KR20210093666A - Air conditioner and method thereof - Google Patents

Abstract

The present invention relates to an air conditioner, and an operating method thereof. According to an embodiment of the present invention, the operating method of the air conditioner comprises: operation of performing a trial run according to a preset condition; operation of acquiring data for each configuration provided in the air conditioner while performing the trial run; operation of determining input data which is a processing target among the data acquired during the trial run; operation of generating a learning model calculating a refrigerant amount by learning the determined input data; and operation of calculating the refrigerant amount by inputting the input data into the learning module acquired while performing the trial run after the learning model is generated. In addition, various embodiments are possible.

Description

Air conditioner and its control method

The present invention relates to an air conditioner and a control method therefor, and more particularly, to an air conditioner capable of calculating an amount of a refrigerant and a control method therefor.

The air conditioner is installed to provide a more comfortable indoor environment to humans by discharging cold and hot air into the room to adjust the indoor temperature and purify the indoor air in order to create a comfortable indoor environment. In general, an air conditioner includes an indoor unit configured as a heat exchanger and installed indoors, and an outdoor unit configured as a compressor and a heat exchanger and supplying refrigerant to the indoor unit.

The air conditioner is operated for cooling or heating according to the flow of refrigerant. During cooling operation, high-temperature and high-pressure liquid refrigerant is supplied to the indoor unit from the outdoor unit's compressor through the outdoor unit's heat exchanger, and as the refrigerant expands and vaporizes in the indoor unit's heat exchanger, the ambient air temperature decreases, and the indoor unit fan rotates. Accordingly, the cold air is discharged into the room. During heating operation, high-temperature and high-pressure gaseous refrigerant is supplied from the compressor of the outdoor unit to the indoor unit, and air heated by the energy released as the high-temperature and high-pressure gaseous refrigerant is liquefied in the heat exchanger of the indoor unit is released into the indoor unit according to the operation of the indoor unit fan. is discharged with

Meanwhile, while circulating in the air conditioner, the amount of refrigerant needs to be maintained at an appropriate level. When the amount of refrigerant is insufficient during operation of the air conditioner, cooling and heating efficiency is reduced, and components such as a compressor or motor are damaged, and when the amount of refrigerant is excessively large, not only the cooling and heating efficiency is reduced, but also the compressor Problems such as an increase in the noise of the air conditioner or an increase in the operating current of the air conditioner may occur.

Conventionally, in order to calculate the amount of refrigerant in an air conditioner, the amount of refrigerant is indirectly estimated depending on whether data such as discharge temperature and room temperature correspond to a certain condition, or data is applied to a regression equation determined by an experiment. By substituting, it is determined whether the amount of refrigerant is insufficient, appropriate, or excessive.

However, according to the conventional method, simply estimating the amount of refrigerant based on whether data meets certain conditions has significantly lower accuracy in calculating the amount of refrigerant, and the regression equation is determined according to the conditions and data of the heating and cooling cycle at the time of the experiment If an appropriate cycle corresponding to the regression equation is not formed at the time of calculating the amount of refrigerant, an error in calculating the amount of refrigerant may increase.

In addition, when there are few types of data substituted for certain conditions or regression equations, there is a problem in that the accuracy of calculating the amount of refrigerant is lowered. If there are too many types of data, time and resources are excessively consumed in calculating the amount of refrigerant. You only need to use the optimal data you need.

In particular, in the case of a multi-type air conditioner for heating and cooling a plurality of indoor spaces, including a plurality of indoor units, the length of the pipe connected to each of the plurality of indoor units varies depending on the installation environment, and the amount of refrigerant flowing through each pipe is different. Since the set values of the plurality of indoor units are set variously during actual operation, the amount of refrigerant is generally calculated using data obtained according to the fixed settings during test operation. However, only temporary data obtained during test operation may be insufficient as basic data for calculating the amount of refrigerant, and when the data is uncertain or insufficient, it is difficult to accurately calculate the amount of refrigerant.

SUMMARY OF THE INVENTION It is an object of the present invention to provide an air conditioner capable of accurately calculating the amount of refrigerant using optimal data required for calculating the amount of refrigerant, and a method for controlling the same.

In order to achieve the above object, an air conditioner according to various embodiments of the present invention includes a control unit that calculates an amount of refrigerant based on data for each configuration provided in the air conditioner, and the control unit is configured to: A learning data acquisition unit that acquires data for each configuration provided in the air conditioner during trial operation, determines input data to be processed among the acquired data, and a learning model that learns the input data to generate a learning model that calculates the amount of refrigerant It may include a learning unit and a refrigerant amount calculating unit for calculating the amount of refrigerant by inputting the input data obtained during the trial operation after the learning unit is created to the generated learning model.

In order to achieve the above object, a method of operating an air conditioner according to various embodiments of the present invention includes an operation of performing a trial run according to a preset condition, and data on each configuration provided in the air conditioner during the trial run. During the operation of acquiring the operation, the operation of determining the input data to be processed among the data acquired during the trial operation, the operation of generating a learning model that calculates the refrigerant amount by learning the determined input data, and the operation of the trial operation after the learning model is created It may include the operation of calculating the amount of refrigerant by inputting the obtained input data to the learning model.

According to various embodiments of the present invention, through machine learning and deep learning, artificial intelligence that selects optimal data required for calculating the amount of refrigerant among data related to each configuration provided in the air conditioner and learns the selected optimal data By configuring and using the algorithm, it is possible to more accurately calculate the amount of refrigerant, thereby improving the reliability of the cooling/heating efficiency and performance of the air conditioner.

Further scope of applicability of the present invention will become apparent from the following detailed description. However, it should be understood that the detailed description and specific embodiments such as preferred embodiments of the present invention are given by way of illustration only, since various changes and modifications within the spirit and scope of the present invention may be clearly understood by those skilled in the art.

1 is a diagram illustrating the configuration of an air conditioner system according to an embodiment of the present invention.
FIG. 2 is a schematic diagram of an outdoor unit and an indoor unit of FIG. 1 .
3 is a block diagram of an air conditioner according to an embodiment of the present invention.
4 is a diagram referenced in the description of deep learning, according to an embodiment of the present invention.
5 to 6B are flowcharts illustrating a method of operating an air conditioner according to an embodiment of the present invention.
7A to 13 are diagrams referenced in the description of the operating method of the air conditioner.

Advantages and features of the present invention and methods of achieving them will become apparent with reference to the embodiments described below in detail in conjunction with the accompanying drawings. However, the present invention is not limited to the embodiments disclosed below, but may be implemented in various different forms, and only these embodiments allow the disclosure of the present invention to be complete, and common knowledge in the art to which the present invention pertains It is provided to fully inform those who have the scope of the invention, and the present invention is only defined by the scope of the claims. Like reference numerals refer to like elements throughout.

Spatially relative terms "below", "beneath", "lower", "above", "upper", etc. It can be used to easily describe the correlation between a component of , and another component. Spatially relative terms should be understood as terms including different orientations of components in use or operation in addition to the orientation shown in the drawings. For example, when a component shown in the drawing is turned over, a component described as “beneath” or “beneath” of another component may be placed “above” of the other component. can Accordingly, the exemplary term “below” may include both directions below and above. Components may also be oriented in other orientations, and thus spatially relative terms may be interpreted according to orientation.

The terminology used herein is for the purpose of describing the embodiments and is not intended to limit the present invention. As used herein, the singular also includes the plural unless specifically stated otherwise in the phrase. As used herein, "comprises" and/or "comprising" means that a referenced component, step and/or action excludes the presence or addition of one or more other components, steps and/or actions. I never do that.

Unless otherwise defined, all terms (including technical and scientific terms) used herein may be used with the meaning commonly understood by those of ordinary skill in the art to which the present invention belongs. In addition, terms defined in a commonly used dictionary are not to be interpreted ideally or excessively unless clearly specifically defined.

In the drawings, the thickness or size of each component is exaggerated, omitted, or schematically illustrated for convenience and clarity of description. In addition, the size and area of each component do not fully reflect the actual size or area.

The suffixes "module" and "~ part" for the components used in the following description are given or mixed in consideration of only the ease of writing the specification, and do not have a meaning or role distinct from each other by themselves.

At this time, it will be understood that each block of the flowchart diagrams and combinations of the flowchart diagrams may be performed by computer program instructions. These computer program instructions may be embodied in a processor of a general purpose computer, special purpose computer, or other programmable data processing equipment, such that the instructions performed by the processor of the computer or other programmable data processing equipment are not described in the flowchart block(s). It creates a means to perform functions. These computer program instructions may also be stored in a computer-usable or computer-readable memory that may direct a computer or other programmable data processing equipment to implement a function in a particular manner, and thus the computer-usable or computer-readable memory. It is also possible that the instructions stored in the flow chart block(s) produce an article of manufacture containing instruction means for performing the function described in the flowchart block(s). The computer program instructions may also be mounted on a computer or other programmable data processing equipment, such that a series of operational steps are performed on the computer or other programmable data processing equipment to create a computer-executed process to create a computer or other programmable data processing equipment. It is also possible that instructions for performing the processing equipment provide steps for performing the functions described in the flowchart block(s).

Additionally, each block may represent a module, segment, or portion of code that includes one or more executable instructions for executing specified logical function(s). It should also be noted that in some alternative implementations it is also possible for the functions recited in blocks to occur out of order. For example, two blocks shown one after another may in fact be performed substantially simultaneously, or it is possible that the blocks are sometimes performed in the reverse order according to the corresponding function.

Also, in this specification, terms such as first and second may be used to describe various elements, but these elements are not limited by these terms. These terms are only used to distinguish one element from another.

1 is a diagram illustrating the configuration of an air conditioner according to an embodiment of the present invention.

Referring to FIG. 1 , the air conditioner 100 according to an embodiment of the present invention may include an outdoor unit 21 and at least one indoor unit 31 connected to the outdoor unit 21 . A plurality of indoor units 31 may be connected to outdoor units 21 , and the number is not limited to the drawings.

The indoor unit 31 may include at least one of a stand-type indoor unit 31a, a wall-mounted indoor unit 31b, and a ceiling-type indoor unit 31c.

Meanwhile, the air conditioner 100 may further include at least one of a ventilator, an air purifier, a humidifier, and a heater, and may operate in conjunction with the operations of the indoor unit 31 and the outdoor unit 21 .

The outdoor unit 21 includes a compressor (not shown) that receives and compresses a refrigerant, an outdoor heat exchanger (not shown) that exchanges heat between the refrigerant and outdoor air, and an accumulator (not shown) that extracts a gaseous refrigerant from the supplied refrigerant and supplies it to the compressor ( (not shown) and a four-way valve (not shown) for selecting a flow path of a refrigerant according to a heating operation. In addition, the outdoor unit 21 may further include a plurality of sensors, valves, and an oil recovery unit.

The outdoor unit 21 may supply the refrigerant to the indoor unit 31 by operating the provided compressor and the outdoor heat exchanger to compress or heat exchange the refrigerant according to a setting. The outdoor unit 21 may be driven by a demand of the central controller 10 or the indoor unit 31 . In this case, as the cooling/heating capacity is varied in response to the driven indoor unit 31 , the number of outdoor units and the number of compressors installed in the outdoor unit may vary.

In this case, the outdoor unit 21 may supply the compressed refrigerant to the connected indoor unit 31 .

The indoor unit 31 may receive refrigerant from the outdoor unit 21 and discharge cold and hot air into the room. The indoor unit 31 may include an indoor heat exchanger (not shown), an indoor unit fan (not shown), an expansion valve (not shown) through which the supplied refrigerant is expanded, and a plurality of sensors (not shown).

At this time, the outdoor unit 21 and the indoor unit 31 may be connected to each other through a communication line to transmit and receive data, and the outdoor unit 21 and the indoor unit 31 are connected to the central controller 10 by wire or wirelessly to the central controller ( 10) may be operated under the control.

The remote controller 41 may be connected to the indoor unit 31 , transmit a user's control command to the indoor unit 31 , and receive and display status information of the indoor unit 31 . In this case, the remote control 41 may communicate with the indoor unit 31 by wire or wirelessly depending on the connection type.

Meanwhile, the air conditioner 100 may further include at least one sensor (not shown) capable of detecting a state of indoor air. For example, the air conditioner 100 may further include a temperature sensor for detecting indoor temperature, a humidity sensor for detecting indoor humidity, an air pressure sensor for detecting indoor air pressure, a sensor for measuring the amount of dust in the indoor air, and the like. In addition, it may include a sensor that can collect various data such as temperature, humidity, atmospheric pressure, and the amount of dust in the air.

Meanwhile, the air conditioner 100 may communicate with the external device 50 and may transmit/receive data to and from each other. For example, the air conditioner 100 may communicate with an external device (eg, a personal computer) by wire, and data on the status of each component provided in the air conditioner 100 and whether an error occurs or not can transmit and receive. The air conditioner 100 may transmit and receive data by accessing a server (not shown) connected to an external network.

FIG. 2 is a schematic diagram of an outdoor unit and an indoor unit of FIG. 1 .

Referring to FIG. 2 , the air conditioner 100 according to an embodiment of the present invention may be largely divided into an outdoor unit 21 and an indoor unit 31 . The air conditioner 100 may include a plurality of indoor units 31a to 31c.

The outdoor unit 21 includes a compressor 110 serving to compress the refrigerant, a compressor motor (not shown) for driving the compressor 110, an outdoor heat exchanger 120 serving to radiate heat from the compressed refrigerant; An accumulator 130 that temporarily stores the vaporized refrigerant to remove moisture and foreign substances and supplies the refrigerant at a constant pressure to the compressor, a cooling/heating switching valve 140 that changes the flow path of the compressed refrigerant, an oil separator 150, An outdoor blower 160 comprising an outdoor fan 161 disposed on one side of the outdoor heat exchanger 120 to promote heat dissipation of the refrigerant and a motor 162 for an outdoor fan rotating the outdoor fan 161, and expands the condensed refrigerant and at least one expansion mechanism (eg, electronic expansion valves (EEV)).

More specifically, the outdoor unit 21 may include a gas pipe service valve 113 to which the gas pipe 182 is connected and a liquid pipe service valve 114 to which the liquid pipe 112 is connected. The gas pipe service valve 113 and the liquid pipe service valve 114 may be connected to the indoor unit 31 through a refrigerant pipe, and may circulate the refrigerant of the outdoor unit 21 .

At least one of an inverter compressor and a constant speed compressor may be used as the compressor 110 .

The outdoor heat exchanger 120 may exchange heat between outdoor air and a refrigerant, and may include a plurality of units 122 and 124 according to an embodiment. The outdoor heat exchanger 120 may operate as a condenser during a cooling operation and as an evaporator during a heating operation.

The outdoor expansion valve 170 may expand the refrigerant flowing into the outdoor heat exchanger 120 during a heating operation, and may pass through the refrigerant without expanding it during a cooling operation. As the outdoor expansion valve 170 , an electronic expansion valve (EEV) capable of adjusting an opening value according to an input signal may be used.

The outdoor expansion valve 170 includes a first outdoor expansion valve 172 that expands the refrigerant flowing into the first outdoor heat exchanger 122 and a second outdoor expansion valve 172 that expands the refrigerant flowing into the second outdoor heat exchanger 174 . It may include an outdoor expansion valve (174).

The outdoor unit 21 may further include a hot gas unit 190 for bypassing the refrigerant supplied to the outdoor heat exchanger 120 to the indoor unit 31 during a heating operation. The hot gas unit 90 may include a hot gas bypass pipe (`, 192) for bypassing the refrigerant, and hot gas valves (193, 194). In this case, the first hot gas valve 193 and the second hot gas valve 194 may be selectively operated. For example, only the first hot gas valve 193 may be opened/closed, or only the second hot gas valve 194 may be opened/closed. Meanwhile, in this embodiment, a lamination valve 195 for laminating the first hot gas bypass pipe 191 and the second hot gas bypass pipe 192 may be disposed.

The outdoor unit 21 may further include a supercooling unit 200 disposed in the liquid pipe 112 . The supercooling unit 200 is bypassed in the supercooling heat exchanger 201 and the liquid pipe 112, and the supercooling bypass pipe 202 connected to the supercooling heat exchanger 201 is disposed in the supercooling bypass pipe 202. The first supercooling expansion valve 203 for selectively expanding the refrigerant flowing and the supercooling-compressor connecting pipe 204 connecting the supercooling heat exchanger 201 and the compressor 110, and/or the supercooling-compressor connecting pipe ( It may include a second supercooling expansion valve 205 disposed in 204 and selectively expanding the flowing refrigerant.

The outdoor unit 21 may further include a receiver 210 disposed in the liquid pipe 112 . The receiver 210 may store liquid refrigerant in order to adjust the amount of circulated refrigerant. The receiver 210 may store the liquid refrigerant separately from the liquid refrigerant being stored in the accumulator 30 . For example, the receiver 210 may supply the refrigerant to the accumulator 130 when the amount of the circulating refrigerant is insufficient, and may recover and store the refrigerant when the amount of the circulated refrigerant is large.

The receiver 210 may include a receiver tank 211 for storing the refrigerant, and receiver valves 213 and 215 for controlling the flow of the refrigerant.

The indoor units 31a to 31c include indoor heat exchangers 33a to 33c which are disposed indoors and perform a cooling/heating function, expansion valves 35a to 35c through which the supplied refrigerant is expanded, and indoor heat exchangers 33a to 33a to 33c), an indoor blower (not shown) comprising an indoor fan (not shown) that promotes heat dissipation of the refrigerant and a motor (not shown) for rotating the indoor fan, a plurality of sensors (not shown), etc. may include At least one indoor heat exchanger (33a to 33c) may be installed in each of the indoor units (31a to 31c).

The air conditioner 100 may be configured as an air conditioner that cools the room, or may be configured as a heat pump that cools or heats the room.

3 is a block diagram of an air conditioner according to an embodiment of the present invention.

Referring to FIG. 3 , the air conditioner 100 includes a sensor unit 310 , a communication unit 320 , a fan driving unit 330 , a compressor driving unit 340 , a storage unit 350 , an input unit 360 , and an output unit. 370 and/or a control unit 380 . The air conditioner 100 according to various embodiments of the present invention may further include various components not shown in FIG. 3 .

The communication unit 310 may include at least one communication module. The communication unit 310 may be provided in each of the outdoor unit 21 and the indoor unit 31 , and the outdoor unit 21 and the indoor unit 31 may transmit/receive data to and from each other. For example, the communication method between the outdoor unit 21 and the indoor unit 31 includes a communication method using a power line, a serial communication method (eg, RS-485 communication), a wired communication method through a refrigerant pipe, as well as a Wi-Fi (Wi- fi), Bluetooth (Bluetooth), beacon (Beacon), may be a wireless communication method such as Zigbee (zigbee).

Meanwhile, the communication unit 310 may transmit/receive data to and from an external device. For example, the communication unit 310 may establish a wireless communication channel with an external device (eg, a mobile terminal), and the state of each configuration provided in the air conditioner 100 through the established wireless communication channel; Data regarding whether or not an error has occurred can be transmitted and received. The communication unit 310 may connect to a server connected to an external network to transmit/receive data.

The sensor unit 320 may include a plurality of sensors, and may transmit data on detection values detected through the plurality of sensors to the controller 370 . For example, the sensor unit 320 is disposed inside the outdoor heat exchanger 120 , a heat exchanger temperature sensor (not shown) that detects a condensation temperature or an evaporation temperature, and each pipe of the air conditioner 100 . A pressure sensor (not shown) for detecting the pressure of the gas refrigerant flowing through the pipe temperature sensor (not shown) for detecting the temperature of the gas refrigerant flowing through each pipe of the air conditioner 100, the temperature of the room It may include an indoor temperature sensor (not shown) for detecting, an outdoor temperature sensor (not shown) for detecting the outdoor temperature, and the like.

The storage unit 330 may store a program for each signal processing and control in the control unit 370 , or may store a signal-processed voice or data signal. For example, the storage unit 330 stores application programs designed for the purpose of performing various tasks that can be processed by the control unit 370, and, upon request of the control unit 370, selectively selects some of the stored application programs. can provide The program stored in the storage unit 330 is not particularly limited as long as it can be executed by the control unit 370 .

Although the embodiment in which the storage unit 330 of FIG. 3 is provided separately from the control unit 370 is illustrated, the scope of the present invention is not limited thereto, and the storage unit 330 may be included in the control unit 370 .

The storage unit 330 may store data related to each component provided in the air conditioner 100 . For example, the storage unit 330 may store data on detection values detected from a plurality of sensors provided in the sensor unit 320 . For example, the storage unit 330 may store data on power consumption, operating frequency, and the like of the compressor 110 . For example, the storage unit 330 may store data on the number of rotations of the fan 351 , the degree of opening of each electronic expansion valve (EEV), the degree of superheating, the degree of subcooling, and the like.

The compressor driving unit 340 may drive the compressor 110 . For example, the compressor driving unit 340 includes a rectifying unit (not shown) for rectifying AC power into DC power and outputting it, a dc stage capacitor for storing the pulsating voltage from the rectifying unit, and a plurality of switching elements, the smoothed DC An inverter (not shown) for converting and outputting power into three-phase AC power of a predetermined frequency and/or a motor for a compressor (not shown) for driving the compressor 110 according to the three-phase AC power output from the inverter can

The compressor driving unit 340 may change the operating frequency of the compressor 110 under the control of the control unit 370 . For example, the compressor driving unit 340 may change the operating frequency of the compressor 110 by changing the frequency of the three-phase AC power output to the compressor motor under the control of the control unit 370 .

The fan driving unit 350 may drive the fan 351 provided in the air conditioner 100 . For example, the fan driving unit 350 may drive the outdoor fan 161 and/or an indoor fan (not shown). For example, the fan driving unit 350 includes a rectifier (not shown) for rectifying AC power into DC power and outputting it, a dc stage capacitor for storing the pulsating voltage from the rectifier, and a plurality of switching elements, and a smoothed DC It may include an inverter (not shown) that converts and outputs power into three-phase AC power of a predetermined frequency and/or a motor that drives a fan according to the three-phase AC power output from the inverter.

Meanwhile, the fan driving unit 350 may be provided with a configuration for driving the outdoor fan 161 and the indoor fan separately.

The fan driver 350 may change the rotation speed of the fan 351 according to the control of the controller 370 . For example, the fan driving unit 350 may change the rotation speed of the outdoor fan 161 by changing the frequency of the three-phase AC power output to the outdoor fan motor under the control of the controller 370 . For example, the fan driving unit 350 may change the rotational speed of the indoor fan by changing the frequency of the three-phase AC power output to the indoor fan motor under the control of the controller 370 .

The output unit 360 may include a display device such as a display (not shown) and a light emitting diode (LED), and related to the operation state of the air conditioner 100 and the occurrence of an error through the display device. Operation status can be displayed.

The output unit 360 may include an audio device such as a speaker and a buzzer, and may output a sound effect for the operating state of the air conditioner 100 through the audio device, and output a predetermined warning sound when an error occurs. can

The controller 370 may be connected to each component provided in the air conditioner 100 and may control the overall operation of each component. The controller 370 may transmit/receive data to and from each component provided in the air conditioner 100 .

The controller 370 may include at least one processor. Here, the processor may be a general processor such as a central processing unit (CPU). Of course, the processor may be a dedicated device such as an ASIC or other hardware-based processor.

The control unit 370 may be provided in at least one of the indoor unit 31 and/or the central controller 10 as well as the outdoor unit 21 .

The controller 370 may acquire data related to each configuration provided in the air conditioner 100 . In this case, the controller 370 may acquire data related to each configuration provided in the air conditioner 100 at a predetermined time interval according to a predetermined period in consideration of the computational load. Here, the data related to each configuration provided in the air conditioner 100 includes the operating frequency of the compressor 110 , the suction temperature of the compressor 110 , the discharge temperature of the compressor 110 , the suction pressure of the compressor 110 , The discharge pressure of the compressor 110, the degree of superheat, the degree of subcooling, the number of rotations of the outdoor fan 161, the condensation temperature or evaporation temperature of the outdoor heat exchanger 120, the outlet temperature of the outdoor heat exchanger 120, the indoor unit (31) the inlet pipe temperature, the outlet pipe temperature of the indoor unit 31, the indoor temperature, the outdoor temperature, the opening degree of each of the plurality of electronic expansion valves (EEV), the liquid pipe 112 temperature, the discharge superheat degree, supercooling It may include an inlet side temperature of the heat exchanger 201 , an outlet side temperature of the supercooling heat exchanger 201 , and the like.

The control unit 370 may perform a trial run according to a preset condition, and may acquire data related to each configuration provided in the air conditioner 100 during the trial run. Here, the test operation is performed under preset conditions to determine whether the state of each component provided in the air conditioner 100 is good, whether the connection state between the components is good, whether the operation according to the operation mode is good, etc. It may mean to operate according to

The controller 370 may check the operation mode of the air conditioner 100 , and may control the operation of each component provided in the air conditioner 100 according to the operation mode of the air conditioner 100 . . For example, the controller 370 may determine whether the operation mode of the air conditioner 100 is a general operation mode for indoor cooling or heating, when the air conditioner 100 is installed or when an error occurs, the air conditioner ( 100), it is possible to check whether it is in the test run mode to check the status.

The controller 370 may calculate the amount of refrigerant in the air conditioner 100 based on data related to each configuration provided in the air conditioner 100 . When the air conditioner 100 is a multi-type air conditioner including a plurality of indoor units 31 , the amount of refrigerant in the air conditioner 100 may be calculated based on data obtained during the test operation.

The control unit 370 learns data related to each configuration provided in the air conditioner 100 through machine learning, such as deep learning, and generates a learning model for calculating the amount of refrigerant. , it is possible to calculate the amount of refrigerant in the air conditioner 100 using the generated learning model. In this case, when generating a learning model for calculating the amount of refrigerant of the air conditioner 100 , the controller 370 may generate a learning model for cooling and a learning model for heating, respectively. Hereinafter, with reference to FIG. 4, deep learning will be described in detail.

4 is a diagram referenced in the description of deep learning, according to an embodiment of the present invention.

Machine learning means that a computer learns from data without a human instructing the computer directly to logic, and allows the computer to solve a problem through this.

Deep learning is a method of teaching a human way of thinking to a computer based on Artificial Neural Networks (ANN), and it refers to an artificial intelligence technology that allows the computer to learn on its own like a human. The artificial neural network (ANN) may be implemented in the form of software or in the form of hardware such as a chip. For example, an artificial neural network (ANN), a deep neural network (DNN), a convolutional neural network (CNN), a recurrent neural network (RNN), a deep belief network; DBN), etc., may include various types of algorithms.

Referring to FIG. 4 , an artificial neural network (ANN) may include an input layer, a hidden layer, and an output layer. Each layer includes a plurality of nodes, each layer may be connected to a next layer, and nodes between adjacent layers may be connected to each other with a weight.

A computer can discover a certain pattern from data to form a feature map, extract low-level features, mid-level features, and high-level features to recognize objects and output the results.

Also, each node may operate based on an activation model, and an output value corresponding to an input value may be determined according to the activation model.

An output value of an arbitrary node, for example, a low-level feature, may be input to a next layer connected to the node, for example, a node of an intermediate-level feature. A node of a next layer, for example, a node of an intermediate-level feature may receive values output from a plurality of nodes of a lower-level feature.

In this case, the input value of each node may be a value in which a weight is applied to the output value of the node of the previous layer. A weight may mean a connection strength between nodes. In addition, the deep learning process can be viewed as a process of finding appropriate weights and biases.

Meanwhile, an output value of an arbitrary node, for example, an intermediate-level feature, may be input to a next layer connected to the corresponding node, for example, a node of a higher-level feature. A node of a next layer, for example, a node of a higher-level feature, may receive values output from a plurality of nodes of an intermediate-level feature.

An artificial neural network (ANN) may extract feature information corresponding to each level by using a learned layer corresponding to each level. An artificial neural network (ANN) may recognize a predetermined target by sequentially abstracting and utilizing feature information of the highest level.

On the other hand, learning of the artificial neural network (ANN) can be made by adjusting the weight of the connection line between nodes so that a desired output is obtained with respect to input data, and a bias value can also be adjusted if necessary. In addition, the artificial neural network (ANN) may continuously update the weight value by learning. In addition, a method such as back-propagation may be used for learning of an artificial neural network (ANN).

The storage unit 330 may store data obtained from each configuration provided in the air conditioner 100 , data for learning an artificial neural network (ANN), and the like. For example, the storage unit 330, for learning the artificial neural network (ANN), a database for each configuration provided in the air conditioner 100, weights and biases constituting the artificial neural network (ANN) structure (bias) can be stored.

The control unit 370 may determine input data that is a processing target used for generating a learning model for calculating the amount of refrigerant from among data related to each configuration provided in the air conditioner 100, and learn the determined input data, It is possible to create a learning model for calculating the amount of refrigerant.

The controller 370 may generate pseudo data by amplifying data related to each configuration provided in the air conditioner 100 . For example, the controller 370 may generate pseudo data by amplifying data related to each configuration provided in the air conditioner 100 in a bootstrap resampling method. Here, the bootstrap resampling method is an algorithm that generates pseudo data by resampling and amplifying data based on statistical data of an existing feature vector (eg, mean, standard deviation, etc. for a feature vector). can mean

The controller 370 may calculate a feature weight for each data related to each configuration provided in the air conditioner 100 . The control unit 370 uses a feature extraction algorithm such as Principal Component Analysis (PCA), Linear Discriminant Analysis (LDA), Neighborhood Component Analysis (NCA), and the like, and a feature selection algorithm such as ReliefF. , a weight may be calculated for each data related to each configuration provided in the air conditioner 100 . For example, when the air conditioner 100 knows the influence of each data on the calculation of the refrigerant amount, by using the NCA algorithm, a weight is applied to each data related to each configuration provided in the air conditioner 100 . can be calculated. For example, when the air conditioner 100 does not know the effect of the data on the refrigerant amount calculation, the weight is applied to each data related to each configuration provided in the air conditioner 100 using the ReliefF algorithm. can be calculated.

The controller 370 may determine the number of data determined as input data based on weights calculated for each data related to each configuration provided in the air conditioner 100 . For example, the control unit 370 learns data related to each configuration provided in the air conditioner 100 while changing the number of data used for learning according to the calculated weight, and learns the plurality of first preliminary learnings. You can create a model. At this time, the controller 370 may check the accuracy of the result of calculating the refrigerant amount by the plurality of first preliminary learning models, respectively, and learn from the first preliminary learning model in which the accuracy of the result of calculating the refrigerant amount is equal to or greater than a preset standard The number of data determined as input data may be determined according to the number of data used for training a model having the smallest number of data used for .

The controller 370 may scale a value of data related to each configuration provided in the air conditioner 100 . The controller 370 may scale a value of data determined as input data among data related to each configuration provided in the air conditioner 100 . Here, the scaling may mean adjusting the scale so that each value of the input data is included in the same range.

For example, the controller 370 may scale the data value based on Equation 1 above. In this case, x may mean a data value, xmax may mean a maximum value of data, and xmin may mean a minimum value of data.

The controller 370 may determine the number of hidden nodes included in the learning model for calculating the amount of refrigerant by using a grid search method.

The controller 370 may generate a plurality of second preliminary learning models based on the input data. In this case, the controller 370 may generate a plurality of second preliminary learning models while changing the number of hidden nodes included in the hidden layer. For example, the controller 370 may include one hidden node in any one of the plurality of second preliminary learning models in the hidden layer, and the other one of the plurality of second preliminary learning models may include two hidden nodes in the hidden layer. It can contain hidden nodes.

Meanwhile, the controller 370 may classify data for learning the learning model into a training set and a validation set. Here, the training set may mean data for learning the learning model, and the verification set may mean data for confirming the learning degree of the learning model.

The controller 370 may compare the accuracy of the refrigerant amount calculation results of the plurality of second preliminary learning models, and may determine the number of hidden nodes of the learning model for calculating the refrigerant amount based on the comparison results. For example, the control unit 370 may calculate the accuracy of the refrigerant amount calculation result of the plurality of second preliminary learning models by using the verification set for each of the plurality of second preliminary learning models, respectively, and the accuracy is the highest. The number of nodes included in the second preliminary learning model may be determined as the number of hidden nodes in the learning model for calculating the refrigerant amount. For example, the control unit 370 may calculate the accuracy of the refrigerant amount calculation result of the plurality of second preliminary learning models by using the verification set for each of the plurality of second preliminary learning models, respectively, and the accuracy is preset. The number of hidden nodes of the second preliminary learning model including the smallest number of hidden nodes among the second preliminary learning models greater than or equal to the reference may be determined as the number of hidden nodes of the learning model for calculating the amount of refrigerant. In this case, the preset criterion for the accuracy may be set by the user or may be determined according to the highest accuracy among the accuracies for the plurality of second preliminary learning models.

The controller 370 may generate a learning model for calculating the amount of refrigerant by controlling the learning model including the determined number of hidden nodes to learn the input data. At this time, the control unit 370 controls the learning model for calculating the amount of refrigerant to learn the correlation between the input data and the amount of refrigerant by using a back propagation algorithm, and the weight and bias constituting the learning model can decide

The controller 370 may output data on the amount of refrigerant calculated through the learning model through the output unit 360 . For example, the controller 370 may output a refrigerant filling rate according to the calculated refrigerant amount through the display.

The controller 370 may output a warning message through the output unit 360 when the amount of refrigerant calculated through the learning model is less than a preset minimum standard or is greater than or equal to a preset maximum standard.

Meanwhile, the control unit 370 may include a learning data obtaining unit (not shown), a model learning unit (not shown), and/or a refrigerant amount calculating unit (not shown).

The learning data acquisition unit may acquire data for each component included in the air conditioner 100 and determine input data to be processed among the acquired data. The model learning unit may generate a learning model for calculating the amount of refrigerant by learning the input data. The refrigerant amount calculating unit may calculate the refrigerant amount by inputting input data obtained after the learning model is generated into the learning model.

5 to 6B are flowcharts illustrating an operating method of the air conditioner according to an embodiment of the present invention, and FIGS. 7A to 13 are diagrams referenced in the description of the operating method of the air conditioner.

Referring to FIG. 5 , in operation S510 , the air conditioner 100 may perform a trial run according to a preset condition. For example, the air conditioner 100 may control each component included in the air conditioner 100 to operate according to a preset condition.

In operation S520 , the air conditioner 100 may acquire data related to each configuration provided in the air conditioner 100 while performing a test operation. For example, the air conditioner 100 includes, as shown in Table 1 below, the operating frequency of the compressor 110 , the suction temperature of the accumulator 130 , the discharge temperature of the compressor 110 , the suction pressure of the compressor 110 , The discharge pressure of the compressor 110, the degree of superheat, the degree of subcooling, the number of rotations of the outdoor fan 161, the condensation temperature or evaporation temperature of the outdoor heat exchanger 120, the outlet temperature of the outdoor heat exchanger 120, the indoor unit The inlet pipe temperature of 31, the outlet pipe temperature of the indoor unit 31, the indoor temperature, the outdoor temperature, the opening degree of the electronic expansion valve (EEV), and the like can be obtained.

Index Data Index Data One discharge pressure of the compressor (110) 14 Temperature of the first outdoor heat exchanger (122) 2 Suction pressure of the compressor (110) 15 Temperature of the second outdoor heat exchanger (124) 3 compression ratio
(discharge pressure/suction pressure) 16 Temperature of the outlet side of the outdoor heat exchanger (120) 4 degree of subcooling 17 Opening degree of the first outdoor expansion valve (172) 5 superheat 18 Opening degree of the second outdoor expansion valve (174) 6 Rotational speed of the outdoor fan 161 19 Opening degree of the first supercooling expansion valve (203) 7 Suction temperature of the accumulator 130 20 Opening degree of the second supercooling expansion valve (205) 8 Discharge temperature of compressor 110 21 Liquid pipe (112) temperature 9 outdoor temperature 22 evaporation temperature 10 room temperature 23 Inlet temperature of the supercooling heat exchanger (201) 11 condensing temperature 24 The temperature of the outlet side of the supercooling heat exchanger (201) 12 Discharge superheat
(Compressor 110 discharge temperature - condenser temperature) 25 Supercooling degree Superheating degree 13 Operating frequency of compressor 110

In operation S530 , the air conditioner 100 may determine input data, which is a processing target, used to generate a learning model for calculating the amount of refrigerant from among data related to each configuration provided in the air conditioner 100 . This will be described with reference to FIG. 6A.

Referring to FIG. 6A , the air conditioner 100 may generate pseudo data by amplifying data related to each component included in the air conditioner 100 in operation S601 . For example, the air conditioner 100 may generate pseudo data by amplifying data related to each configuration provided in the air conditioner 100 in a bootstrap resampling method.

In operation S602 , the air conditioner 100 may calculate a weight for each data related to each configuration based on data related to each configuration provided in the air conditioner 100 including the pseudo data. For example, the air conditioner 100 may calculate a weight for each data related to each configuration provided in the air conditioner 100 using the ReliefF algorithm.

In operation S603 , the air conditioner 100 may scale a value of data related to each component included in the air conditioner 100 including the pseudo data. For example, the air conditioner 100 may scale each value of the data to be included in the same range based on the maximum value of the data, the minimum value of the data, and the like.

The air conditioner 100 learns data related to each configuration provided in the air conditioner 100 including pseudo data while changing the number of data used for learning according to the calculated weight in operation S604. , a plurality of first preliminary learning models may be generated.

In operation S605 , the air conditioner 100 may determine a specific model from among the plurality of first preliminary learning models according to a preset criterion. For example, the air conditioner 100 may check the accuracy of the result of calculating the amount of refrigerant by the plurality of first preliminary learning models, respectively, and the accuracy of the result of calculating the amount of refrigerant among the plurality of first preliminary learning models It may be determined whether or not there is at least one model of which is equal to or greater than a preset standard.

At this time, the air conditioner 100 determines the number of data used for learning the model having the smallest number of data used for learning among the first preliminary learning models in which the accuracy of the result of calculating the amount of refrigerant is equal to or greater than the preset standard The number of data determined as input data may be determined.

In operation S606 , the air conditioner 100 receives input data from among data related to each configuration provided in the air conditioner 100 according to the number of data determined as input data in consideration of a weight for each data. can decide

7A is a graph of the weights of data obtained when the air conditioner 100 is tested according to the cooling mode, and FIG. 7B is the accuracy of the result of calculating the amount of refrigerant by the plurality of first preliminary learning models according to the training set , and FIG. 7c is a graph of the accuracy of the result of calculating the refrigerant amount by the plurality of first preliminary learning models according to the test set.

Referring to FIGS. 7A to 7C , when the test operation is performed according to the cooling mode, weights of data related to each configuration provided in the air conditioner 100 may be calculated in the order shown in Table 2 below.

Rank Data Rank Data One outdoor temperature 14 superheat 2 Temperature of the outlet side of the outdoor heat exchanger (120) 15 Discharge temperature of compressor 110 3 Liquid pipe (112) temperature 16 Discharge superheat
(Compressor 110 discharge temperature - condenser temperature) 4 The temperature of the outlet side of the supercooling heat exchanger (201) 17 Opening degree of the first outdoor expansion valve (172) 5 Rotational speed of the outdoor fan 161 18 Opening degree of the second outdoor expansion valve (174) 6 Suction temperature of the accumulator 130 19 compression ratio
(discharge pressure/suction pressure) 7 condensing temperature 20 Temperature of the first outdoor heat exchanger (122) 8 Opening degree of the second supercooling expansion valve (205) 21 Inlet temperature of the supercooling heat exchanger (201) 9 Suction pressure of the compressor (110) 22 room temperature 10 discharge pressure of the compressor (110) 23 Operating frequency of compressor 110 11 Supercooling degree Superheating degree 24 Temperature of the second outdoor heat exchanger (124) 12 degree of subcooling 25 Opening degree of the first supercooling expansion valve (203) 13 evaporation temperature

The air conditioner 100 may generate a plurality of first preliminary learning models while changing the number of data used for learning according to the calculated weight. For example, the air conditioner 100 may change the number of data while excluding data having the lowest weight one by one.

At this time, when the plurality of first preliminary learning models calculate the amount of refrigerant according to the training set, it can be confirmed that the accuracy of 70% or more appears from the first preliminary learning model using 10 data for learning, and the plurality of first preliminary learning models It can be confirmed that the same is true even when the model calculates the amount of refrigerant according to the test set.

Therefore, the number of data determined as input data for calculating the amount of refrigerant during trial operation according to the cooling mode may be determined to be 10, and in the order of weights, from the outdoor temperature to the discharge pressure of the compressor 110 during trial operation according to the cooling mode It can be determined from the input data of

8A is a graph of the weights of data obtained when the air conditioner 100 is test-run according to the heating mode, and FIG. 8B is the accuracy of the result of calculating the amount of refrigerant by the plurality of first preliminary learning models according to the training set , and FIG. 8C is a graph of the accuracy of the result of calculating the amount of refrigerant by the plurality of first preliminary learning models according to the test set.

Referring to FIGS. 8A to 8C , when a test operation is performed according to a heating mode, weights of data related to each configuration provided in the air conditioner 100 may be calculated in the order shown in Table 3 below.

Rank Data Rank Data One outdoor temperature 14 evaporation temperature 2 Liquid pipe (112) temperature 15 Suction pressure of the compressor (110) 3 discharge pressure of the compressor (110) 16 Rotational speed of the outdoor fan 161 4 superheat 17 Opening degree of the second supercooling expansion valve (205) 5 The temperature of the outlet side of the supercooling heat exchanger (201) 18 compression ratio
(discharge pressure/suction pressure) 6 Temperature of the outlet side of the outdoor heat exchanger (120) 19 room temperature 7 condensing temperature 20 Opening degree of the second outdoor expansion valve (174) 8 Discharge superheat
(Compressor 110 discharge temperature - condenser temperature) 21 Inlet temperature of the supercooling heat exchanger (201) 9 Supercooling degree Superheating degree 22 Temperature of the second outdoor heat exchanger (124) 10 degree of subcooling 23 Operating frequency of compressor 110 11 Suction temperature of the accumulator 130 24 Temperature of the first outdoor heat exchanger (122) 12 Opening degree of the first outdoor expansion valve (172) 25 Opening degree of the first supercooling expansion valve (203) 13 Discharge temperature of compressor 110

The air conditioner 100 may generate a plurality of first preliminary learning models while changing the number of data used for learning according to the calculated weight. For example, the air conditioner 100 may change the number of data while excluding data having the lowest weight one by one.

At this time, when the plurality of first preliminary learning models calculate the amount of refrigerant according to the training set, it can be confirmed that the accuracy of 80% or more appears from the first preliminary learning model using five data for learning, and the plurality of first preliminary learning models It can be confirmed that the same is true when the model calculates the refrigerant amount according to the test set.

Accordingly, the number of data determined as input data for calculating the amount of refrigerant during trial operation according to the heating mode may be determined to be five, and in the order of weight, from the outdoor temperature to the outlet temperature of the supercooling heat exchanger 201 in the heating mode. It may be determined based on input data at the time of the test run.

Referring back to FIG. 5 , the air conditioner 100 may generate a learning model for calculating the amount of refrigerant in operation S540 . This will be described with reference to FIG. 6B.

Referring to FIG. 6B , the air conditioner 100 may scale the determined value of the input data in operation S611. For example, the air conditioner 100 may scale each value of the input data to be included in the same range based on the maximum value of the input data, the minimum value of the input data, and the like.

The air conditioner 100 may generate a plurality of second preliminary learning models based on the input data in operation S612 . In this case, the air conditioner 100 may generate a plurality of second preliminary learning models while changing the number of hidden nodes included in the hidden layer.

The air conditioner 100 may compare the accuracy of the refrigerant amount calculation result of the plurality of second preliminary learning models in operation S613, and based on the comparison result, determine the number of hidden nodes of the learning model for calculating the refrigerant amount there is. For example, the air conditioner 100 may calculate the accuracy of the refrigerant amount calculation result of the plurality of second preliminary learning models by using the verification set for each of the plurality of second preliminary learning models, respectively, and Based on the result of comparing the accuracy of the refrigerant amount calculation result of the second preliminary learning model, it is possible to determine the number of hidden nodes of the learning model for calculating the refrigerant amount.

The air conditioner 100 may generate a learning model for calculating the amount of refrigerant by controlling the learning model including the determined number of hidden nodes to learn the input data in operation S614 . In this case, in generating the learning model for calculating the amount of refrigerant, the air conditioner 100 may generate a learning model during trial operation according to the cooling mode and a learning model during trial operation according to the heating mode, respectively.

9A is a graph of the number of parameters used for learning the learning model according to the number of hidden layers constituting the learning model, and FIG. 9B is a plurality of second preliminary learning models according to the training set. It is a graph of the accuracy of the result of calculating the amount of refrigerant, and FIG. 9C is a graph of the accuracy of the result of calculating the amount of refrigerant by the plurality of second preliminary learning models according to the test set.

Referring to FIG. 9A , based on the graph 901 when the number of hidden layers is one, even if the number of hidden nodes included in the hidden layer increases, the number of parameters used for learning the learning model is significantly different. However, based on the graph 902 when the number of hidden layers is two, as the number of hidden nodes included in the hidden layer increases, the number of parameters used for learning the learning model increases exponentially. that can be checked Accordingly, the air conditioner 100 may determine the number of hidden layers constituting the learning model as one.

Referring to FIG. 9B , when the plurality of second preliminary learning models calculate the amount of refrigerant according to the training set, it can be seen that the accuracy of more than 90% appears from the second preliminary learning model including five hidden nodes.

On the other hand, referring to FIG. 9C , when the plurality of second preliminary learning models calculate the amount of refrigerant according to the test set, the accuracy of the second preliminary learning model including five hidden nodes does not reach 70%, and 10 It can be seen that the accuracy of 75% or more appears from the second preliminary learning model including the hidden node.

At this time, the air conditioner 100 selects the smallest number of hidden nodes from among the second preliminary learning models showing the accuracy above a predetermined standard based on the accuracy when the amount of refrigerant is calculated using the learning set and the test set. 10, which is the number of hidden nodes of the included model, may be determined as the number of hidden nodes of the learning model for calculating the amount of refrigerant.

Referring back to FIG. 5 , in operation S550 , the air conditioner 100 may determine whether a trial run is performed again according to a preset condition. For example, the air conditioner 100 may determine whether a trial operation is performed again while a normal operation is performed after generating a learning model for calculating the amount of refrigerant.

In operation S560 , when the test run is performed again according to a preset condition, the air conditioner 100 may acquire data related to each configuration provided in the air conditioner 100 during the test run.

The air conditioner 100 may calculate the amount of refrigerant based on the generated learning model in operation S570. For example, the air conditioner 100 may calculate the amount of refrigerant based on data determined as input data among data related to each configuration provided in the air conditioner 100 .

Referring to FIG. 10A , when the training set is input to the learning model generated during test operation according to the cooling mode ( 1012 ), it can be confirmed that the amount of refrigerant is calculated with an accuracy of 70.3% when compared with the actual amount of refrigerant ( 1011 ). In addition, referring to FIG. 10B , even when the test set is input to the learning model generated during trial operation according to the cooling mode (1022), it can be confirmed that the amount of refrigerant is calculated with an accuracy of 70.4% when compared with the actual amount of refrigerant 1021 there is.

On the other hand, referring to FIG. 11A , when the training set is input to the learning model generated during trial operation according to the heating mode (1112), it can be confirmed that the amount of refrigerant is calculated with an accuracy of 91.2% when compared with the actual amount of refrigerant (1011). . In addition, referring to FIG. 11b , even when the test set is input to the learning model generated during trial operation according to the heating mode ( 1122 ), it can be confirmed that the amount of refrigerant is calculated with an accuracy of 91.0% when compared with the actual amount of refrigerant ( 1121 ). there is.

Referring to FIG. 12 , the air conditioner 100 displays data on the amount of refrigerant calculated through the learning model on the screen 1200 output through the display, data related to each configuration provided in the air conditioner 100 . can be displayed with In this case, the air conditioner 100 may display the current refrigerant supply state 1210 according to the refrigerant amount calculated through the learning model.

Meanwhile, referring to FIG. 13 , the air conditioner 100 may transmit data on the amount of refrigerant calculated through the learning model to the mobile terminal 1300 through the communication unit 310 . Alternatively, the air conditioner 100 may transmit, through the communication unit 310, data on the amount of refrigerant calculated through the learning model to a server (not shown), and the mobile terminal 1300 may transmit the data on the amount of refrigerant calculated through the learning model. Data on the amount of refrigerant may be received from the server.

In this case, the mobile terminal 1300 may display a screen for an application providing various functions related to the air conditioner 100 , and the user may display a screen corresponding to the function of checking the operating state of the air conditioner 100 . Item 1310 may be selected.

Among the items displayed on the screen of the mobile terminal 1300 , when the item 1310 corresponding to the function of checking the operating state of the air conditioner 100 is selected, the mobile terminal 1300 performs the learning of the air conditioner 100 . A screen indicating the operating state of the air conditioner 100 including data on the amount of refrigerant calculated through the model may be displayed. In this case, the mobile terminal 1300 may display the current refrigerant supply state 1320 according to the refrigerant amount calculated through the learning model of the air conditioner 100 .

As described above, according to various embodiments of the present invention, through machine learning and deep learning, from among data related to each configuration provided in the air conditioner 100, optimal data required for calculating the amount of refrigerant is selected, and the selected By constructing and using an artificial intelligence algorithm that learns optimal data, the amount of refrigerant can be more accurately calculated, thereby improving the reliability of the air conditioning efficiency and performance of the air conditioner 100 .

In addition, in the case of a multi-type air conditioner that includes a plurality of indoor units and calculates the amount of refrigerant using data obtained according to a fixed setting during test operation, temporary data obtained during trial operation alone may not be sufficient as basic data for calculating the amount of refrigerant. However, according to various embodiments of the present invention, since the training model is trained by generating pseudo data by amplifying data obtained during trial operation, the amount of refrigerant can be accurately calculated even during trial operation.

The accompanying drawings are only for easy understanding of the embodiments disclosed in the present specification, and the technical spirit disclosed herein is not limited by the accompanying drawings, and all changes and equivalents included in the spirit and scope of the present invention It should be understood to include water or substitutes.

Likewise, although acts are depicted in the drawings in a particular order, it should not be construed that such acts must be performed in that particular order or sequential order shown, or that all depicted acts must be performed in order to achieve desirable results. . In certain cases, multitasking and parallel processing may be advantageous.

In addition, although preferred embodiments of the present invention have been illustrated and described above, the present invention is not limited to the specific embodiments described above, and the technical field to which the present invention belongs without departing from the gist of the present invention as claimed in the claims In addition, various modifications may be made by those of ordinary skill in the art, and these modifications should not be individually understood from the technical spirit or perspective of the present invention.

Claims (8)

A method of operating an air conditioner including a plurality of indoor units, the method comprising:
performing a trial run according to a preset condition;
acquiring data for each configuration provided in the air conditioner while performing the test operation;
determining input data to be processed among the data for each configuration obtained during the trial run;
generating a learning model for calculating the amount of refrigerant by learning the determined input data; and
and calculating the amount of refrigerant by inputting the input data obtained during the trial run after the learning model is generated into the learning model. According to claim 1,
The operation of determining the input data is
generating pseudo data by amplifying the data for each configuration obtained during the trial run in a bootstrap resampling method; and
and determining the input data based on the generated pseudo data. 3. The method of claim 2,
Determining the input data based on the generated pseudo data includes:
calculating a weight for each of the data for each configuration based on the generated pseudo data;
generating a plurality of first preliminary learning models based on the generated pseudo data while changing the number of data used for learning according to the calculated weight; and
Among the plurality of generated first preliminary learning models, when there is at least one model in which the accuracy of the calculation result of the refrigerant amount is equal to or greater than the preset standard, the model in which the accuracy of the result of calculating the refrigerant amount is equal to or greater than the preset standard determining a first preliminary learning model having the smallest number of data used for the learning among them; and
The method of operating the air conditioner according to the determined number of data used for learning of the first preliminary learning model, further comprising the operation of determining the input data. 4. The method of claim 3,
The operation of calculating the weight comprises calculating the weight according to a ReliefF algorithm. 5. The method of claim 4,
The operation of generating the learning model is,
scaling a value of the input data based on a minimum value and a maximum value of each of the data determined as the input data;
generating a plurality of second preliminary learning models while changing the number of hidden nodes included in a hidden layer based on the scaled input data; and
and determining the number of hidden nodes of a model having the highest accuracy with respect to the result of calculating the amount of refrigerant among the generated plurality of second preliminary learning models as the number of hidden nodes of the learning model How to operate an air conditioner. 6. The method of claim 5,
The operation of generating the learning model is,
Air conditioning, characterized in that it further comprises the operation of determining a weight and a bias by learning a correlation between the scaled input data and the amount of refrigerant according to an error back-propagation method how the machine works. 7. The method of claim 6,
outputting a message about the calculated amount of refrigerant through an output unit included in the air conditioner; and
and outputting a warning message through the output unit when the calculated amount of refrigerant is less than a preset minimum standard or when the calculated amount of refrigerant is greater than or equal to a preset maximum standard. An air conditioner including a plurality of indoor units, the air conditioner comprising:
a control unit for calculating the amount of refrigerant based on data for each configuration provided in the air conditioner;
The control unit is
a learning data acquisition unit configured to acquire data for each configuration provided in the air conditioner while performing a test operation according to a preset condition, and to determine input data to be processed among the acquired data;
a model learning unit for generating a learning model for calculating the amount of refrigerant by learning the input data; and
and a refrigerant amount calculation unit configured to calculate the refrigerant amount by inputting the input data obtained during the trial run after the learning model is generated into the created learning model.

Images