image data of the adsorption filter 31A captured by the camera 50.

[0081] The analysis unit 62 analyzes the image data and evaluates the current performance. One example of the performance is the pollutant removal efficiency. That is, the analysis unit 62 analyzes the image data and outputs the pollutant removal efficiency. The function of the analysis unit 62 will be described in detail later.

[0082] The prediction unit 63 records the contaminant removal efficiency in chronological order and calculates an approximate curve from the change in removal efficiency over time since the start of use of the adsorbent. When the approximate curve is extrapolated, the prediction unit 63 predicts the time when the removal efficiency will fall below a threshold value as the replacement time.

[0083] The notification unit 64 notifies the person in charge of the replacement time. The person in charge may be notified of the replacement time each time it is predicted, or may be notified when the remaining time until the replacement time falls below a certain amount.

[0084] <About the functions of the learning device> 7 is a functional block diagram of a learning device 500 according to an embodiment of the present disclosure. The learning device 500 includes a learning data acquisition unit 502, a learning data storage unit 503, a feature detection unit 504, and a learning unit 505. These functional units of the learning device 500 are functions or means realized by the processor 221 of the learning device 500 executing instructions of a program loaded in the memory 222. The learning device 500 is used in the learning phase of machine learning.

[0085] The learning data acquisition unit 502 acquires learning data 501. The learning data 501 has a plurality of sets of image data and contaminant removal efficiencies, with one set consisting of image data and contaminant removal efficiencies (performance).

[0086] The learning data storage unit 503 stores the learning data 501 acquired by the learning data acquisition unit 502 .

[0087] The feature amount detection unit 504 analyzes the image data and first quantifies the following: 1. Number of white pixels 2. Particle size information (area, perimeter, maximum diameter, minimum diameter, average particle diameter) The absolute values ​​of the number of white pixels and particle size information may vary even in the initial state immediately after the adsorption filter 31 is attached to the pollutant removal device 40. Therefore, in this disclosure, the amount of change in the number of white pixels and particle size information is used as a feature. The feature may be referred to as an explanatory variable, and the removal efficiency as a target variable (sometimes referred to as training data). Note that the learning device 500 may learn the correspondence between the image data itself and the removal efficiency without extracting feature values ​​from the image data.

[0088] The learning unit 505 learns the learning data using various machine learning algorithms to generate a performance evaluation model 506. The performance evaluation model 506 is correspondence information that associates the feature amount of image data with the removal efficiency. In other words, the performance evaluation model 506 outputs the removal efficiency in response to the input of the feature amount of image data.

[0089] Machine learning is a technology that allows computers to acquire human-like learning capabilities. It is a technology in which a computer autonomously generates algorithms necessary for making judgments, such as data classification, from previously acquired training data and applies these algorithms to new data to make predictions. The learning method for machine learning may be any of supervised learning, unsupervised learning, semi-supervised learning, reinforcement learning, and deep learning, or may be a combination of these learning methods. The performance evaluation model 506 can be realized by a regression model, which is a type of supervised learning. Regression models include multiple regression, neural networks, ridge regression, lasso regression, and elastic net regression, and are not limited to the methods described in this disclosure.

[0090] <About the analysis unit functions> Next, functions of the analysis unit 62 will be described with reference to Fig. 8. Fig. 8 is a functional block diagram of the analysis unit 62 according to an embodiment of the present disclosure. The analysis unit 62 includes an image data acquisition unit 512, a feature amount detection unit 504, and an inference unit 513. The analysis unit 62 is used in the inference phase of machine learning.

[0091] The image data acquisition unit 512 acquires image data captured by the camera 50 from the communication device 51. The image data acquisition unit 512 may receive the data from the communication device 51, or may acquire image data stored in the auxiliary storage device 223. Since a sudden decrease in the performance of the adsorbent 32 rarely occurs, the performance may be evaluated by batch processing that is performed at a fixed time each day.

[0092] Similar to the learning device 500, the feature detection unit 504 detects features from image data.

[0093] The inference unit 513 inputs the feature amount detected from the image data into the performance evaluation model 506 and outputs the pollutant removal efficiency 514. The pollutant removal efficiency 514 is estimated as the performance of the pollutant removal device 40.

[0094] <Relationship between change in number of white pixels and removal efficiency> Next, we will explain the relationship between the change in the number of white pixels and the removal efficiency. When ammonia, for example, is adsorbed onto the adsorbent 32, salt precipitates over time, turning the adsorbent white. Because the adsorption of ammonia reduces performance, the performance of the pollutant removal device 40 can be estimated based on the color of the adsorbent 32.

[0095] Figure 9 shows an example of the experimentally determined relationship between the elapsed time since the start of use of the adsorbent, image data, and pollutant removal efficiency. The elapsed time is the time elapsed since the adsorption filter 31 was attached to the pollutant removal device 40. The removal efficiency is assumed to be a reliable value, such as a value measured by the manufacturer. The correspondence between the elapsed time, the number of white pixels, and the removal efficiency is assumed to be as follows: After 15 minutes 28531 pieces 99% After 30 minutes 30949 pieces 98% After 60 minutes 30957 pieces 93% After 120 minutes 32395 pieces 42% After 180 minutes 32469 pieces 0% In this way, the performance of the adsorbent 32 correlates with the discoloration state of the adsorbent 32. Note that the experimental results in Figure 9 are the results of an accelerated test. In reality, it takes six months to a year or more for the removal efficiency to reach 0%.

[0096] FIG. 10 is a scatter plot showing the correspondence between the number of white pixels and the removal efficiency of contaminants. FIG. 10 also shows an approximation curve that approximates the correspondence between the number of white pixels and the removal efficiency. According to FIG. 10, performance gradually decreases as the number of white pixels increases, and then performance rapidly decreases as the number of white pixels increases. Note that this approximation may vary depending on the combination of adsorbent and contaminant. If the learning unit 505 learns the correspondence between the number of white pixels and the removal efficiency using a regression model, the current removal efficiency can be inferred from the number of white pixels. However, the number of white pixels changes each time the adsorption filter 31 is attached, even if the elapsed time is 0 minutes. Therefore, in this disclosure, the change in the number of white pixels from the initial state is used as a feature.

[0097] In the present disclosure, the amount of change in the number of white pixels is used as the feature, but the amount of change in each density of the grayscale may be used as the feature, instead of the two colors of white and black. Also, a color histogram may be created and the amount of change in each color may be used as the feature.

[0098] Figure 11 shows a schematic diagram of the relationship between the change in the number of white pixels and the ammonia removal efficiency. Note that Figure 11 does not show experimental results, but rather shows the estimated relationship between the change in the number of white pixels and the removal efficiency (performance). The change in the number of white pixels is calculated as follows: Change in number of white pixels = (current number of white pixels - number of white pixels at the initial state when elapsed time is 0 minutes) / number of white pixels at the initial state when elapsed time is 0 minutes The learning unit 505 generates a performance evaluation model 506 by learning the correspondence between the change in the number of white pixels and the contaminant removal efficiency using a regression model. The performance evaluation model 506 outputs the removal efficiency of the adsorption filter 31 arranged in the contaminant removal device 40. As will be described later, the learning unit 505 may learn the correspondence between the change in the number of white pixels and the change in mass (or volume) and the contaminant removal efficiency, in addition to the change in the number of white pixels.

[0099] Alternatively, the feature detection unit 504 may digitize the ratio of the number of white pixels to the total number of pixels instead of the number of white pixels. If the correspondence between the change in the number of white pixels and the removal efficiency is learned, there is a risk that the learning device 500 will need to re-learn if the type or performance of the camera 50 changes. In contrast, the ratio of the number of white pixels to the total number of pixels is unlikely to change even if the type or performance of the camera 50 changes, so the need for re-learning can be reduced.

[0100] <How to detect the number of white pixels> 12 is a flowchart illustrating the process of quantifying the number of white pixels by the feature amount detection unit 504. The method of quantifying the number of white pixels may be the same in the learning phase and the inference phase.

[0101] First, the feature amount detection unit 504 converts the color image data captured by the camera 50 into grayscale (S1).

[0102] Next, the feature detection unit 504 binarizes the grayscale image data using a predetermined threshold (S2). As a binarization method, there are static binarization, variable threshold processing, Otsu's binarization, etc., and an appropriate method is adopted.

[0103] The feature detection unit 504 counts the number of white pixels (S3). The number of white pixels may be counted as the number of pixels as is, or multiple pixels may be counted as one block, and if a certain number or more of the pixels in the block are white, the entire block may be counted as one white pixel.

[0104] <Relationship between mass change and removal efficiency> Next, the relationship between the amount of change in mass or volume and the removal efficiency will be described. Since mass or volume cannot be directly detected from image data, in this disclosure, the feature detection unit 504 digitizes particle size information from the image data and converts the particle size information into mass or volume.

[0105] Figure 13 is an illustration that explains the increase in mass and volume due to the adsorption of pollutants. As shown in Figure 13, when ammonia is adsorbed onto adsorbent particles 125, a reaction produces salt 123 and water, and the mass and volume of particle 125 increase.

[0106] The mass or volume correlates with particle size information. The particle size information includes the area, perimeter, maximum diameter, minimum diameter, average particle size, etc. of each particle of the adsorbent 32. When ammonia is adsorbed onto the adsorbent 32, its performance decreases, so the performance of the adsorbent 32 correlates with the mass or volume of the adsorbent 32 or the particle size information.

[0107] FIG. 14 shows an example of image data and particle size information of the adsorbent 32. FIG. 14(a) shows image data of the adsorbent 32, captured using a general-purpose camera 50 (visible light). FIG. 14(b) shows the particle size distribution of the adsorbent 32, and is a histogram of diameters with a BIN width of 10 mm. As will be described later, each particle of the adsorbent 32 is detected by image processing, and the area, perimeter, maximum diameter, minimum diameter, and average particle size can be determined.

[0108] FIG. 15 is a flowchart illustrating the process of detecting granularity information.

[0109] First, the feature detection unit 504 performs image preprocessing on the image data (S11). The feature detection unit 504 performs preprocessing by, for example, removing noise from the image data and adjusting contrast and brightness so that particles can be easily identified.

[0110] Next, the feature detection unit 504 uses a segmentation technique to separate the particles of the adsorbent 32 into individual particles (S12). Known segmentation techniques include the watershed method and thresholding. The watershed method is an image processing algorithm that separates (segments) contacting objects. Thresholding is an algorithm that generates a binary image divided into pixels with intensities below a specific threshold and pixels with intensities above the threshold. These techniques are existing or well known, so a description thereof will be omitted.

[0111] Next, the feature detection unit 504 calculates the area, perimeter, maximum diameter, minimum diameter, and average particle size for each particle separated by segmentation (S13). The area is calculated by the number of pixels occupying one particle, the perimeter is calculated by the number of pixels on the outer edge of one particle, and the diameter is calculated by the length of a line passing through the center of one particle, etc., through image processing.

[0112] Next, the feature detection unit 504 converts the particle size information into mass or volume using one or more of these pieces of particle size information (S14). In the present disclosure, it is assumed that correlation coefficients p and q that correspond to particle size information and mass or volume have been obtained. The feature detection unit 504 converts the particle size information into mass or volume using the correlation coefficients p and q. Note that the particle size information may be one or more of area, perimeter, maximum diameter, minimum diameter, and average particle size. Mass = p x particle size information Volume = q x particle size information The correlation coefficient may be determined by machine learning. That is, a model that has learned the correspondence between particle size information and mass or volume receives particle size information as input and outputs mass or volume.

[0113] If the learning unit 505 learns the correspondence between mass or volume and removal efficiency using a regression model, the current removal efficiency can be inferred from the mass or volume. However, the mass or volume of the adsorbent 32 filled in the adsorption filter 31 is unlikely to be the same even in the initial state. Therefore, in the present disclosure, the change in mass or volume from the initial state is used as a feature quantity.

[0114] Figure 16 shows a schematic diagram of the relationship between the amount of change in mass or volume and the removal efficiency. Note that Figure 16 does not show experimental results, but rather shows the estimated relationship between the amount of change in mass or volume and the removal efficiency (performance). The amount of change in mass is calculated as follows: Change in mass = (current mass - mass at initial state when time is 0 minutes) / mass at initial state when time is 0 minutes The change in volume is calculated as follows: Volume change = (current volume - volume at initial state when elapsed time is 0 minutes) / volume at initial state when elapsed time is 0 minutes The learning unit 505 generates the performance evaluation model 506 by learning the correspondence between the amount of change in mass or volume and the efficiency of contaminant removal using a regression model. The learning unit 505 may also learn the correspondence between the amount of change in the number of white pixels and the amount of change in mass (or volume) and the efficiency of contaminant removal. This allows for only one performance evaluation model 506, thereby reducing the processing load.

[0115] <An example of a performance evaluation model> The performance evaluation model 506 obtained by regression analysis is, for example, as follows:

[0116]

number

[0117] Furthermore, a performance evaluation model 506 may be generated using a neural network, as shown in Fig. 17. Fig. 17 shows a performance evaluation model 506 configured using a neural network 170. The neural network 170 is a form of artificial intelligence (AI) model that trains a computer to process data in a manner that mimics the workings of the human brain. An existing configuration is used for the neural network 170.

[0118] The neural network 170 corresponds to the performance evaluation model 506. The neural network 170 has an input layer 171, an intermediate layer 172, and an output layer 173. In the neural network 170 of FIG. 17, L layers are fully connected from node 175 of the input layer 171 to node 176 of the output layer 173. A neural network with a deep hierarchy is called a DNN (Deep Neural Network). The layer between the input layer 171 and the output layer 173 is called an intermediate layer 172. The number of intermediate layers 172 and the number of nodes are merely examples. The inputs to the input layer 171 are the change in the number of white pixels, the change in mass or volume, and a bias (assumed to be 1). The output from the output layer 173 is the pollutant removal efficiency.

[0119] Weights are set for the connections between nodes, and the value obtained by multiplying the output from a node by the weight is transmitted to the node in the next layer. The node in the next layer receives the outputs of all the nodes in the previous layer, so the node in the next layer sums the outputs of all the nodes in the previous layer. The node in the next layer activates the summed output using an activation function and transmits it to the next node. This process is repeated until the value is transmitted all the way to the output layer 173.

[0120] In this embodiment, since we want to infer the removal efficiency, the neural network 170 is a regression model (other models include classification models). For this reason, the output layer 173 is provided with one node 176 that outputs the removal efficiency.

[0121] It is assumed that the neural network 170 is trained by an existing method such as the error backpropagation method in the learning phase. That is, the performance evaluation model 506 learns the correspondence between the change amount of the number of white pixels and the change amount of mass or volume and the removal efficiency. However, only one of the change amount of the number of white pixels or the change amount of mass or volume may be used for learning. Further, other information may be input.

[0122] In the inference phase using the performance evaluation model 506, for example, the change amount of the number of white pixels and the change amount of mass or volume are input to the input layer 171. The output layer 173 calculates (infers) the removal efficiency. Since there is a correlation between mass or volume and particle size information, instead of mass or volume, or the change amount of particle size information may be input together with mass or volume.

[0123] For the input of the change amount of the number of white pixels and the change amount of mass or volume detected from one image data, the output layer 173 outputs one removal efficiency. By repeatedly outputting the removal efficiency by the performance evaluation model 506, the removal efficiency associated with the elapsed time can be obtained. The prediction unit 63 can calculate an approximate curve from the relationship between the elapsed time and the removal efficiency and predict the lifetime until the extrapolated removal efficiency becomes less than the threshold value.

[0124] [[ID=H12]]<Performance Evaluation Model Using CNN> As shown in FIG. 18, a CNN (Convolutional Neural Network) may be used for the performance evaluation model 506. FIG. 18 shows a configuration example of the performance evaluation model 506 using the CNN 60. As an example, the CNN 60 has convolutional layers 72, 74, pooling layers 73, 75, and a fully connected layer 70. The input image 71 is image data captured by the camera 50. The input image 71 is processed in the order of the convolutional layer 72, the pooling layer 73, the convolutional layer 74, the pooling layer 75, and the fully connected layer 70.

[0125] The convolutional layers 72 and 74 convert lattice-like numerical data called kernels (or filters) and numerical data of partial images (called windows) of the same size as the kernels into a single numerical value by calculating the sum of the products of each element. The convolutional layers 72 and 74 convert this conversion process into small lattice-like numerical data (i.e., tensors) by shifting the windows little by little. The lattice-like numerical data is activated by an activation function and input to the pooling layers 73 and 75.

[0126] The pooling layers 73 and 75 are used to create a single numerical value from the activated numerical data. Examples include maximum pooling, which selects the maximum value in a window, and average pooling, which selects the average value in a window. The convolutional layers 72 and 74 extract features from the image data, and the pooling layers 73 and 75 reduce the sensitivity to the position of the object. The activation function is a function that nonlinearly transforms (activates) the input (for example, ReLU, tanh, sigmoid, etc.).

[0127] The output of the pooling layer 75 is input to the input layer 171 of the fully connected layer 70 (neural network). The information input to the input layer 171 propagates through the intermediate layer 172, and the node 176 in the output layer 173 outputs the removal efficiency. The structure of the neural network was explained in Fig. 17, but it is preferable that the number of layers and the number of nodes are optimized for the CNN.

[0128] In this way, even if image data is input directly (or after preprocessing) to the performance evaluation model 506, the output layer 173 can infer the removal efficiency. However, learning with extracted features is expected to require less training data. Since measuring performance at the stage of preparing training data is costly, having less training data is one advantage.

[0129] <Replacement time prediction> Next, a method for predicting the replacement time of the adsorption filter 31 will be described with reference to FIG. 19 and other figures. FIG. 19 is a graph showing the relationship between elapsed time and removal efficiency. The elapsed time is the time elapsed since the adsorption filter 31 was attached to the contaminant removal device 40. The vertical axis represents the removal efficiency inferred by the inference unit 513 using the performance evaluation model 506. In FIG. 19, the inference unit 513 inferred the removal efficiency at times t1, t2, and t3. The prediction unit 63 approximates the relationship between elapsed time and removal efficiency with a pre-set approximation curve 121. The approximation curve 121 intersects with the removal efficiency (e.g., 50%) that serves as a guideline for replacing the adsorption filter 31, and the time tx at which the approximation curve 121 intersects with the time axis is the replacement time (also referred to as the lifespan). Furthermore, with t3 being the current time, tx - t3 is the remaining time until replacement.

[0130] <Action or Processing> FIG. 20 is a sequence diagram illustrating the flow of processing performed by the performance evaluation system 100 to predict the replacement time of the adsorption filter 31.

[0131] S31: The camera 50 captures images of the adsorption filter 31 at a preset cycle. The cycle may be a frequency that corresponds to the lifespan of the adsorbent 32, such as once a day. The camera 50 may capture images irregularly, or may capture images in response to an instruction from the user terminal 5. The communication device 51 transmits image data of the adsorption filter 31 to the server device 60.

[0132] S32: The communication unit 61 of the server device 60 receives the image data. The feature detection unit 504 analyzes the image data and digitizes the number of white pixels and the particle size information. The feature detection unit 504 further converts the particle size information into mass or volume. The feature detection unit 504 calculates the difference between the recorded number of white pixels in the initial state (the number of white pixels in the initial state when the elapsed time is 0 minutes) and the current number of white pixels as the amount of change in the number of white pixels. The feature detection unit 504 calculates the difference between the recorded mass or volume in the initial state (the mass or volume in the initial state when the elapsed time is 0 minutes) and the current mass or volume as the amount of change in mass or volume.

[0133] S33: The inference unit 513 inputs the feature amount into the performance evaluation model 506 and infers the removal efficiency of pollutants.

[0134] S34: The prediction unit 63 approximates the relationship between elapsed time and removal efficiency using a pre-set approximation curve 121, and predicts the replacement time (lifespan) at which the curve intersects with the removal efficiency (e.g., 50%) that serves as a guideline for replacing the adsorption filter 31.

[0135] S35: The notification unit 64 determines whether the remaining time until the end of the service life is equal to or less than a threshold value. If it is equal to or less than the threshold value, the notification unit 64 notifies the user terminal 5 of the recommendation to replace the adsorption filter. The notification unit 64 may send the notification by email, or may notify via social media or the like. The notification may also be sent to a customer representative or a service technician.

[0136] <Major Effects> As described above, the performance evaluation system of the present disclosure includes: (i) Since the server device 60 can evaluate the performance of the adsorbent 32 from a remote location, the timing for replacing the adsorbent 32 can be predicted without a service technician having to visit the site, which enables cost reduction. (ii) The server device 60 can constantly evaluate the performance of the adsorbent 32, making it easier to determine the appropriate time to replace the adsorbent 32. (iii) Since the server device 60 can predict the replacement time, excessive replacement and delays in replacement can be suppressed.

[0137] [Second embodiment] In this embodiment, the inference of the removal efficiency of the second- and third-stage adsorption filters 31B, 31C will be described. In the first embodiment, it was necessary to place cameras 50 on all of the adsorption filters 31A to 31C, or to replace the adsorption filters 31B, 31C based on the performance of the first-stage adsorption filter 31A. In this embodiment, the server device 60 uses the removal efficiency inferred for the first-stage adsorption filter 31A and surrounding environment data to infer the removal efficiencies of the second- and third-stage adsorption filters 31B, 31C.

[0138] <Summary> Assume that three adsorption filters 31A to 31C are installed in the pollutant removal device 40. The removal efficiency 531 of the first-stage adsorption filter 31A can be inferred in the same manner as in the first embodiment. Because cameras 50 are not installed on the second-stage adsorption filter 31B and the third-stage adsorption filter 31C, in order to infer their performance, in this disclosure, the temperature, humidity, and air volume of the air passing through the adsorption filters 31B and 31C are used as ambient environment data.

[0139] First, the removal efficiency of the first-stage adsorption filter 31A, the second-stage adsorption filter 31B, and the third-stage adsorption filter 31C, as well as surrounding environment data (temperature, humidity, air volume), are obtained as learning data. Note that the removal efficiencies of the adsorption filters 31A to 31C are assumed to be reliable values, such as values ​​measured by manufacturers. The learning device 500 generates a second-stage model and a third-stage model using the following as the objective variable (teaching data) and explanatory variables: The second-stage model is a performance evaluation model that infers the removal efficiency of the second-stage adsorption filter 31B, and the third-stage model is a performance evaluation model that infers the removal efficiency of the third-stage adsorption filter 31C. Second stage model (an example of the second model) Objective variable: Removal efficiency of the second-stage adsorption filter Explanatory variables: performance of the first stage adsorption filter, temperature, humidity, air volume Third stage model Objective variable: removal efficiency of the third-stage adsorption filter Explanatory variables: performance of the first stage adsorption filter, temperature, humidity, air volume During inference, the performance of the first-stage adsorption filter 31A and the surrounding environment data inferred by the performance evaluation model 506 are input into the second-stage model and the third-stage model, whereby the performance of the second-stage adsorption filter 31B and the performance of the third-stage adsorption filter 31C are inferred.

[0140] The ambient environment data is controlled by an air conditioner, so it is expected that there will be little fluctuation throughout the year. However, it is more preferable to use the ambient environment data, which is an explanatory variable, as an integrated value since the adsorption filter 31 was installed, or to add time to the explanatory variable.

[0141] <Example of temperature sensor, humidity sensor, and airflow sensor placement> 21 shows an example of the arrangement of temperature sensor 21, humidity sensor 22, and air flow sensor 23. The temperature sensor 21, humidity sensor 22, and air flow sensor 23 need only be located in the air flow path, and the diagram is just an example. A contaminant removal device 40 is installed at the end of a duct 35 from which air flows out. Contaminants are removed from the air flowing out of duct 35 by contaminant removal device 40, and the air passes through mesh 36 before flowing into the clean room. Therefore, contaminant removal device 40 is installed in storage space 37 just before the clean room.

[0142] The temperature sensor 21, humidity sensor 22, and air volume sensor 23 are arranged, for example, in a mesh 36. The temperature sensor 21, humidity sensor 22, and air volume sensor 23 are connected to a communication device 51, and the communication device 51 transmits the temperature, humidity, and air volume to a server device 60.

[0143] <About the function> Fig. 22 is an example of a functional block diagram for explaining the functions of the performance evaluation system 100 of this embodiment, divided into blocks. The explanation of Fig. 22 will mainly focus on the differences from Fig. 6. A temperature sensor 21, a humidity sensor 22, and an air volume sensor 23 are connected to a communication device 51. The communication device 51 transmits the temperature, humidity, and air volume together with image data captured by a camera 50 to a server device 60.

[0144] The imaging period of the image data may be different from the measurement period of the temperature sensor 21, humidity sensor 22, and air volume sensor 23. This is because the temperature, humidity, and air volume fluctuate little throughout the year. If the temperature, humidity, and air volume may fluctuate, it is preferable for the communication device 51 to transmit the temperature, humidity, and air volume to the server device 60 at a shorter period than the image data. The server device 60, for example, integrates the received temperature, humidity, and air volume and uses the integrated value for inference.

[0145] The server device 60 newly includes a removal efficiency integrating unit 65. The removal efficiency integrating unit 65 integrates the removal efficiency of the first-stage adsorption filter 31A, the second-stage adsorption filter 31B, and the third-stage adsorption filter 31C to calculate the removal efficiency (performance) of the entire contaminant removal device 40. Details will be described later.

[0146] <About the functions of the learning device> Figure 23 is a functional block diagram of a learning device 500 according to an embodiment of the present disclosure. The explanation of Figure 23 will mainly focus on the differences from Figure 7. Of the functions possessed by learning device 500, the function for generating performance evaluation model 506 may be the same as that of Figure 7. Here, the function for generating second-stage model 507 and third-stage model 508 will be explained.

[0147] The training data 501 for generating the second-stage model 507 includes the performance of the first-stage adsorption filter 31A, the performance of the second-stage adsorption filter 31B, and ambient environment data (temperature, humidity, air volume). A plurality of sets of these are prepared.

[0148] The training data 501 for generating the third-stage model 508 includes the performance of the first-stage adsorption filter 31A, the performance of the third-stage adsorption filter 31C, and ambient environment data (temperature, humidity, air volume). A plurality of sets of these are prepared.

[0149] The learning unit 505 learns the correspondence between the performance and surrounding environment data of the first-stage adsorption filter 31A and the removal efficiency of the second-stage adsorption filter 31B, thereby generating a second-stage model 507. The second-stage model 507 is correspondence information that associates the performance and surrounding environment data of the first-stage adsorption filter 31A with the removal efficiency of the second-stage adsorption filter 31B.

[0150] The learning unit 505 learns the correspondence between the performance and surrounding environment data of the first-stage adsorption filter 31A and the removal efficiency of the third-stage adsorption filter 31C to generate a third-stage model 508. The third-stage model 508 is correspondence information that associates the performance and surrounding environment data of the first-stage adsorption filter 31A with the removal efficiency of the third-stage adsorption filter 31C.

[0151] <About the analysis unit functions> Figure 24 is a functional block diagram of an analysis unit 62 according to an embodiment of the present disclosure. The explanation of Figure 24 will mainly focus on the differences from Figure 8. The analysis unit 62 newly includes a surrounding environment data acquisition unit 522, a first inference unit 523, a second inference unit 524, and a third inference unit 525. The first inference unit 523 has the same function as the inference unit 513 in Figure 8, and inputs the feature amount detected from image data 511 to a performance evaluation model 506 and outputs the removal efficiency of the first-stage adsorption filter 31A.

[0152] The ambient environment data acquisition unit 522 acquires the temperature, humidity, and air volume (ambient environment data 521) measured by the temperature sensor 21, humidity sensor 22, and air volume sensor 23 from the communication device 51.

[0153] The second inference unit 524 inputs the surrounding environment data 521 acquired by the surrounding environment data acquisition unit 522 and the removal efficiency 531 of the first stage adsorption filter 31A inferred by the first inference unit 523 into the second stage model 507, and outputs the removal efficiency 532 of the second stage adsorption filter 31B.

[0154] The third inference unit 525 inputs the surrounding environment data 521 acquired by the surrounding environment data acquisition unit 522 and the removal efficiency 531 of the first stage adsorption filter 31A inferred by the first inference unit 523 into the third stage model 508, and outputs the removal efficiency 533 of the third stage adsorption filter 31C.

[0155] <Examples of the second and third models> The second model 507 or the third model 508 obtained by the regression analysis is, for example, as follows: Second model

[0156]

number

[0157]

number

[0158] Fig. 25 shows a second-stage model 507 configured by the neural network 182. Compared to Fig. 17, the removal efficiency 531 of the first-stage adsorption filter 31A inferred by the first inference unit 523, the temperature, humidity, and air volume are input to the input layer 171. Furthermore, the output layer 173 outputs the removal efficiency of the second-stage adsorption filter 31B.

[0159] In the inference phase, first, the first inference unit 523 infers the removal efficiency 531 of the first-stage adsorption filter 31A using the performance evaluation model 506. Next, the second inference unit 524 inputs the removal efficiency 531 of the first-stage adsorption filter 31A inferred by the first inference unit 523, the temperature, humidity, and air volume into the second-stage model 507, and the output layer 173 can calculate (infer) the removal efficiency 532 of the second-stage adsorption filter 31B.

[0160] Each time the image data and the surrounding environment data are transmitted, the second inference unit 524 outputs the removal efficiency. Therefore, the removal efficiency is associated with the elapsed time. The prediction unit 63 calculates an approximation curve from the relationship between the elapsed time and the removal efficiency, and can predict the replacement time (lifespan) when the removal efficiency will be below the threshold.

[0161] The configuration of the neural network 182 for the third model is similar, so it is not shown in the figure.

[0162] <Overall removal efficiency of the pollutant removal device> Therefore, in this disclosure, removal efficiency curves 41 to 43 are obtained for each of the three adsorption filters 31A to 31C, as shown in Figure 26. Figure 26 shows an example of the removal efficiency curves of the adsorption filters 31A to 31C versus elapsed time. Figure 26 shows not only the removal efficiency curves 41 to 43 for each of the adsorption filters 31A to 31C, but also an integrated removal efficiency curve 44 that integrates these.

[0163] A method for integrating multiple removal efficiencies will now be described. Equation (4) is a calculation formula for the integrated removal efficiency curve 44.

[0164]

number

[0165] The weights w(i) are calculated using equation (5). The sum of the weights w(i) is 1.

[0166]

number

[0167] S1 and S2 are determined so that S1 + S2 = 1. The weighting coefficients for each factor are a1 and a2. From the above, the removal efficiency integration unit 65 can calculate the integrated removal efficiency curve 44.

[0168] <Action or Processing> FIG. 27 is a sequence diagram illustrating the flow of processing performed by the performance evaluation system to predict the replacement time of the adsorption filter.

[0169] S51: The camera 50 captures images of the adsorption filter 31 at a preset cycle. The temperature sensor 21 detects the temperature, the humidity sensor 22 detects the humidity, and the air volume sensor 23 detects the air volume. The communication device 51 transmits image data of the adsorption filter 31 and data on the surrounding environment to the server device 60.

[0170] S52: The communication unit 61 of the server device 60 receives the image data and the surrounding environment data. The feature amount detection unit 504 detects feature amounts in the same manner as in step S32.

[0171] S53: The first inference unit 523 inputs the feature amount into the performance evaluation model 506 and infers the removal efficiency 531 of the first-stage adsorption filter 31A.

[0172] S54: The second inference unit 524 inputs the removal efficiency 531 of the first-stage adsorption filter 31A and the surrounding environment data into the second-stage model 507, and infers the removal efficiency 532 of the second-stage adsorption filter 31B.

[0173] S55: The third inference unit 525 inputs the removal efficiency 531 of the first-stage adsorption filter 31A and the surrounding environment data into the third-stage model 508, and infers the removal efficiency 533 of the third-stage adsorption filter 31C.

[0174] S56: The removal efficiency integration unit 65 calculates the overall removal efficiency from the three removal efficiencies.

[0175] S57: The prediction unit 63 approximates the relationship between the elapsed time and the removal efficiency with a preset approximation curve, and predicts the time (life) at which the curve intersects with the removal efficiency (for example, 50%) that is the guideline for replacing the adsorption filter 31.

[0176] S58: The notification unit 64 determines whether the remaining time until the end of the service life is equal to or less than a threshold value. If it is equal to or less than the threshold value, the notification unit 64 notifies the user terminal 5 of the recommendation to replace the adsorption filter. The notification unit 64 may send the notification by email, or may notify via social media or the like. The notification may also be sent to a customer representative or a service technician.

[0177] <Major Effects> As described above, the performance evaluation system 100 of the present disclosure can infer the performance of adsorption filters 31B and 31C based on the performance of adsorption filter 31A and surrounding environment data. Furthermore, the performance evaluation system 100 can integrate the performance of adsorption filters 31A to 31C to calculate the performance of the entire contaminant removal device 40 (all of adsorption filters 31A to 31C), allowing a service technician to replace adsorption filters 31A to 31C all at once.

[0178] <Reasons for the effect> The first aspect of the present disclosure "analyzes the image data to obtain feature quantities related to the performance of the adsorbent, infers the performance of the adsorbent by inputting the feature quantities into correspondence information that associates the feature quantities with the performance of the adsorbent, and predicts the replacement time of the adsorbent based on the relationship between the elapsed time since the start of use of the adsorbent and the inferred performance of the adsorbent," thereby making it possible to predict the appropriate replacement time of the adsorbent.

[0179] In the second aspect of the present disclosure, "the feature quantity related to the performance of an adsorbent is a feature quantity that changes depending on the amount of salt precipitated on the adsorbent," and therefore the amount of salt that correlates with the performance of the adsorbent can be used as a feature quantity related to performance, thereby enabling the performance of the adsorbent to be inferred.

[0180] In the third aspect of the present disclosure, "the feature is the amount of change in color of the adsorbent, and the control unit inputs the amount of change in color into the correspondence information to infer the performance of the adsorbent," so that the amount of change in color due to salt precipitation can be used as a feature to infer the performance of the adsorbent.

[0181] The fourth aspect of the present disclosure is "the amount of change in mass or volume of the adsorbent, and the control unit inputs the amount of change in mass or volume of the adsorbent into the correspondence information to infer the performance of the adsorbent," so that the amount of change in mass or volume due to salt precipitation can be used as a feature to infer the performance of the adsorbent.

[0182] A fifth aspect of the present disclosure provides a method for detecting a feature amount, the feature amount being a change amount of particle size information of the adsorbent; The control unit inputs the amount of change in the particle size information into the correspondence information and infers the performance of the adsorbent, so the amount of change in the particle size information that changes due to salt precipitation can be used as a feature to infer the performance of the adsorbent.

[0183] The sixth aspect of the present disclosure "notifies the terminal device of the predicted replacement time," so that the user operating the terminal device can know when the replacement time will come.

[0184] In the seventh aspect of the present disclosure, "the correspondence information is a first model that is machine-learned to determine the correspondence between the feature and the performance of the adsorbent," and therefore, the correspondence information can be generated using a machine learning method implemented in a tool, etc.

[0185] An eighth aspect of the present disclosure is that "the control unit infers the performance of the adsorbent at the second installation location by inputting the performance of the adsorbent at the first installation location inferred by the first model and the surrounding environmental data of the pollutant removal device into a second model that has learned the correspondence between the performance of the adsorbent at the first installation location and the surrounding environmental data, and the performance of the adsorbent at the second installation location," so that the performance of the adsorbent at the first installation location can be inferred from the performance of the adsorbent at the first installation location and the surrounding environmental data of the pollutant removal device, without having to provide a camera for each adsorbent.

[0186] The ninth aspect of the present disclosure "calculates the overall performance of the adsorbent in the pollutant removal device by integrating the weighted performance of the adsorbent in the first placement location and the performance of the adsorbent in the second placement location," so that even if the inferred performance differs depending on the placement location, the performance of the adsorbent in the entire device can be calculated and multiple adsorbents can be replaced at once.

[0187] In the tenth aspect of the present disclosure, "the imaging means has sensitivity to the wavelength of light that images the salt precipitated on the adsorbent," making it easy to image the precipitated salt. [Explanation of symbols]

[0188] 40 Pollutant removal equipment 60 Server equipment 100 Performance Evaluation System

Claims

1. A performance evaluation system in which a pollutant removal device that removes pollutants from the air and an information processing device can communicate with each other via a network, The pollutant removal device includes: an adsorbent that adsorbs the contaminants; imaging means for imaging the adsorbent; a communication device that transmits image data of the adsorbent captured by the imaging means to the information processing device, the information processing device has a control unit, the control unit analyzes the image data to acquire a feature amount related to the performance of the adsorbent; Inferring the performance of the adsorbent by inputting the feature into correspondence information that associates the feature with the performance of the adsorbent; predicting the time to replace the adsorbent based on the relationship between the elapsed time since the start of use of the adsorbent and the inferred performance of the adsorbent; Performance evaluation system.

2. the feature quantity related to the performance of the adsorbent is a feature quantity that changes depending on the amount of salt precipitated on the adsorbent; The performance evaluation system according to claim 1 .

3. the characteristic amount is a change in color of the adsorbent, the control unit inputs the amount of change in color into the correspondence information to infer the performance of the adsorbent. A performance evaluation system according to claim 1 or 2.

4. the characteristic amount is a change in mass or volume of the adsorbent, the control unit inputs a change in mass or volume of the adsorbent into the correspondence information to infer the performance of the adsorbent. A performance evaluation system according to claim 1 or 2.

5. the feature amount is a change amount of particle size information of the adsorbent, the control unit inputs the amount of change in the particle size information into the correspondence information to infer the performance of the adsorbent. A performance evaluation system according to claim 1 or 2.

6. The control unit notifies a terminal device of the predicted replacement time. The performance evaluation system according to claim 1 .

7. The correspondence information is a first model obtained by machine learning the correspondence between the feature amount and the performance of the adsorbent. The performance evaluation system according to claim 1 .

8. the contaminant removal device has the adsorbent disposed at a first location and a second location; the communication device transmits the environmental data surrounding the pollutant removal device to the information processing device; the control unit calculates a second model that has learned a correspondence between the performance of the adsorbent at the first placement location and the surrounding environment data, and the performance of the adsorbent at the second placement location, inputting the performance of the adsorbent at the first location estimated by the first model and environmental data surrounding the pollutant removal device to estimate the performance of the adsorbent at the second location; The performance evaluation system according to claim 7 .

9. the control unit weights the performance of the adsorbent at the first placement location estimated by the first model by a flow rate of air or a deterioration rate of the adsorbent; weighting the performance of the adsorbent at the second location by the air flow rate or the degradation rate of the adsorbent; calculating an overall performance of the adsorbent for the pollutant removal device by combining the weighted performance of the adsorbent at the first location and the weighted performance of the adsorbent at the second location; The performance evaluation system according to claim 8 .

10. the imaging means has sensitivity to a wavelength of light for imaging the salt precipitated on the adsorbent. The performance evaluation system according to claim 1 .

11. A pollutant removal device for removing pollutants from air, comprising: an adsorbent that adsorbs the contaminants; an imaging means for imaging the adsorbent; a control unit; the control unit analyzes image data of the adsorbent captured by the imaging means to acquire feature quantities related to the performance of the adsorbent; Inferring the performance of the adsorbent by inputting the feature into correspondence information that associates the feature with the performance of the adsorbent; predicting the time to replace the adsorbent based on the relationship between the elapsed time since the start of use of the adsorbent and the inferred performance of the adsorbent; Contaminant removal equipment.

12. an adsorbent that adsorbs contaminants; an imaging means for imaging the adsorbent; and a communication device that transmits image data of the adsorbent captured by the imaging means to an information processing device, the information processing device being capable of communicating with a pollutant removal device that removes pollutants in the air via a network, A control unit is provided. the control unit analyzes the image data to acquire a feature amount related to the performance of the adsorbent; Inferring the performance of the adsorbent by inputting the feature into correspondence information that associates the feature with the performance of the adsorbent; predicting the time to replace the adsorbent based on the relationship between the elapsed time since the start of use of the adsorbent and the inferred performance of the adsorbent; Information processing device.

Images