KR20210152613A - Artificial Intelligence Fine Dust Reduction Device - Google Patents

Abstract

The present invention relates to a fine dust reduction device implemented with artificial intelligence. More particularly, the present invention relates to an artificial intelligence fine dust reduction device that controls the sensors, functional devices, and auxiliary devices constituting the fine dust reduction device to not only remove fine dust present in the air but also control indoor temperature and humidity.

Description

Artificial Intelligence Fine Dust Reduction Device

The present invention relates to a fine dust reduction device that implements artificial intelligence, and more specifically, an artificial intelligence fine dust reduction device in which artificial intelligence controls individual sensors and devices constituting the fine dust reduction device to not only purify air but also to control temperature and humidity. It relates to a dust reduction device.

Because the size of fine dust is too small like PM10 (10㎛ diameter) and PM2.5 (2.5㎛ diameter), it causes lung and brain diseases in humans. The ㎛ unit is called a micrometer and indicates ^{10 -6 m.}

PM10 fine dust consists of pollen, mold, dust, various oils generated during food cooking, tire powder generated when moving automobiles, and powder generated in the production process in factories. It is composed of organic compounds and heavy metals that are released when burned.

Currently, fine dust is reduced by using the following technologies.

① Use various types of air purifiers or air purifiers (simply ' air purifier ').

② Use plants with air purifying function.

Air purifiers are largely divided into electrostatic precipitation and filtering methods according to the method of removing fine dust. The electrostatic precipitation method generates negative ions and attaches them to the fine dust with electrostatic force to remove them, while the filtering method uses a fan to suck in air and filter the fine dust through a HEPA filter or Ulpa filter to remove it. The electrostatic precipitation method has a side effect of low efficiency and harmful ozone generation, so air purifiers usually use a relatively harmless filtering method.

Because the structure of the air purifier is extremely simple, it works by pressing the power button, manual mode button, and automatic mode button. Based on this principle of operation, an air purifier installed in a public place should be turned on and closed by the manager during the morning and evening opening and closing hours.

The concentration of fine dust is highest near the source of fine dust, and since it moves through the wind, the concentration of fine dust around the air purifier is lowest. Also, there is a capacity for air purifiers, so there is a limit to the indoor space that the air purifier can handle. have. In addition, the filter type air purifier needs to be replaced on a regular basis. And if you use the air purifier for a long time in an enclosed space, side effects such as lack of oxygen or generation of carbon dioxide and ozone also occur, so you need to ventilate the indoor air frequently.

In addition to the filtering method, water is used as a method to remove fine dust. According to the research data of the Colorado Mining School in the US in April 1976, the effect of capturing fine dust varies somewhat depending on humidity, temperature, pressure, and the electric charge of fine dust, but in general, "water particles with a size similar to that of fine dust" (droplets) have the effect of collecting fine dust." Water particles similar in size to fine dust are called dry fog . Dry fog usually refers to the case where the size of water particles is 5㎛ ~ 20㎛. Wet fog refers to water particles having a size of 20 μm to 50 μm.

Dry fog used to collect fine dust mainly uses water particles with a size of 10 μm, similar to the size of fine dust. These dry fogs do not easily liquefy even when they collide with objects, do not cause short circuits or wet nearby people and machines, and can deliver moisture over relatively long distances.

However, dry fog has a limitation in capturing ultrafine dust with a size of 2.5 μm or less. And when generating dry fog and spraying it around, the water used must be purified or distilled water. If using non-purified water, tap water, occurs baekbun phenomenon of increasing concentration of the fine dust by the mineral and the chlorine component contained in the water. Also, when sulfur dioxide and dry fog meet in a stagnant atmosphere, it is a major cause of smog. And droplets that cause infectious diseases have a size similar to that of dry fog. As a result, there are many difficulties in using dry fog indoors.

The size of fine dust is 0.1㎛ ~ 10㎛, whereas the size of a water molecule (H ₂ O) called water vapor is 2.8Å (Å is called angstrom and has a size of ^{10 -10 m).} Water vapor and particulate matter are joined through the van der Waals forces (van der Waals force) and hydrogen bonding (hycdrogen bond). The van der Waals force is a force acting between all molecules, and it is an electrostatic force that attracts the positive charge of the nucleus of one molecule and the negative charge of the outermost electron of another neighboring molecule. Hydrogen bonding refers to the electrostatic force that occurs between adjacent hydrogen atoms.

Unlike dry fog and water mist, water vapor is pure water molecules that are not polluted. Therefore, even if people breathe in the steam in a sauna or steam bath, it is not harmful to the human body. The change of liquid water into water vapor is called vaporization or evaporation. When water molecules turn to water vapor, they leave with heat. The heat taken is called heat of vaporization or heat of evaporation. The movement of water molecules from a high concentration to a low concentration is called diffusion.

The method of removing fine dust using plants is environmentally friendly, and it is possible to remove harmful substances such as benzene, formaldehyde and trichloroethylene in addition to fine dust. You can enjoy the phosphorus effect. And fine dust and organic compounds absorbed by plants are decomposed and destroyed by microorganisms in the soil.

The method of removing fine dust using plants is slower than an air purifier, despite various advantages. In addition, since plants are a living being, not a mechanical device, nutrients along with water must be supplied periodically.

Meanwhile, interest in artificial intelligence (AI) technology has recently increased in the home appliance market. Nevertheless, the reason why artificial intelligence cannot be used in air purifiers is that “in the case of an air purifier, it does not require any other functions other than supplying and shutting off power and controlling the speed of the fan according to the concentration of fine dust.”

Republic of Korea Patent Publication No. 10-1945314 "fine dust reduction device using fine dust prediction module based on artificial intelligence algorithm" Republic of Korea Patent Publication No. 10-2001782 "A method for controlling the operation of an air purifier based on an input sound using artificial intelligence and an air purifier implementing the same" Republic of Korea Patent Publication No. 10-2040505 "Artificial intelligence-based air purifier and air purifying system including the same"

Colorado School of Mines Article

Through an embodiment of the present invention, an action generated by applying various environmental information to an artificial intelligence algorithm is applied to the fine dust reduction device to provide a fine dust reduction device capable of obtaining the best fine dust reduction effect in relation to the input energy and sleep

Through an embodiment of the present invention, using information obtained by learning values measured from various types of sensors constituting the fine dust reduction device, fans, motors, and devices such as a spray device, the fine dust reduction device is installed for multiple use It is intended to provide a fine dust reduction device that reduces fine dust by controlling functional devices and auxiliary devices according to the environment of the facility and controls the surrounding temperature and humidity.

According to an embodiment of the present invention, it is an object to provide a fine dust reduction device that can reduce noise and minimize energy consumption by installing a fine dust reduction device suitable for a multi-use facility.

The fine dust reduction device according to an embodiment of the present invention has five basic components as follows.

① Functional device and auxiliary device to remove fine dust in a mechanical way

② Environment sensor that detects and collects information related to fine dust around the device

③ Central processing unit that collects information from sensors and individual devices, sends them to the server, and receives instructions from the server to control individual environmental sensors, functional devices, and auxiliary devices (hereinafter referred to as ' sensors and devices ')

④ A server that collects and analyzes information transmitted from an institution related to fine dust, such as a meteorological center, and the central processing unit, and sends the necessary tasks to the central processing unit

⑤ Input/output device that receives tasks requested by the user and outputs fine dust concentration and weather information

① Functional device that removes fine dust refers to a device used for the purpose of directly reducing fine dust. The inside of the fine dust reduction device is divided into a basic set and an extended set. The basic set is the smallest unit that purifies the air containing fine dust. It includes an adsorption filter that can adsorb water particles and fine dust; a filter box that can contain the adsorption filter; one or more spraying devices for spraying the water particles; and a circulation fan that sucks and distributes the air containing the fine dust and the water particles (hereinafter referred to as ' wet air ') into the filter box. And one or more basic sets are mounted on the ramen structure using a coupling member as shown in FIGS. 1 and 2 . Among the basic sets of electronically controlled functional devices are the atomizer and the circulation fan .

On the other hand, the extended set is additionally equipped with a drip tray in addition to the basic set as shown in FIG. 8 . The drip tray provides a function to prevent water that condenses and falls inside the fine dust reduction device and water particles that collide with the filter box from going out to the outside of the device, and can be used as a flowerpot, aquarium, or fountain if necessary. In the case of Figure 2, the fine dust reduction device does not require a separate extension set because the pyramid-shaped filter boxes are symmetrically interlocked with each other.

In addition, ① auxiliary device assists the above functional device, and in order to reduce fine dust existing outside the fine dust reduction device, an exhaust device that sprays purified air and water vapor from the inside of the device to the surroundings, and prevention of freezing of pipes in winter There is a hot water supply device and a hot air supply device that generate natural water or hot water or steam and hot air for rapid vapor evaporation. The auxiliary device is an optional device used to improve the function of the fine dust reduction device and further expand the application range.

Therefore, depending on how the above-mentioned functional device and auxiliary device (hereinafter referred to as ' device ') are combined, high performance can be achieved with little energy and a comfortable indoor environment can be created. For example, if many basic sets are mounted on a ramen structure, its own performance will increase, but energy will be used more and noise will also increase. In order to solve this problem and to widen the area for reducing fine dust, the basic set was slightly reduced, and an exhaust device was separately attached to the top of the ramen structure to naturally adsorb water vapor to the fine dust and drop it on the floor to remove it. The purified air is discharged to the outside. At this time, in order to increase the amount of water vapor used, a hot water supply device that can use steam, hot water, and natural water, or a hot air supply device that generates hot air by coping with a hot hot water supply device if necessary, can be installed under the raven structure. .

② The environmental sensor, which collects environmental information related to fine dust around the device, is a fine dust sensor that can measure the concentration of fine dust from the air in real time and a temperature/humidity sensor ( options) are available. The fine dust sensor and temperature/humidity sensor provide measured data as the magnitude of voltage or amount of current, or the concentration of fine dust in μg/m3 that humans can perceive, Celsius or Fahrenheit, and relative It is also converted into humidity (%) and provided as various communication signals.

And due to the characteristic that fine dust moves with the wind, the concentration of fine dust is collected by collecting data from a number of fine dust sensors scattered throughout the room and averaged using a statistical method. The measurement period of fine dust for calculating statistical values can be set to 0.5 seconds or more than 2 seconds, but it is usually set to be provided in units of 1 second. The concentration of fine dust to be finally obtained is obtained by collecting the measured values provided in units of 1 second, obtaining a normal distribution in units of 1 minute, 10 minutes, and 60 minutes, respectively, and calculating statistics such as the average value and standard deviation. Unlike fine dust, temperature and humidity are calculated per second or averaged per minute, but there is no significant difference. Therefore, temperature and humidity are measured every 10 seconds or every 20 seconds, and the average value and standard deviation are calculated in units of 1 minute, 10 minutes, and 60 minutes.

The remaining environmental sensors are used when other functional devices are additionally introduced or when the environmental characteristics of multi-use facilities are additionally reflected. For example, if a hot water supply device is additionally installed in the fine dust reduction device, it is recommended to add a water level sensor so that water can be replenished smoothly. In order to analyze dust and fine dust coming from outdoors, a proximity sensor that can additionally track the movement or location of people or objects can be used as an auxiliary means. In addition, the environmental sensor includes a noise sensor and an illuminance sensor . And each sensor generates a normal distribution considering its characteristics and obtains statistics.

③ The central processing unit, which collects the measurement values of environmental sensors and individual devices, is equipped with a Wi-Fi or Bluetooth module or Ethernet or LTE or 5G module according to the installed environment for communication with the server, and a timer and main process that can measure time Embedded computer equipped with; and. If the sensor provides measurements in voltage and current, put a separate sensing unit (optional) that can be made into integers or real numbers that can be understood and handled by a person or process. The embedded computer can be installed with more than one sensing unit and control unit depending on the installation location.

The control of the sensor and device is managed for each process in the embedded computer. The fine dust reduction device projects a key performance indicator called KPI (Key Performance Indicator) that indicates the ability of individual devices to maximize the performance of functional devices and auxiliary devices that perform their respective roles. For example, the spray device measures the performance and effect by the amount of spray sprayed per minute (ml/min: amount of water particles). For fans such as circulation fans or exhaust fans, the most important measurement is wind speed or air volume . However, the air volume and wind speed depend on the size of the fan and the shape of the fan, and it is complicated because a separate sensor is required to directly measure it. Therefore, the wind speed is calculated using the fan rotation speed for each model, which is a more generally calculable index, and the air volume is calculated by multiplying the rotation area of the fan. Of course, the wind volume and wind speed may be slightly different depending on the installed conditions at the site, but the insignificant difference is ignored.

And the central processing unit measures the environmental sensors, functional devices, and measurement values for each auxiliary device: for example, the concentration, temperature, and humidity of fine dust, the measurement value of the spray device (spray amount), and the measurement value of the circulation fan (rotation speed or If the number of revolutions is unknown, voltage) is collected and the measured values are collected for a predetermined elapsed time (eg, 1 minute, 10 minute, and 60 minute units) of fine dust together with the original to make a normal distribution to obtain statistics and store them in the internal storage. It is saved and sent to the server at the same time. This is referred to as the first process. And the stored file is used for retransmission after communication with the server is cut off or used to analyze the cause of the failure that was not recorded in the server when a failure occurs. If the embedded computer in the central processing unit does not have enough memory capacity and CPU performance to accumulate data or calculate normal distributions and statistics, it sends all measured values to the server, works on the server, and receives actions from the server for processing.

In general, an embedded computer has a function to control sensors and devices, but if the number of sensors and functional devices to be controlled is large or it is difficult to control because the number of sensors and functional devices to be controlled is too far away from the embedded computer, the fine dust reduction device is separately installed. It can be used separately from the embedded computer by making a control unit.

The central processing unit controls individual environmental sensors, functional devices, and auxiliary devices through a second process of downloading actions from the server in real time and applying them to the process of controlling individual devices in real time. This is called the third process. The central processing unit repeats steps 1, 2, and 3 to reach the environmental improvement goal of the multi-use facility set by the server (referred to simply as ' environmental target').

⑤ Input/output device that receives user's desired task or outputs fine dust and weather information is installed in front of the device and displays fine dust concentration, temperature/humidity, weather forecast, breaking news, etc. A display panel or a separate screen is provided. As shown in FIG. 3 , the direction in which the discharge device faces is the front of the fine dust reduction device. And, if necessary, an LED device can be used to display different colors depending on the concentration of fine dust. This is optional for the device.

In addition, the administrator can instruct the fine dust reduction device by touching the panel one by one for instructions. A website can be provided.

Accordingly, the administrator directly accesses the fine dust reduction device remotely using a smartphone or PC, or changes to the desired mode ( mode refers to a code that classifies multi-use facilities) in an indirect way through a server, or Change the state of the abatement device ( status refers to the work in progress of the fine dust reduction device) or change the initial action (for example, a menu-type action like the previously provided “quiet course” or an action recommended by the server) to another You can change the settings of the fine dust reduction device by changing it to one action. This is called manual management .

If the administrator directly touches the panel or instructs a specific operation by using the smartphone as a remote control, all the instructed contents (mode setting, status change, and action selection) are transmitted to the server through the central processing unit. And the administrator's instruction is delayed until a confirmation message from the server arrives at the central processing unit again. That is, all modes, actions, and states are only possible when synchronized in real time between the central processing unit and the server.

Even in manual management, if the administrator does not give a separate instruction after a certain period of time has elapsed, the automatic management is carried out. Automatic management refers to a method of delegating all function settings to a server or central processing unit. For example, if the rate at which fine dust flows into the room from the outside is faster than the rate at which the fine dust reduction device reduces fine dust indoors, there is a problem in that fine dust is not properly reduced by the previous action set for manual management. Therefore, in this case, the indoor fine dust concentration is measured abnormally high compared to the increase in the local fine dust forecast provided by the meteorological center, or a high-level action under the manager's approval according to the fine dust warning or warning issued by a public institution can be selected.

And ④ server, which collects and analyzes information and sends necessary tasks to the central processing unit, provides time information (current time, manager's commute time, visitor entry time, curfew time, elapsed time since the start of the action, end time of the action, etc.) and all parts information and details (part model, part replacement date, repair date, and cleaning date) used by the fine dust reduction device, and location, regional forecasts of fine dust and meteorological conditions announced by organizations related to fine dust such as the Meteorological Center, the concentration of fine dust transmitted from the central processing unit, the measured value of the circulation fan, the measured value of the atomizing device, and the measured value of the auxiliary device , target energy efficiency, noise level (db), execution time for each detailed step of the action, repetition period and number of repetitions of each step (hereinafter referred to as ' environmental information '), etc. One action is produced using the same algorithm.

Actions are further subdivided into detailed work steps that each functional device and auxiliary device must perform individually . The action calculated by the server is downloaded by the embedded computer in the central processing unit and delivered to the process that manages individual sensors and devices, and simultaneously performs the tasks to be processed according to the schedule. In addition, the measurement values measured by the environmental sensor and the measurement values collected from functional devices and auxiliary devices that perform step-by-step operations become variable values for calculating actions and are transmitted to the server. In this case, it is recommended to use integers rather than real numbers for convergence as much as possible, and to limit the number of decimal places of real numbers to a few decimal places so as not to go below a specific value when obtaining statistical values.

And at this time, depending on which variable is highlighted and which variable is ignored according to the algorithms of various servers, the action to be selected varies. For example, to achieve good performance, energy consumption is increased and noise is also increased. In some cases, this does not meet the environmental standards required by multi-use facilities, so an appropriate balance is required in accordance with laws and regulations. In addition, the temperature and humidity measured by the temperature sensor and humidity sensor use the formula of the dew point temperature and the discomfort index to control the temperature of the water and the amount of spraying of the main equipment like steam, hot water, and natural water so that an unpleasant feeling does not occur.

All without any manual management or division of automatic management measures (fine dust concentration, the rotational speed of the fan, temperature / humidity, illumination, spraying of the spraying device, the number of revolutions of the pump, if necessary, etc.) supervised learning (Supervised learning), or enhance learning It is used as a sample for reinforcement learning.

The present invention creates a comfortable environment for people to live in by controlling temperature and humidity while removing fine dust.

Through one embodiment of the present invention, it is possible to save the time of the manager and reduce the required labor.

Through one embodiment of the present invention, it is possible to reduce fine dust while saving energy according to the installed indoor environment.

1 is a perspective view of an open-type fine dust reduction device;
2 is a perspective view of a closed fine dust reduction device;
3 is a front view of an open-type fine dust reduction device;
4 is a block diagram of the device, the central processing unit and the sensor;
5 is a flow chart of information and control in the central processing unit;
6 is an action input value and a result value;
7 is a perspective view of the discharge device;
8 is an extension set with a drip tray;
9 is a configuration diagram of a central processing unit and a server;
10 is a rotation angle of the left and right rotating plates of the fine dust reduction device parallel operation and the exhaust fan;
11 is a rotation angle of the vertical rotation port of the exhaust fan facing the front door;

Hereinafter, with reference to the accompanying drawings, the embodiments of the present invention will be described in detail so that those skilled in the art can easily carry out the present invention. However, the present invention may be embodied in several different forms and is not limited to the embodiments described herein. In order to clearly explain the present invention in the drawings, parts irrelevant to the description are omitted, and the same reference numerals are assigned to the same or similar components throughout the specification.

1. Components of a fine dust reduction device

The device for reducing fine dust according to an embodiment of the present invention can be divided into an open type fine dust reducing device as shown in FIG. 1 and a closed fine dust reducing device as shown in FIG. 2 . The open type fine dust reduction device is not a basic set, but an extended set with a drip tray 300 as shown in FIG. 8, and there is only a difference in processing capacity compared to the closed type fine dust reduction device, but there is no significant difference in function. Therefore, the present invention intends to refer to the open-type fine dust reduction device as a reference.

The basic set of the open-type fine dust reduction device includes an adsorption filter 311 capable of adsorbing water particles and fine dust, a filter box 310 capable of containing the adsorption filter; one or more atomizers 601 for spraying the water particles; A circulation fan 301 for sucking and distributing the air containing the fine dust and the water particles (hereinafter referred to as ' wet air ') into the filter box; this is a basic set and is a coupling member inside the ramen structure 200 (201) is used to attach. In the basic set, the water particles are attached to the adsorption filter to narrow the filter gap, and also serve as a kind of wet air purifier that filters fine dust by capillary action.

Among the basic sets, the functional devices that can be controlled by electronic signals are the atomizer and circulation fan. In general, in ^{the case of a small office of about 50 m 2} , one of the basic sets is mounted on the ramen structure. However, when the area is 50 m ² or more, the ramen structure is equipped with a plurality of the basic sets. Of course, this standard is based on its own model, and if it is a different model or a different model, the standard may also be different.

In FIG. 3, the fine dust reduction device is equipped with an extended set, not the basic set, and additionally a heating wire or a heating roll (hereinafter referred to as ' heating wire ') is mounted, and heat to supply purified water to the spray device 601 In order to supply hot air into the hot water supply device 600 and the filter box 310, a hot air supply device 402 equipped with a heating wire (not shown), a vent 401 and a hot air fan (not shown) is an auxiliary device. It may be mounted on the lower portion of the furnace ramen structure 200 . However, in the case of the fine dust reduction device according to another embodiment of the present invention, the hot water supply device can cope with a water pipe wound with a heating wire together with a solenoid valve. The solenoid valve provides the function of opening the valve when an electrical signal is received and closing the valve when there is no electrical signal. The solenoid valve 602 present in FIG. 3 is used for replenishing water to the hot water supply device 600 in the form of a water storage tank connected to a water pipe regardless of the above-mentioned.

The reason for using a heating wire is to increase the production of water vapor by making natural water or hot water or steam, to prevent freezing of pipes in winter, and to control the temperature and relative humidity around the device. And if you use the hot water supply device 600 manufactured in the form of a water storage tank as shown in FIG. 3, since water may be insufficient at any time, a water level sensor (not shown) may be additionally attached, and if necessary, the water temperature can be measured. A separate water temperature sensor (not shown) can be attached. In general, even if a water temperature sensor is not used, the temperature of water varies depending on the temperature of the natural water used and the number of lines and specifications of the heating wire. If you write a program that supplies power periodically by using the obtained data as a function, you can adjust the desired water temperature at any time.

And, the hot air supply device 402 is attached to the adsorption filter 311 using hot air in an environment where the hot and hot water supply device 600 cannot generate hot water or steam depending on the indoor environment in which the fine dust reduction device is installed. Used to evaporate moisture. Steam or high heat is used for the purpose of sterilizing bacteria or mold adhering to the adsorption filter and accelerating the evaporation of moisture attached to the adsorption filter. In addition, auxiliary devices used in the present invention include a discharge device (a discharge fan, a left and right rotation plate, and a set of vertical rotation ports). In addition, the heating wire used in the hot water supply device and the hot air supply device uses the temperature of water or the temperature of the heating wire as the unit of measurement and collects the measured values for a given time (for example, 1 minute or 10 minutes) based on the temperature measured every second. Calculate the statistics of the normal distribution (if the distribution is non-normal, make it normal). If the temperature cannot be measured, deal with it with the amount of power supplied to the heating wire for a given time or the cycle of power supply.

The fine dust reduction device may be managed manually by an administrator (hereinafter referred to as ' manual management ') or automatically (hereinafter referred to as ' automatic management ') using an embedded computer. In addition, the fine dust reduction device can be changed to automatic management at any time by setting a timer even if the manager performs manual management. Of course, even in the automatic management state, the administrator can artificially manipulate individual devices manually. In other words, manual management takes precedence over automatic management for all tasks .

For automatic management, the fine dust reduction device uses various types of environmental sensing sensors, functional devices, and auxiliary devices (hereinafter, collectively referred to as ' sensors and devices ') as shown in FIG. 4 . The most important sensor among the environmental sensors is the fine dust sensor 107, which is a sensor capable of measuring the concentration of fine dust from the surrounding air in real time. The remaining environmental sensors are used as sensors that support the fine dust sensor, or are used for data collection, statistics, and measurement of environmental changes around the device. And the concentration of fine dust measured by the fine dust sensor is transmitted to the embedded computer 102 in the central processing unit 100 .

In addition, the environmental sensors include the essential fine dust sensor, optional temperature/humidity sensor, water temperature sensor, water level sensor, distance proximity sensor, illuminance sensor, power meter, and noise sensor. The power meter is used for the purpose of obtaining statistics or charging according to the rental. As a temperature sensor and a humidity sensor, a sensor capable of simultaneously measuring temperature and humidity (hereinafter referred to as a ' temperature/humidity sensor ' 108) is mainly used.

The fine dust sensor 107 measures the number or mass of fine dust spread in the atmosphere using a light scattering method or a beta-ray absorption method and a gravimetric method and provides it as a voltage or an amount of current, but most of the fine dust The sensor uses an internal circuit to convert an integer or real number in μg/m3 that can be detected by a human again. to provide.

In addition, most environmental sensors, including temperature/humidity sensors, provide environmental changes (eg temperature and humidity) in the amount of voltage or current, like fine dust sensors, or can be detected by humans or embedded computers. It is converted into a numerical value such as an integer or real number and provided through various communication signals. Therefore, if the sensor provides the measured values with voltage and current, since the embedded computer cannot directly receive it, a separate sensing unit 103 is placed in the central processing unit 100 as shown in FIG. 9 to convert it into an integer or real number.

The feature that the environmental sensor and the central processing unit can send and receive messages through wired and wireless communication is the advantage of separating one or more various types of sensors from the fine dust reduction device and attaching them to various places indoors to enable more precise measurement. provides In particular, fine dust is moved by the wind, so it is better to install a fine dust sensor inside a fine dust reduction device that is relatively well purifying, so it is better to install it in a place where there is a lot of people or where fine dust is expected to enter and occur. It is better to install The fine dust reduction device can identify the source of indoor fine dust from the fine dust concentration measured by one or more fine dust sensors, and more fundamentally identify and remove the source of fine dust .

The concentration of fine dust in the area affected by the fine dust reduction device is calculated based on statistics by receiving signals from fine dust sensors transmitted from various places. The measurement period of the fine dust sensor for calculating statistics varies depending on the setting, but is usually set to be provided in units of 1 second. And finally, the concentration of fine dust to be obtained is collected in units of 1 minute, 10 minutes, and 60 minutes, respectively, by collecting the measurement values of all the fine dust sensors provided in units of 1 second (the time interval can be changed according to convenience). Create a statistic by creating a normal distribution. The normal distribution and the statistic are also used for the purpose of discriminating abnormal values, normal values, and noise, and determining whether a device is faulty.

According to the World Health Organization (WHO) standard, the fine dust concentration is 0 ~ 15 ㎍/㎥, and the fine dust concentration is 0 ~ 30 ㎍/㎥. The normal level of ultrafine dust is 16 ~ 35 ㎍/㎥, and that of fine dust is usually 31 ~ 80 ㎍/㎥. The badness of ultrafine dust is 36 ~ 75 ㎍/㎥, and the bad of fine dust is 81 ~ 151 ㎍/㎥. The very bad of ultrafine dust is 76 ㎍/㎥ or more, and the very bad of fine dust is 151 ㎍/㎥ or more.

In the present invention, grades such as good, normal, and bad of ultrafine dust or fine dust can be recognized as entry steps for starting and ending the fine dust reduction device and for strengthening/weaking individual functional devices. That is, the fine dust grade of the present invention is a unit of measurement and an environmental improvement goal (or environmental goal ) pursued by the mode, and is also an action suggested by various algorithms used for reinforcement learning to solve environmental problems. is used as a variable value for calculating .

For example, as in the example of Figure 6 (Variable 1): When the change in the concentration of external fine dust per hour is 2 μg/m3 or less, (Variable 2): On average, one person moves within a 10 meter radius of the surrounding area per minute (Variable 3 ): In the underground shopping mall (Variable 4): The average concentration of fine dust measured for 1 minute is 35 μg/m3, (Variable 5): The average indoor temperature measured for 1 minute is 10℃ (Variable 6): Measured for 1 minute The average relative humidity is 40%, and (Variable 7) is in a silent state, (Variable 8): the rotation speed of the circulation fan is 1000 RPM, and (Variable 9): The spray device is in a state where there are no water particles currently being sprayed If it is (non-operational state), (Action 1): To perform the work of removing fine dust for 10 minutes (Action 1-Step 1), make the silent state to a normal state and (Action 1-Step 2): Apply power, (Action 1-Step 3): 1600RPM circulation fan rotation speed and (Action 1-Step 4): Circulation fan operation time is 10 minutes, (Action 1-Step 5): Spraying device for 1 minute Spray water particles at 10ml/min and work during the action time with a 1-minute rest cycle. As a result after 10 minutes (calculation result 1): The average concentration of fine dust measured in the previous 1 minute is lowered to 25㎍/㎥ (calculation result 2 ): The average temperature measured in the previous 1 minute rises to 12℃, and ( Calculation result 3 ) : The average relative humidity measured in the previous 1 minute rises to 60%. And (Calculation Result 4): There is no change in the remaining variable values. The above steps are one action and each device is executed at the same time. The measurement of the effect of an action is judged based on the measurement value collected every second and the normal distribution generated by the measurement value collected within a set time. Accordingly, if the fine dust concentration rises even before the action ends, a more powerful fine dust reduction action is triggered.

In the above example, examples of additional auxiliary devices are omitted as an example for better understanding. The aforementioned spraying device 601 may be manufactured with the following parts.

● Ultrasonic humidifier module, hose and fan that uses ultrasonic waves to create water mist

● A nozzle that can spray water particles and a water pump (or an air pump and a pressurized pump)

● Devices that can generate other water particles, such as artificial waterfalls or watermills

As mentioned above, the most important measurement value of the spray device is the spray amount of water particles sprayed per minute (ml/min). And the spray amount is obtained by inversely calculating the frequency (Hz) of the ultrasonic module, the rotation speed or voltage of the pump, and the cycle of PWM through an experiment rather than directly measuring it. Since the frequency or rotation speed may increase or decrease depending on the action time, it cannot be evaluated collectively. maximum value, minimum value, etc.).

The most important measurement value of the circulation fan is wind speed (m/s) or air volume (m ³ /s). However, while the motor measures the number of revolutions (rpm), the air volume and wind speed depend on the size and shape of the fan, and if you want to measure it directly, a separate sensor is required. Therefore, by using the fan rotation speed (rpm), which is a more generally calculable index, the wind volume and wind speed are calculated inversely for each model. And since the rotation speed of the circulation fan is also adjusted according to the influence of numerical values such as noise rather than rotating at the same speed during the action time, a normal distribution is obtained for a given time and statistics are calculated.

As in the example above, if the concentration of fine dust is reduced from 35 ㎍/㎥ to 25㎍/㎥ by one action operation, the 25㎍/㎥ fine dust grade in the multi-use facility is 35㎍/㎥ fine dust under all the same conditions. A variable for calculating an action goal that can be improved with one action from a rating and another action to reach the mode's environmental goal (environmental improvement goal that can be reached in a multi-use facility at the same time, such as fine dust or quiet) be the value Of course, the action differs depending on the subway or library's multi-use facility, even if they have the same environment value.

The fine dust reduction device may be equipped with an exhaust device that sprays the air purified inside the device to the surroundings in order to reduce fine dust existing outside the device. As shown in FIG. 7, the exhaust device includes an exhaust fan 302 that controls wind speed and air volume with electronic signals, and is attached to the bottom of the exhaust fan 302 and has an open hemispherical vertical rotation sphere 306 and the vertical rotation opening (306) is attached to the bottom and is composed of a set of left and right rotational plate 305 that is installed at the top of the ramen structure.

In addition, the central processing unit 100 determines the location where the fine dust sensor showing the highest concentration among the fine dust concentrations measured by the fine dust sensors installed in the vicinity according to the action of the server is installed, and the left and right rotation plate 305 and the vertical rotation port Moving to 306, the purified air is discharged to the vicinity where the fine dust sensor is installed, and the fine dust is removed using water vapor.

The vertical rotation tool and the left and right rotation plate perform a necessary rotation operation using a cross shaft gear or a worm gear and a step motor or a servo motor. The rotation angle can take a range of 180 degrees left and right or -180 degrees counterclockwise with respect to a) the front of the raven structure as shown in FIG. So b) rotate upwards and when the front of the exhaust fan is in a straight line with the raven structure, it is recognized as -90 degrees, c) rotates downward so that the front of the exhaust fan is in a straight line with the raven structure and it becomes 90 degrees when facing downward. Realistically, the above-mentioned angle becomes the theoretical rotation angle, and the realistic angle is less than that angle.

More precisely, the exhaust fan 302 delivers the purified air in a cross-saturation form based on the place where the fine dust reduction device is installed even if a separate fine dust sensor 107 does not exist in the vicinity as shown in FIG. 10 of the exhaust fan 302 . It is schematically discharged from the center at an angle of 45 degrees left and right or discharged at an angle of 35 degrees downward from the horizontal to discharge a specific location (doors, entrances, and windows of a building) as shown in FIG. 11 . The purified air discharged from the bottom comes up again because the temperature is higher than the air coming in from the outside. Therefore, if there is no movement of people at the door, it is necessary to discharge the purified air while rotating up and down in order to circulate the heated air from the top down again.

And the central processing unit 100 is the measured value of the exhaust fan (here it refers to the number of rotations of the exhaust fan) and the measured value of the left and right rotation plate (if it is a fixed position, the current angle or if it moves left and right, the rotation angle per second and rotation per minute) Period: The rotation period per minute can be calculated by the rotation angle per second) and the measured value of the vertical rotation ball (the current angle if the position is fixed or the rotation angle per second and the rotation period per minute if it moves up and down) to the server, and the According to the action received from the server, the direction is determined using the left and right rotation plates and the vertical rotation port, and the wind speed (m/second or m/minute) or air volume (m ³ /second or m) for each model through the rotation speed of the exhaust fan. ³ /minute) is converted to use. And if the angle of the left and right rotation plate and the vertical rotation ball moved while performing the action does not change, the measured value for the positioned angle is not transmitted.

The method for the central processing unit to receive the measured values from the fine dust sensor and the temperature/humidity sensor has already been mentioned. In addition, a water level sensor (not shown) that is additionally mounted to the hot water supply device 600 is attached to a specific position in the case of a contact type, so that electricity does not flow when water is insufficient, so that 0 V is notified to the central processing unit. And when the water goes above a certain level, electricity flows, so the water level sensor provides an electrical signal of 1V or more to the central processing unit. More complex water level sensors include an ultrasonic sensor and an infrared laser sensor that can directly measure the water level. And if there is a water shortage, the central processing unit notifies the manager of the water shortage through the server.

The proximity sensor 110 measures the return time after sound waves or infrared rays collide with an object using an ultrasonic sensor or an infrared laser sensor, and converts the approach distance of the object into centimeters or millimeters, and the central processing unit 100 provided to Accordingly, the proximity sensor calculates a normal distribution of the number of people who stayed within the measurement range for a given time. The noise sensor (not shown) and the illuminance sensor (not shown) also provide values measured in voltage and current depending on the volume and light quantity, or in decibels (db) and lux (Lux) that can be recognized by humans. The noise sensor and the illuminance sensor also collect the measured values within a given time to calculate a normal distribution.

The above-mentioned measured values are stored in the internal storage, and for further analysis, the raws of all measured values measured by sensors and devices and statistics about the normal distribution (non-normal distribution is transformed into a normal distribution) (e.g. For example, average value, standard deviation, maximum value, minimum value, cumulative value) and the log recorded by the central processing unit are attached to their product code and transmitted to the server 400 . In addition, all files stored in the central processing unit are kept for several days or months in consideration of the storage capacity and deleted periodically. The deletion cycle of individual files differs according to their importance. In addition, the archived file is used for retransmission in the process of recovery after a failure of the server and communication module or to analyze the cause of the failure that is not recorded in the server when a failure occurs.

As a result, in the present invention, environmental information is the basic data for calculating the action for controlling sensors and devices, and includes time information, information and details of parts constituting the fine dust reduction device, legal regulations required by the environment in which the fine dust reduction device is installed, or The increase/decrease (or movement) number of people present within the surrounding radius measured by the proximity sensor (the more people move, the more scattered dust increases), the local forecast of fine dust and weather announced by the Meteorological Center ( The weather is good or cloudy, or the outdoor fine dust concentration increases or decreases depending on it), the indoor fine dust concentration measured by the fine dust reduction device, the temperature, the humidity, and the measured value of the circulation fan (rotation speed, PWM period or voltage, etc.), the measured value of the atomizer (spray amount, the number of rotations of the pump or the frequency of the ultrasonic module), the measured value of the discharge device (the number of rotations of the fan, the rotation angle and rotation period of the left and right rotation plates, the rotation of the vertical rotation tool) angle and rotation period), measured value of hot water supply device (water temperature, heating wire temperature), measured value of hot air supply device (heating wire temperature, fan rotation speed), required energy efficiency (wattage), noise generation (db) , illuminance (lux), the execution time for each detailed step of the action, the number of repetitions and the repetition period, etc., the action to be selected differs depending on which variable is highlighted and which variable is ignored.

The action is further divided into detailed work steps to be performed by the central processing unit, the functional unit, and the auxiliary unit according to their respective roles . And the above-mentioned sensor and device measurement values become variables for calculating actions again.

The central processing unit 100 uses a Wi-Fi or Bluetooth module or Ethernet or LTE or 5G module for communication with the server 400 according to the environment in which the fine dust reduction device is installed. Each sensor and measurement value provided by each sensor inside the fine dust reduction device becomes an independent variable, and each independent process in the central processing unit manages the independent variable. Information passing between processes uses internal communication or shared memory. The processes executed inside the central processing unit 100 leave all self-managed contents in log files, and the log files are also sent to the server to determine whether the actual processes performed properly or whether individual parts in the central processing unit are abnormal, and Analyze in real time whether there was a third party hacking attempt.

In addition, the central processing unit 100 is the area received from the server, such as the display panel 109 installed in front of the device by the users, including the administrator, or a large screen (not shown) installed in a separate location, fine dust concentration, temperature/humidity around the device. It is output through the input/output device so that the fine dust forecast, local weather forecast, breaking news, etc. can be viewed. Here, the temperature/humidity is output when an optional sensor is installed. And, if necessary, an LED device can be used to display different colors depending on the concentration of fine dust. This is optional for the device.

The administrator can give instructions by touching the panel for the necessary instructions for the fine dust reduction device. Separately, an app is provided for convenient control like a remote control with a smartphone or a website that can be used by accessing the Internet site with a web browser can provide

Accordingly, the administrator directly accesses the fine dust reduction device remotely using a smartphone or PC, or changes to the desired mode (mode refers to the setting method according to the multi-use facility) or an initial action ( For example, the setting of the fine dust reduction device can be changed by changing a menu-type action like the previously provided “quiet course” or an action recommended by the server) to another action recommended by the server.

If the administrator directly touches the panel or instructs a specific operation by using the smartphone as a remote control, all the instructed contents (mode setting, status change, and action selection) are transmitted to the server through the central processing unit. And the administrator's instruction is delayed until a confirmation message from the server arrives at the central processing unit again. In other words, application of all modes, actions, and states is possible only when synchronized in real time between the central processing unit and the server.

In general, embedded computers have a function to control sensors and devices. However, if the number of sensors and devices to be controlled is large or distant from the embedded computer, the fine dust reduction device separates the control unit 104 that performs the following functions separately from the embedded computer 102 as shown in FIG. can be installed.

● Apply power to each sensor and device just before use, and turn off the power to each sensor and device when not in use. A circuit that supports these functions is called a relay board.

● A PWM (Pulse Width Modulation) control fan and a PWM control pump can be used to separately control the rotation speed of the powered fan and pump. PWM control fans and pumps provide a function to control the rotation speed of the fan according to the voltage cycle, or to have the central processing unit know the rotation speed of the fan in reverse. A circuit that supports these functions is called a PWM driver module.

● If a PWM control fan or pump is not used, a circuit to boost or step down the required voltage together with a fan and a pump (usually a brushless motor or brush motor) that can control the rotation speed of the fan or pump by the voltage difference. use. For example, in the case of a specific fan, if a normal voltage of 12V is applied, it rotates at 1800RPM, but if a low voltage of 9V is given, it rotates at a speed of 100RPM. In this case, the number of rotations of the fan and motor is used by making a function by identifying the number of rotations versus the voltage provided by each model.

● If you use a step motor, use a step motor driver that can control the rotation speed and direction of a separate motor.

● The hot water supply device and hot air supply device using a heating wire adjust the voltage, current, and power supply cycle so that the temperature of the heating wire is kept constant.

As shown in FIGS. 1 to 2 , the temperature and humidity of the purified air measured from the temperature/humidity sensor 108 attached to the top of the ramen structure is the basis for calculating the dew point temperature. The dew point temperature is the temperature at which air is saturated and water vapor condenses. And the fine dust reduction device controls the spray device and the hot and hot water supply device or the hot air supply device so that the dew point temperature of the purified air is at least equal to or higher than the local outdoor temperature announced by the meteorological center. The function of removing dust, hot and humid purified air and fine dust coming in from the outside by colliding near the door or front door as shown in Fig. to provide.

The formula for calculating the dew point temperature is as follows.

Dew point temperature = 243.12 x {ln(relative humidity/100) + (17.62 x temperature}/(243.12 + temperature)} / (17.62 - {ln(relative humidity/100) + (17.62 x temperature}/(243.12 + temperature)) }

The above relative humidity and temperature are measured by a temperature/humidity sensor located near the exhaust fan. Since there is a difference from the location where the actual collision occurs, there may be a slight difference in temperature and humidity, so adjust it with an appropriate correction value or ignore it.

In addition, the temperature and humidity of the purified air are 70 or less to prevent unpleasant sensations by additionally calculating the discomfort index in summer or spring and controlling the functional device and the exhaust device to remove the humidity in the purified air that a person breathes directly. (For example, if the discomfort index is between 70 and 83, some people feel unpleasant, and if it is over 83, most people feel unpleasant.) And in winter, the wind speed of the circulation fan and the wind speed of the exhaust device and the power supplied to the hot water supply device 600 or the heating wire in the hot air supply device so that the temperature of the purified air using the sensible temperature is equal to or higher than the outdoor temperature It provides a function that controls the supply so that the user entering through the door feels warm (higher than the outdoor temperature) when in contact with the purified air.

The formula for calculating the discomfort index and the sensible temperature is as follows.

Discomfort index = 9/5 x temperature - 0.55 x (1 - relative humidity) x (9/5 x temperature - 26) x 32

Feeling temperature = 13.12 + 0.6215 x temperature - 11.37 x wind speed ^0.16 + 0.3965 x wind speed ^0.16 x temperature

{Here, the unit of wind speed is km/h, and the wind speed measured through actual measurements at the time of installation of the fine dust reduction device at the location of the door or window is used to create a function using the rotation speed of the exhaust fan.}

2. Mode and state management

The fine dust reduction device manages the detailed state as follows. In the present invention, status management is explained by changing the current work content, which is difficult to explain to an administrator or developer only with numerical values, into the expression of bit flags called status.

① Sleep state : This refers to a state in which power is supplied only to the embedded computer to minimize power consumption. In the sleep state, power supply to the sensing unit, the control unit, the sensor, the device, and the display panel is stopped. Even in the sleep state, if the outdoor temperature in winter falls below zero or a freeze warning is issued, it will wake up periodically to prevent freezing of the pipe and make it into a measurement state, preheating state, or hot air state.

② Measurement state : It refers to the state in which the concentration of fine dust or temperature/humidity is measured using the environmental sensor. The sensor to measure can be selected according to conditions and needs. For example, even in the measurement state, the water level sensor does not measure separately unless it is in the injection state.

③ Output state : It refers to the state in which the information the user wants to see is output through the display panel.

④ Silent state : It is a state in which sensors and devices are operated without noise. Noise-free refers to the noise (dB) below the allowable level at night as a residential standard.

⑤ Emergency : To achieve a high reduction effect, the functional devices and auxiliary devices are fully operated at the highest level. When the group leader issues a crisis alert or emergency alert, the content of the alert is transmitted to the manager's terminal and is operated with the manager's consent. If an emergency state is permitted according to the mode, no separate administrator approval is required.

⑥ Fan operation condition : The circulation fan and the exhaust fan are operated at the same time as the fan rotation speed is within the range allowed by room noise. It is mainly used to dry the adsorption filter containing moisture when using the fan alone or with hot air.

⑦ Aiming state : It is used to align the direction of the exhaust fan in a specific direction using the left and right rotary plates and up and down rotary tools.

⑧ Preheating state : It is used to preheat to a certain temperature by supplying power to the inside of the hot water supply device.

⑨ Steam state : It refers to the state in which steam is sprayed through the atomizer. Because steam is dangerous, it is managed in a separate state.

⑩ Spray state : It refers to the state in which water particles are sprayed at room temperature through the spray device. When it is a natural number, the power provided to all heating wires is cut off. On the other hand, hot water supplies power periodically to keep the temperature of the heating wire constant.

⑪ Hot air condition : It refers to the condition in which the hot air supply device is heated and the hot air is generated and dissipated in a specific direction through the tuyere. If there is no hot air supply device, the hot air state is omitted.

⑫ Inspection state : It is a state to check whether the sensor and device are normal or abnormal. The fine dust reduction device initializes each sensor and device according to the self-inspection action and performs a test to determine whether it is normal or abnormal.

⑬ Abnormal state : Abnormal state is the state before conversion to maintenance state, which means that maintenance is required due to a failure. Even in an abnormal state, if a failure occurs in one basic set and all other basic sets operate normally, the two states, the normal and abnormal states, are shared, and all operations operate normally until transferred to the maintenance state.

⑭ Maintenance status : Refers to the status in which parts are being replaced or the inside of the device is being cleaned.

⑮ Dangerous state : It refers to a state in which a disaster is predicted to have occurred, such as a fire.

The above state derives only the most basic state. And the above state is used by making one state or a state in which 2 to 6 are bundled (for example, fan operation state + hot air state or measurement state + fan operation state + injection state + output state = normal state). And even in the same state, it is classified in detail according to the action calculated according to the mode and artificial intelligence algorithm. For example, the fine dust reduction device is in the same fan operation state, but the fan rotation speed is different depending on the mode and action, and even in the same spraying state, the number of rotations of the water pump 603 or the number of vibrations of the ultrasonic module (not shown) is distinguished This makes the amount of sprayed water particles different. If the adsorption filter contains 100% moisture and can no longer be adsorbed, the spray device no longer needs to spray water particles. At this time, the injection state also appears and disappears repeatedly for each time period.

Multi-use facilities such as subway stations, bus terminals, airports, schools, auditoriums and libraries have different usage environments and different sources of fine dust.

The fine dust reduction device is an environmental target that meets the set mode as follows, for example, in the case of library mode, as an environmental standard, 'fine dust is good (fine dust 15㎍/㎥ or less)' or 'humidity comfortable (relative humidity 50)'%)','quiet' or 'quiet' (noise less than 50db), each functional device by creating one or more actions (each action has an action goal distinct from environmental goals) that finds the optimal way to reach and control auxiliary devices. In the present invention, the environmental target generally complies with the laws and regulations of multi-use facilities, and refers to the aim of achieving a WHO fine dust rating of good.

● Library mode refers to a space that requires a quiet environment, such as a library or cathedral. In the library mode, even if the fine dust level is measured as bad, the functional and auxiliary devices are operated in a silent state when opened. These places opt for the action of a quiet, prolonged clean-up rather than a noise-causing quick clean-up.

● Public place mode purifies the surrounding air in a short time by activating functional devices and auxiliary devices according to the action set according to the high and low fine dust concentration while complying with the indoor noise standards in places frequented by people. If commuting exists or in the case of a public place where access to the building is controlled at night time, the sleeping state or measurement state is made during the control period. In other words, if there are no people, operating the device becomes meaningless no matter how high the concentration of fine dust, and in this case, energy efficiency comes first.

● In school mode , even if the fine dust level is measured as bad according to the class time of 90 minutes followed by 10 minutes of rest or 50 minutes of lessons followed by 10 minutes of rest, fine dust During the 10-minute break, it performs the task of removing fine dust in a normal state according to the concentration of fine dust. Also, for classes that take place suddenly without a schedule, such as a vacation, by setting a time in advance, functional devices and auxiliary devices are operated in a silent state to purify fine dust.

● In the exhibition room mode , even if the fine dust grade is measured as bad according to the exhibition time period (9 am to 6 pm), the functional and auxiliary devices are operated in a silent state to purify fine dust. In addition, during times other than the opening hours when there are no visitors, it is made into a sleep state or a measurement state to save energy.

● User mode removes fine dust through manual management by the user. Generally, it refers to a mod created artificially by an administrator. In this case, a problem arises in deriving an optimal action due to insufficient training samples.

The mode mentioned above is the most important environment option and is set together with the product code in the environment file that is executed when the central processing unit reboots. The product code is a unique number that identifies each device (external to the device, a barcode or QR code is used to recognize the number), and also includes GPS (latitude and longitude) information to determine the installation location related to maintenance. connected This GPS information informs the fact that the installed location is a school, library, subway station, etc. connected to a map or map. Therefore, in the above-mentioned action for each mode, the KPI target for each step required for each device at the present time is designated and delivered to the process executed by the central processing unit.

Fine dust reduction devices installed in schools or libraries additionally attach a noise sensor (not shown) that measures noise separately to control each functional device so as not to exceed a certain level of noise in accordance with the law (noise and vibration standards) Needs to be. As mentioned above, even if there is no noise sensor, it is possible to comply with the noise standards required by laws and regulations by adjusting the rotation speed and spray volume of the fan based on the noise of the circulation fan and spraying device measured for each model. However, since noise is generated by adding up noise from other nearby devices or construction sites, it depends on the location where the fine dust reduction device is installed. If more precise measurement is required, a noise sensor is required.

And if the exhibition room mode is also necessary like the school mode, use the illuminance sensor (not shown) or the proximity sensor 110 for more than 10 minutes in the presence or absence of light, or when there is no person or when there is a normal state or sleep state. can be maintained as In other words, even during working hours, depending on the presence of people and objects, or the illuminance detected by the illuminance sensor, the amount of light that can distinguish objects based on the brightness of warehouse lights (10-20 lux) or emergency lights (1-10 lux) Depending on the presence or absence, the power of the functional device, the auxiliary device, and the input/output device may be applied or cut off through the control unit. Of course, even if there is no illuminance sensor or noise sensor, the fine dust reduction device can be switched to a sleep state, a measurement state, a normal state, and a silent state periodically or as needed depending on the working hours or curfew hours operated by the institution. For reference, the transition from one state to another ( referred to simply as 'state transformation ') is accomplished through one action.

And the above-mentioned state is when the action of “switch to another state from the server” comes down, or when the state is manually changed from the administrator’s smartphone or PC to the Internet, or when the preset time has elapsed, the action A transition can be made from the current state to another state through Of course, the administrator can select and change the mode or action, and the setting action and mode for initialization before connection with the server after booting are recorded in the environment file in the central processing unit. The state changes every moment, so it is not stored as a separate file.

The fine dust sensor 107 generally measures the fine dust concentration once per second. For the measured fine dust information, a normal distribution is calculated from the original fine dust measurement values in units of 1 minute, 10 minutes, and 60 minutes, and statistics are obtained.

The fine dust reduction device is “one-minute average value of ultra-fine dust + 1-minute standard deviation of ultra-fine dust > good ultra-fine dust (15㎍/㎥)” or “one-minute average value of fine dust + 1-minute standard deviation of fine dust” > When fine dust is good (30㎍/㎥)”, turn on the power of the functional device and start operation, or start automatically according to the time when the external fine dust concentration is expected to be bad in the meteorological center. And if “10-minute average value of ultra-fine dust + 10-minute standard deviation of ultra-fine dust <= fine for ultra-fine dust (15㎍/㎥)” and at the same time “10-minute average value of fine dust + 10-minute standard deviation of fine dust < = Fine dust good (30㎍/㎥)”, cut off the power to the circulation fan and the atomizer and stop the operation of the fine dust reduction device. However, if necessary, the fine dust reduction device can be operated when the “concentration of ultrafine dust > 10㎍/㎥” measured every second.

If the administrator sets separate operating conditions (conditions to apply and cut off power), the conditions are followed. In other words, comparison operators (<, >, <=, >=) and logic for variables such as fine dust concentration, temperature, humidity, noise, presence of people around, time information, local fine dust forecast, and local weather forecast Use the operator (or, and) to arbitrarily set the operating conditions (e.g., “noise < 40db or temperature < 20℃”) Block.”, it can be set automatically using the latest character recognition technology.

The administrator directly accesses the fine dust reduction device remotely using a smartphone or PC, or changes to the desired mode (mode can only be changed by the administrator) in an indirect way through the server, or The automatic management setting of the fine dust reduction device can be changed to manual management by changing the state or changing the currently executed action to one of the actions recommended by the past server. The action may change the state of the fine dust reduction device from time to time as needed.

However, in the present invention, all actions or state management is set to 24 hours . Therefore, even if the administrator sets manual management, even if all the step-by-step goals set in the action are reached or the action is not reset to manual management after a certain period of time has elapsed, the setting is transferred to automatic management. For example, in the case of a multi-use facility with commuting time, even if a specific action is selected for manual management, the setting may be changed to automatic management after midnight, and the action may be changed. And if the light flickers for more than 10 minutes or there is no one around, the setting goes to automatic management under the instructions of the server. The concept of artificial intelligence fine dust reduction device is “ The action of manual management is performed on the assumption that a person exists around the device .”

In addition , even with manual management , the rate of fine dust entering the room from the outside is faster than the rate at which fine dust is reduced indoors, so there is a problem that the fine dust is not reduced properly with the previous action, or a fine dust warning or warning is issued at the weather center. If issued, actions at a higher level beyond the mod's allowable range can be selected with the permission of the administrator.

And the administrator can delegate all settings to the server or central processing unit through automatic management. Accordingly, the server 400 sends an appropriate action to the fine dust reduction device under normal circumstances. For example, even in the same mode, it is set to sleep or measured during lunchtime (from 12:00 am to 1:00 pm), at night when there is no one around, or during quiet holidays when the entrance is closed, so that unnecessary energy is generated even if the concentration of fine dust is high. will reduce consumption.

The central processing unit 100 must be equipped with an algorithm for detecting noise received from the sensor and the device. For example, the easiest algorithm is to threshold “2 x standard deviation” or “3 x standard deviation” according to the 68-95-99.7 rule of normal distribution (or 3 sigma rule) and usually within “mean ± threshold” of the recognition of the measured value to an appropriate value or the mean value theorem having range, obtain the gradient from the mean value is calculated continuously as a (theorem Roll) can be calculated by an appropriate amount to "tilt x mean ± threshold value" and the extent Values outside the range can be recognized as noise Since the measured noise is related to the reliability problem of the product, it can be used as an index to determine the possibility of failure.

However, the problem is that when the measured value measured in units of 1 second or 0.5 seconds is measured higher than the threshold or lower than the threshold at any moment, the value is recognized as noise and stored in a separate storage rather than deleted to prevent noise or appropriateness. Depending on whether it is a number, it should be discarded or included in calculating the mean and standard deviation. For example, if welding or fire occurs in a special place, the fine dust concentration instantly jumps over 1000 times higher than the previously measured fine dust concentration. Since it is necessary to determine whether this value is noise or an actual measured value, continuity must be ensured through continuous measurement (usually continuous measurement for more than 10 seconds) to give reliability to the measured value and to include it as a statistical value. In addition, in order to increase reliability and further reinforce the value, it is possible to determine the appropriateness of the numerical value by detecting another change using a proximity sensor or an illuminance sensor as well as a temperature/humidity sensor. As mentioned before, if there is no report of a specific task, it is necessary to notify the manager of the danger status by e-mail or text message.

3. Machine Learning of Fine Dust Reduction Device

The fine dust reduction device performs learning on all measured values regardless of manual management or automatic management. However, there is a method of learning inside the fine dust reduction device and a method of learning using a separate server.

Learning through the server can save the cost of additional hardware required for each fine dust reduction device and share the learning contents with all other fine dust reduction devices. In addition, the server obtains all information that cannot be measured with an environmental sensor (for example, from a weather center or public institution, outdoor fine dust concentration or weather forecast in the area, information on wind speed and volume, etc.) It provides the advantage of being able to make predictions about Of course, it can be executed in the central processing unit, but security and additional hardware are burdened.

The server collects environmental information and makes the necessary learning results into actions for each mode and downloads them to the central processing unit in the fine dust reduction device to re-execute the process of controlling individual functional devices and auxiliary devices, or to periodically upgrade and execute the program. do. In addition, the server can create and distribute a new action in real time even if a given environment is changed.

If the fine dust reduction device is used in an environment where internet is not available, only limited variables are learned inside the fine dust reduction device, so many of the benefits mentioned above disappear.

Actions are provided by an algorithm, called the decision tree in a variety of experiments and experience in the development process, the best action is reinforced by the machine learning called like reinforcement learning (Reinforcement Learning for Recommendation) utilizing a continuous sample after the installation do.

In reinforcement learning, the fine dust reduction device becomes an agent, calculates one action using the measurement values provided by the sensor and the device, operates the functional device and the auxiliary device, and then measures the measured value again Receive evaluation or reward according to the results of

If you want to perform reinforcement learning, you need a large number of samples obtained through manual management of administrators. Automatic management uses samples that have already been secured, so the value of the samples is weak. However, there is a difficulty in relying on reinforcement learning for all actions due to the problem that it is difficult to obtain a variety of initial samples. Therefore, by using the scenarios obtained based on the development, experimentation, and experience process for each mode mentioned above, a decision tree and a random forest are created and provided as an initial menu, and the user's selection and actions are modified by the administrator. Samples are continuously made and provided by measuring the results according to the

100: central processing unit 101: wire 102: embedded computer
103: sensing unit 104: control unit 105: power supply device
106: antenna 107: fine dust sensor 108: temperature/humidity sensor
109: display panel 110: distance proximity sensor 200: raven structure
201: coupling member 300: drip tray 301: circulation fan
302: exhaust fan 303: cylindrical guide 305: left and right rotating plate
306: vertical rotation port 307: steel wire 310: filter box
311: adsorption filter 312: upper pre-filter 313: lower pre-filter
400: server 401: air outlet 402: hot air supply device
500: safety fence 600: hot and hot water supply device 601: spray device
602: solenoid valve 603: pump 604: water hose

Claims (10)

an environment sensor that detects changes in the environment and expresses them as environmental values;
Functional device for reducing fine dust;
a ramen structure in which one or more functional devices are mounted using a coupling member;
a discharge device attached to the top of the raven structure to discharge the purified air pushed up by the functional device to a predetermined position to reduce the fine dust; and
(a) an original copy of the environmental value and the measured value of the functional device and the measured value of the exhaust device;
(b) including the statistics of the normal distribution corresponding to the raw material accumulated for a predetermined elapsed time
(c) the first process of transmitting the log recorded by the process to the server;
A second process of downloading an action generated based on environment information from the server and applying it to the process in real time;
Including a third process of controlling the environment sensor, the functional device, and the discharge device
and a central processing unit that repeats the first process, the second process, and the third process to achieve the environmental goal.
The functional device is
at least one atomizing device provided at the lower part of the filter box having a built-in adsorption filter to generate water particles; and
Artificial intelligence fine dust reduction device comprising a; a circulation fan provided at the upper portion of the filter box to flow the moist air containing the fine dust to the adsorption filter to be adsorbed. The method of claim 1,
The environmental sensor
It includes a fine dust sensor that measures the concentration of fine dust.
Temperature/humidity sensor to measure temperature and humidity, proximity sensor to measure movement of people and objects, illuminance sensor to measure illuminance, water level sensor to measure water level, water temperature sensor to measure water temperature, It further includes at least one or more of a power meter for measuring the amount of electricity, and a noise sensor for measuring noise,
The environmental information is
Time information, information on parts constituting the fine dust reduction device, laws and regulations to be observed at the installation site, local fine dust forecast, regional weather forecast, the environmental value, the measured value of the functional device, the measured value of the exhaust device, the above An artificial intelligence fine dust reduction device comprising a log, and any one or more of the detailed step-by-step execution time and repetition number of the action. The method of claim 1,
The discharge device is
exhaust fan with adjustable air volume and speed;
an open hemispherical vertical rotation port attached to the bottom of the exhaust fan to enable vertical rotation of the exhaust fan based on the horizontal and to collect air coming up from the bottom and discharge it to the front; and
Artificial intelligence fine dust reduction device comprising a; is attached to the bottom of the vertical rotation port, is installed on the top of the ramen structure, and can rotate left and right with respect to the front side. 3. The method of claim 2,
The central processing unit is
An artificial intelligence fine dust reduction device that applies power to the functional device and the discharge device when a result of adding the average value and the standard deviation derived from the normal distribution of the fine dust is greater than a set value. According to claim 1,
A hot water supply device for generating and supplying any one of natural water, hot water, or steam to the spray device; and
The central processing unit additionally transmits, to the server, the measured value of the hot water supply device and a statistic of a normal distribution corresponding to the measured value accumulated for a predetermined elapsed time, and downloads the action from the server to perform the purification. An artificial intelligence fine dust reduction device that controls the hot water supply device so that the dew point temperature of the air is equal to or higher than the outdoor temperature. 6. The method of claim 5,
The dew point temperature is calculated by the following [Equation 1] and [Equation 2]
[Equation 1]
243.12 x [Equation 2]/ (17.62 - [Equation 2])
[Equation 2]
ln(relative humidity/100) + (17.62 x temperature}/(243.12 + temperature)
Artificial intelligence fine dust reduction device. According to claim 1,
and a hot air supply device for supplying hot air into the filter box.
The central processing unit additionally transmits, to the server, the measured value of the hot air supply device and a statistic of a normal distribution corresponding to the measured value accumulated for a predetermined elapsed time, and downloads the action from the server to obtain the purified air An artificial intelligence fine dust reduction device that additionally controls the hot air supply device so that the sensible temperature of the user is equal to or higher than the outdoor temperature. According to claim 1,
The server is an artificial intelligence fine dust reduction device that notifies the manager to execute the action exceeding the allowable range of the mode in response to the crisis alert and emergency alert issued by the meteorological center and receives approval. 3. The method of claim 2,
The central processing unit is an artificial intelligence fine dust reduction device that periodically puts the artificial intelligence fine dust reduction device into a sleep state by the action generated based on any one of the movement of the person and the object, the time information, or the illuminance. . According to claim 1,
The central processing unit is an artificial intelligence fine dust reduction device for identifying noise generated from the measurement values of the environment detection sensor and the functional device and the measurement values of the emission device.

Images