KR20210055998A - Air conditioner controlling method and air conditioner controlling apparatus - Google Patents

Abstract

Disclosed are an air-conditioner control method and an air-conditioner control device, in which an air conditioner and a server can communicate in a 5G communication environment by executing a built-in artificial intelligence (AI) algorithm and/or a machine learning algorithm. The air-conditioner control method according to an embodiment of the present invention comprises the steps of: obtaining compressor operation data from a compressor included in an air conditioner; classifying state information of the compressor corresponding to the compressor operation data using an unsupervised learning-based deep learning model pre-trained to classify the state information of the compressor by inputting the compressor operation data as an input; repeating the classification step at regular intervals; acquiring the state information classification result of the compressor based on a batch basis; determining a final state of the compressor including one of a normal operation of the compressor, the failure of the compressor, and the imminent failure of the compressor on the basis of statistical information on the state information classification result of the compressor, which is acquired based on the batch basis; and actively controlling the operation of the compressor by receiving the final state of the compressor as feedback.

Description

Air conditioner control method and air conditioner control device {AIR CONDITIONER CONTROLLING METHOD AND AIR CONDITIONER CONTROLLING APPARATUS}

The present invention relates to an air conditioner control method and an air conditioner control device, and more particularly, to diagnose a compressor failure by performing unsupervised learning based on a deep learning model on operation data of a compressor included in the air conditioner. And an air conditioner control method and an air conditioner control device for predicting failure of a compressor.

In general, an air conditioner includes a compressor, a condenser, an expansion valve, and an evaporator connected by a refrigerant circulation line. The refrigerant sequentially moves to a compressor, a condenser, an expansion valve, and an evaporator along the refrigerant circulation line, and is fed back to the compressor to circulate inside the air conditioner. In the process of refrigerant circulation, the gaseous refrigerant is compressed to high temperature and high pressure in the compressor, the high temperature and high pressure gaseous refrigerant is cooled to a low temperature and high pressure liquid state in the condenser. Is changed to a liquid state (capillary phenomenon), and in the evaporator, a low-temperature, low-pressure liquid refrigerant is heated to a gaseous state. When the air conditioner is disposed, the compressor and the condenser may be disposed outdoors, the evaporator may be disposed indoors, and the expansion valve may be disposed on the refrigerant circulation line. Accordingly, the compressor and condenser may be referred to as "outdoor units", and the evaporator may be referred to as "indoor units". The air conditioner continuously performs cooling by heat exchange of refrigerants in the "indoor" and "outdoor units". In "indoor", the refrigerant is vaporized and heat is taken from the indoor air by state change energy, and in "outdoor unit", the refrigerant is liquefied and heat is released to the outside by state change energy.

Prior Art 1 (Korean Patent Registration No. 10-1927213 (2018.12.04)) measures vibration by attaching a vibration sensor to the air compressor in order to diagnose the possibility of failure of an air compressor employing a single screw method. It analyzes, installs a gap sensor at each end of the rotation shaft of the main rotor and gate rotor, measures the axial displacement, analyzes the change trend, and determines the abnormality.

Prior Art 2 (Korean Patent Registration No. 10-0996630 (2010.11.19)) has a motor protector that stops the compressor motor when it detects that the refrigeration/air conditioning system is out of the characteristic state, and monitors the state of the motor protector to prevent the compressor. Disclosed is the content of diagnosing the type of problem.

It is difficult to obtain compressor failure data from an air conditioner, and the reason is as follows. Since most air conditioners are not connected to an external network, there is a method of acquiring compressor operation data directly through the product during service, but the amount of data that can be acquired is limited and the phenomenon before compressor failure is resolved. There is a problem that it is not sufficiently reflected in the data. In addition, even for an air conditioner connected to an external network, it is practically impossible for an administrator to individually determine and label whether the data received from the server is faulty data, impending fault data, or normal data. In addition, acquiring the compressor failure data through the experiment involves many risk factors in the compressor failure reproducing experiment, and the experiment cost increases exponentially because the compressor replacement for each experimental condition must be repeatedly performed.

The above-described background technology is technical information possessed by the inventor for derivation of the present invention or acquired during the derivation process of the present invention, and is not necessarily known to be known to the general public prior to filing the present invention.

An object of the present invention is to diagnose compressor failure and predict compressor failure by performing unsupervised learning based on a deep learning model on operation data of a compressor.

An object of the present invention is to predict the failure of the compressor at a time before the occurrence of the failure of the compressor without using a series of sensors for determining the failure of the compressor.

An object of the present invention is to automatically classify failure data of a compressor by performing unsupervised learning based on a deep learning model on operation data of a compressor.

An object of the present invention is to additionally control the operation of a compressor by feeding back a compressor diagnosis result.

An air conditioner control method according to an embodiment of the present invention includes performing unsupervised learning based on a deep learning model on operation data of a compressor in real time to diagnose a failure of a compressor and predict a failure of the compressor. Can include.

Specifically, the air conditioner control method according to the present embodiment includes the steps of acquiring compressor operation data from a compressor included in the air conditioner, and unsupervised learning trained in advance to classify state information of the compressor by inputting the compressor operation data. The step of classifying the state information of the compressor corresponding to the compressor operation data using the based deep learning model, repeating the classifying step at regular intervals, and obtaining the classification result of the state information of the compressor in batch units. Determining a final state of the compressor including one of a normal operation of the compressor, a failure of the compressor, and an imminent compressor failure, based on the step and statistical information on the classification result of the condition information of the compressor obtained in batch units; and It may include the step of actively controlling the operation of the compressor by receiving the final state of the compressor as a feedback.

Through the air conditioner control method according to the present embodiment, by performing unsupervised learning based on a deep learning model on the operation data of the compressor, diagnosing the failure of the compressor and predicting the failure of the compressor, the service engineer Since there is no need for pre-inspection, service cost and inspection time can be reduced, customer reliability can be improved, and unstable downtime can be eliminated by not operating the air conditioner.

The air conditioner control apparatus according to the present embodiment includes an acquisition unit that obtains compressor operation data from a compressor included in the air conditioner, and an unsupervised learning-based pretrained to classify state information of the compressor by inputting the compressor operation data. By using a deep learning model to classify the state information of the compressor corresponding to the compressor operation data, the classification operation of the classification unit is repeated every predetermined period, and the classification result of the state information of the compressor is obtained in batch units, A determination unit that determines the final state of the compressor including one of normal operation of the compressor, failure of the compressor, and impending failure of the compressor, based on statistical information on the classification result of the condition information of the compressor obtained in batch units, and It may include a compressor control unit for actively controlling the operation of the compressor by receiving the final state as feedback.

An air conditioner control apparatus according to another embodiment includes one or more processors and a memory connected to the one or more processors, and the memory acquires compressor operation data from a compressor included in the air conditioner, and inputs compressor operation data Classify the state information of the compressor corresponding to the compressor operation data using a deep learning model based on unsupervised learning that has been trained in advance to classify the state information of the compressor by using Status information classification results are obtained in batch units, and based on statistical information on the classification results of compressor status information obtained in batch units, one of the following is included: normal operation of the compressor, failure of the compressor, and impending compressor failure. A command that causes the final state of the compressor to be determined and the final state of the compressor to be input as a feedback to be input to actively control the operation of the compressor may be stored.

In addition, another method for implementing the present invention, another system, and a computer-readable recording medium storing a computer program for executing the method may be further provided.

Other aspects, features, and advantages other than those described above will become apparent from the following drawings, claims, and detailed description of the invention.

According to the present invention, a compressor failure is diagnosed by performing unsupervised learning based on a deep learning model on the operation data of a compressor, and a failure of the compressor is predicted, thereby eliminating the need for a preliminary inspection by a service engineer It can reduce the overhaul and inspection time.

In addition, by performing unsupervised learning based on a deep learning model on the operation data of the compressor, the failure of the compressor can be diagnosed, and the failure of the compressor can be predicted to improve the customer's reliability. Unfortunate downtime can be eliminated.

It can also disperse service complaints before the season of increasing air conditioner use.

In addition, there is no need for additional sensors (for example, vibration sensor, noise measurement sensor, image sensor, etc.) for the compressor failure plate, and it is possible to predict the failure of the compressor in advance at the time before the failure of the compressor, thereby improving customer reliability. , Can reduce service cost and maintenance time.

In addition, it is possible to improve the power efficiency of the air conditioner by diagnosing a failure of the compressor using only the optimal processor resource and predicting the failure of the compressor.

The effects of the present invention are not limited to those mentioned above, and other effects not mentioned will be clearly understood by those skilled in the art from the following description.

1 is an exemplary diagram of an air conditioner control environment including an air conditioner, a user terminal, a server, and a network connecting them to each other according to an embodiment of the present invention.
2 is a schematic block diagram of an apparatus for controlling an air conditioner according to an embodiment of the present invention.
3 is a schematic block diagram of a diagnosis unit according to an embodiment of the air conditioner control apparatus of FIG. 2.
4 and 5 are exemplary diagrams illustrating output results of a classification unit of the diagnosis unit of FIG. 3.
6 is an exemplary view of output data of a classification unit for determining a final state of a compressor by a determination unit of the diagnosis unit of FIG. 3.
7 is a flowchart illustrating a method of controlling an air conditioner according to an embodiment of the present invention.

Advantages and features of the present invention, and a method of achieving them will be apparent with reference to embodiments described in detail together with the accompanying drawings. However, it should be understood that the present invention is not limited to the embodiments presented below, but may be implemented in various different forms, and includes all transformations, equivalents, and substitutes included in the spirit and scope of the present invention. . The embodiments presented below are provided to complete the disclosure of the present invention, and to completely inform the scope of the invention to those of ordinary skill in the art to which the present invention pertains. In describing the present invention, when it is determined that a detailed description of a related known technology may obscure the subject matter of the present invention, a detailed description thereof will be omitted.

The terms used in the present application are only used to describe specific embodiments, and are not intended to limit the present invention. Singular expressions include plural expressions unless the context clearly indicates otherwise. In the present application, terms such as "comprise" or "have" are intended to designate the presence of features, numbers, steps, actions, components, parts, or combinations thereof described in the specification, but one or more other features. It is to be understood that the presence or addition of elements or numbers, steps, actions, components, parts, or combinations thereof does not preclude in advance. Terms such as first and second may be used to describe various elements, but the elements should not be limited by the terms. The above terms are used only for the purpose of distinguishing one component from another component.

Hereinafter, embodiments according to the present invention will be described in detail with reference to the accompanying drawings, and in the description with reference to the accompanying drawings, identical or corresponding components are assigned the same reference numbers, and redundant descriptions thereof will be omitted. I will do it.

1 is an exemplary diagram of an air conditioner control environment including an air conditioner, a user terminal, a server, and a network connecting them to each other according to an embodiment of the present invention. Referring to FIG. 1, an air conditioner control environment may include an air conditioner 100, a user terminal 200, a server 300, and a network 400. The air conditioner 100, the user terminal 200, and the server 300 may be connected to each other in a 5G communication environment.

The air conditioner 100 may have a cooling mode and a heating mode according to a refrigerant circulation direction. Hereinafter, for convenience of description, a case in which the air conditioner 100 of the present invention operates in a cooling mode will be described as an example. The air conditioner 100 according to the present embodiment may include an indoor unit 110, an outdoor unit 120, and a refrigerant circulation line (130 of FIG. 2 ).

The indoor unit 110 is disposed indoors, the outdoor unit 120 is disposed outdoors, and the indoor unit 110 and the outdoor unit 120 may be connected by a refrigerant circulation line 130. The refrigerant may circulate the indoor unit 110 and the outdoor unit 120 through the refrigerant circulation line 130.

Referring to FIG. 2, the indoor unit 110 may include an external case, an evaporator 111 that is an internal component, and an indoor unit fan 112. The outdoor unit 120 may include an external case, a compressor 121, a condenser 122 and an outdoor unit fan 123, which are internal components.

The refrigerant circulation line 130 may be a pipe that sequentially connects the evaporator 111, the compressor 121, and the condenser 122, and is fed back to the evaporator 111. In addition, an expansion valve 131 may be disposed between the condenser 122 and the evaporator 111 on the refrigerant circulation line 130.

The refrigerant may be fed back to the evaporator 111 by passing through the evaporator 111, the compressor 121, the condenser 122, and the expansion valve 131 in sequence along the refrigerant circulation line 130.

In the evaporator 111, a low-temperature, low-pressure liquid refrigerant may be evaporated through heat exchange with indoor air. That is, the refrigerant in the evaporator 111 may cool the indoor air by absorbing heat from the indoor air.

In the compressor 121, the refrigerant in the gaseous state heated in the evaporator 111 is compressed by mechanical driving, so that the refrigerant in the gaseous state of high temperature and high pressure may be changed. For example, the compressor 121 may include a rotary motor and an electric cylinder (not shown). In this case, the compressor 121 can compress the refrigerant by pressurizing the electric cylinder by rotation of the rotary motor.

In the condenser 122, the high-temperature, high-pressure gaseous refrigerant compressed by the compressor 121 may be condensed through heat exchange with outdoor air. That is, the refrigerant in the condenser 122 may heat the outdoor air by discharging heat to the outdoor air.

In the expansion valve 131, the low-temperature, high-pressure liquid refrigerant condensed in the condenser 122 passes through the capillary tube, and may be changed into a low-temperature, low-pressure liquid refrigerant.

The indoor unit fan 112 flows indoor air to increase a heat exchange rate between the refrigerant of the evaporator 111 and the indoor air, and blows the cooled indoor air throughout the room. For example, the indoor unit fan 112 may be an impeller that rotates by a rotation motor.

The outdoor unit fan 123 may flow outdoor air to increase a heat exchange rate between the refrigerant of the condenser 122 and the outdoor air, and discharge the heated outdoor air to the outside. For example, the outdoor unit fan 123 may be an impeller that rotates by a rotation motor.

In this embodiment, the air conditioner 100 performs a process of cooling the room by obtaining heat from the indoor unit 110 and discharging the heat obtained from the indoor unit 110 through the outdoor unit 120 by circulation of a refrigerant. It can be repeated continuously to cool the room. The controller 170 may control the indoor unit 110 and the outdoor unit 120 based on the operation mode.

Returning to FIG. 1, the air conditioner 100 may acquire compressor operation data in order to diagnose a failure of the compressor 121 and predict a failure of the compressor 121. The compressor operation data may be stored in a memory (160 in FIG. 2) or may be obtained from the compressor 121 in real time.

The air conditioner 100 may classify state information of a compressor corresponding to the compressor operation data by using an unsupervised learning-based deep learning model trained in advance to classify state information of the compressor by inputting the compressor operation data as input.

The air conditioner 100 may repeat classification of state information of the compressor corresponding to the compressor operation data at a predetermined period by using a deep learning model based on unsupervised learning.

The air conditioner 100 may obtain a result of classifying status information of the compressor 121 in a batch unit.

The air conditioner 100 performs normal operation of the compressor 121, failure of the compressor 121, and failure of the compressor 121 based on statistical information on the classification result of the status information of the compressor 121 obtained in batch units. It is possible to determine the final state of the compressor 121 including one of the imminent ones.

The air conditioner 100 may actively control the operation of the compressor 121 by receiving the final state of the compressor 121 as a feedback. For example, the air conditioner 100 may control the compressor 121 under a phase current limiting condition in response to a final state of the compressor 121 including the impending failure of the compressor 121.

After accessing the air conditioner control application or the air conditioner control site, the user terminal 200 monitors the status information of the air conditioner 100, drives or controls the air conditioner 100 through an authentication process. You can receive services that can be provided. In this embodiment, the user terminal 200 that has completed the authentication process operates the air conditioner 100 by determining an operation mode of the air conditioner 100 or controls the operation of the air conditioner 100. can do.

The user terminal 200 may include a communication terminal capable of performing a function of a computing device (not shown). In this embodiment, the user terminal 200 is a desktop computer, a smartphone, a laptop computer, and Tablet PC, smart TV, mobile phone, personal digital assistant (PDA), laptop, media player, micro server, global positioning system (GPS) device, e-book terminal, digital broadcasting terminal, navigation, kiosk, MP3 player, digital camera, home appliance Devices and other mobile or non-mobile computing devices may be, but are not limited thereto. In addition, the user terminal 200 may be a wearable terminal such as a watch, glasses, hair band, and ring having a communication function and a data processing function. The user terminal 200 is not limited to the above-described contents, and a terminal capable of web browsing may be borrowed without limitation.

The server 300 may be a database server that provides big data required to apply various artificial intelligence algorithms and data related to control of the air conditioner 100. In addition, the server 300 may include a web server or an application server that enables remote control of the air conditioner 100 using an application installed in the user terminal 200 or a web browser.

Here, artificial intelligence (AI) is a field of computer science and information technology that studies how computers can do the thinking, learning, and self-development that human intelligence can do. It could mean being able to imitate intelligent behavior.

In addition, artificial intelligence does not exist by itself, but is directly or indirectly related to other fields of computer science. In particular, in modern times, attempts to introduce artificial intelligence elements in various fields of information technology and use them to solve problems in the field are being made very actively.

Machine learning is a branch of artificial intelligence and can include a field of research that gives computers the ability to learn without explicit programming. Specifically, machine learning can be said to be a technology that studies and builds a system that learns based on empirical data, performs prediction, and improves its own performance, and algorithms for it. Rather than executing strictly defined static program instructions, machine learning algorithms can take a way to build specific models to derive predictions or decisions based on input data.

The server 300 may receive compressor operation data from the air conditioner 100. The server 300 may classify the state information of the compressor 121 corresponding to the compressor operation data using an unsupervised learning-based deep learning model trained in advance to classify the state information of the compressor by inputting the compressor operation data as input. . The server 300 performs normal operation of the compressor 121, failure of the compressor 121, and failure of the compressor 121 based on statistical information on the classification result of the status information of the compressor 121 obtained in a batch unit. It is possible to determine the final state of the compressor, including one of the imminent failures. The server 300 may receive the final state of the compressor as a feedback, generate a compressor control signal, and transmit it to the air conditioner 100.

Depending on the processing capability of the air conditioner 100, at least some of the classification result of the condition information of the compressor 121, determination of the final condition of the compressor 121, and control of the compressor 121 are performed by the air conditioner 100. Can be done by

The network 400 may serve to connect the air conditioner 100, the user terminal 200, and the server 300. Such networks 400 include wired networks such as local area networks (LANs), wide area networks (WANs), metropolitan area networks (MANs), and integrated service digital networks (ISDNs), wireless LANs, CDMA, Bluetooth, and satellite communications. It may cover a wireless network such as, but the scope of the present invention is not limited thereto. In addition, the network 400 may transmit and receive information using short-range communication and/or long-distance communication. Here, short-range communication may include Bluetooth, radio frequency identification (RFID), infrared data association (IrDA), ultra-wideband (UWB), ZigBee, and wireless fidelity (Wi-Fi) technologies, and Communication includes code division multiple access (CDMA), frequency division multiple access (FDMA), time division multiple access (TDMA), orthogonal frequency division multiple access (OFDMA), single carrier frequency division multiple access (SC-FDMA) technology. I can.

Network 400 may include connections of network elements such as hubs, bridges, routers, switches and gateways. Network 400 may include one or more connected networks, such as a multi-network environment, including a public network such as the Internet and a private network such as a secure corporate private network. Access to network 400 may be provided through one or more wired or wireless access networks. Furthermore, the network 400 may support an Internet of Things (IoT) network and/or 5G communication that exchanges and processes information between distributed components such as objects.

2 is a schematic block diagram of an apparatus for controlling an air conditioner according to an embodiment of the present invention. In the following description, portions that overlap with the description of FIG. 1 will be omitted. 2, the air conditioner 100 includes an indoor unit 110, an outdoor unit 120 including a compressor 121, a refrigerant circulation line 130, a communication unit 140, a diagnosis unit 150, and a memory ( 160) and a control unit 170 may be included. Descriptions of the indoor unit 110, outdoor unit 120, and refrigerant circulation line 130 are the same as those described above, and thus will be omitted.

The communication unit 140 interworks with the network 400 to provide a communication interface necessary to provide a transmission/reception signal between the air conditioner 100 and/or the user terminal 200 and/or the server 300 in the form of packet data. I can. Furthermore, the communication unit 140 may serve to receive a predetermined information request signal from the user terminal 200 and transmit information processed by the air conditioner 100 to the user terminal 200. I can. In addition, the communication unit 140 may transmit a predetermined information request signal from the user terminal 200 to the server 300, receive a response signal processed by the server 300, and transmit it to the user terminal 200. In addition, the communication unit 140 may be a device including hardware and software necessary for transmitting and receiving a signal such as a control signal or a data signal through a wired or wireless connection with another network device.

In addition, the communication unit 140 may support various intelligent communication of things (internet of things (IoT), internet of everything (IoE), internet of small things (IoST), etc.)), and machine to machine (M2M) communication, V2X ( vehicle to everything communication) communication, device to device (D2D) communication, etc. may be supported.

The diagnosis unit 150 may obtain compressor operation data in order to diagnose a failure of the compressor 121 and predict a failure of the compressor 121.

The diagnosis unit 150 may classify the state information of the compressor 121 corresponding to the compressor operation data using a deep learning model based on unsupervised learning that has been trained in advance to classify the state information of the compressor by inputting the compressor operation data as input. have.

The diagnosis unit 150 may repeat the classification operation of the classification unit every predetermined period, and obtain a classification result of state information of the compressor in batch units.

The diagnosis unit 150 determines the normal operation of the compressor 121, a failure of the compressor 121, and the compressor 121 based on the statistical information on the classification result of the status information of the compressor 121 obtained in a batch unit. It is possible to determine the final state of the compressor 121 including one of the impending failure of ).

The diagnosis unit 150 may actively control the operation of the compressor 121 by receiving the final state of the compressor 121 as a feedback.

The diagnosis unit 150 may normalize the obtained compressor operation data to a certain range and input it to a deep learning model based on unsupervised learning. When the final state of the compressor 121 is one of a failure of the compressor 121 or an imminent failure of the compressor 121, the diagnosis unit 150 may generate an error code and display it on the air conditioner 100. The diagnostic unit 150 transmits the final state of the compressor 121 to the server 300, and the server 300 receives the final state of the compressor 121 and remotely manages and controls the air conditioner 100. I can.

In the present embodiment, the diagnosis unit 150 may perform learning in connection with the control unit 170 or may receive a learning result from the control unit 170. In this embodiment, the diagnosis unit 150 may be provided outside the control unit 170 as shown in FIG. 2, or may be provided inside the control unit 170 to operate like the control unit 170, or the server of FIG. 1 It may be provided inside (300). Details of the diagnosis unit 150 will be described below with reference to FIGS. 3 to 6.

The memory 160 may store various types of information necessary for the operation of the air conditioner 100 and store control software capable of operating the air conditioner 100, and may include a volatile or nonvolatile recording medium. have. In addition, the memory 160 may store information processed by the diagnosis unit 150.

In addition, the memory 160 is a deep learning model based on unsupervised learning that is pretrained to classify state information of the compressor by inputting a command to be executed by the diagnostic unit 150, for example, a command for acquiring compressor operation data, and compressor operation data. Using a command for classifying the status information of the compressor 121 corresponding to the compressor operation data, a command for repeating the status information of the compressor 121 at regular intervals, and a batch of the status information classification result of the compressor 121 Normal operation of the compressor 121, failure of the compressor 121, and impending failure of the compressor 121 based on the command acquired in units and the statistical information on the classification result of the status information of the compressor 121 acquired in batch units. A command for determining the final state of the compressor including one of the commands, a command for actively controlling the operation of the compressor by receiving the final state of the compressor 121 as a feedback, a command for normalizing the compressor operation data to a certain range, the compressor 121 A command for transmitting the final state of) to the server 300, and when the final state of the compressor 121 is one of failure or imminent failure, an error code may be generated and displayed on the air conditioner, and the like may be stored.

Here, the memory 160 may include a magnetic storage medium or a flash storage medium, but the scope of the present invention is not limited thereto. The memory 160 may include internal memory and/or external memory, and volatile memory such as DRAM, SRAM, or SDRAM, one time programmable ROM (OTPROM), PROM, EPROM, EEPROM, mask ROM, flash ROM, Non-volatile memory such as NAND flash memory, or NOR flash memory, SSD. A flash drive such as a compact flash (CF) card, an SD card, a Micro-SD card, a Mini-SD card, an Xd card, or a memory stick, or a storage device such as an HDD.

As a kind of central processing unit, the control unit 170 may control the operation of the entire air conditioner 100 by driving control software installed in the memory 160. The control unit 170 may include all types of devices capable of processing data, such as a processor. Here, the'processor' may refer to a data processing device embedded in hardware, which has a circuit physically structured to perform a function represented by a code or instruction included in a program. As an example of a data processing device built into the hardware as described above, a microprocessor, a central processing unit (CPU), a processor core, a multiprocessor, and an application-specific integrated (ASIC) circuit), a field programmable gate array (FPGA), and the like, but the scope of the present invention is not limited thereto.

In this embodiment, the control unit 170 may perform machine learning such as deep learning on the compressor operation data so that the air conditioner 100 can determine the state (normal, failure, impending failure) of the compressor 121. (machine learning) may be performed, and the memory 160 may store data used for machine learning, result data, and the like.

Deep learning technology, a kind of machine learning, can learn by going down to the deep level in multiple stages based on data. Deep learning can represent a set of machine learning algorithms that extract core data from a plurality of data as the level increases.

The deep learning structure may include an artificial neural network (ANN), and for example, the deep learning structure consists of a deep neural network (DNN) such as a convolutional neural network (CNN), a recurrent neural network (RNN), and a deep belief network (DBN). Can be. The deep learning structure according to the present embodiment may use various known structures. For example, the deep learning structure according to the present invention may include CNN, RNN, DBN, and the like. RNN is widely used for natural language processing, etc., and is an effective structure for processing time-series data that changes with the passage of time, and can construct an artificial neural network structure by stacking layers every moment. The DBN may include a deep learning structure constituted by stacking RBM (restricted boltzman machine), which is a deep learning technique, in multiple layers. By repeating RBM learning, when a certain number of layers is reached, a DBN having the corresponding number of layers can be configured. CNN can include a model that simulates human brain function, which is made based on the assumption that when a person recognizes an object, the basic features of the object are extracted, and then the brain undergoes complex calculations and recognizes the object based on the result. have.

On the other hand, learning of the artificial neural network can be accomplished by adjusting the weight of the connection line between nodes (if necessary, adjusting the bias value) so that a desired output is produced for a given input. In addition, the artificial neural network can continuously update the weight value by learning. In addition, a method such as back propagation may be used for learning of the artificial neural network.

Meanwhile, the controller 170 may be equipped with an artificial neural network, and grouping of state information of the compressor based on machine learning using received compressor operation data as input data may be performed.

The controller 170 may include an artificial neural network, for example, a deep neural network (DNN) such as CNN, RNN, and DBN, and may learn a deep neural network. Supervised learning, unsupervised learning, and reinforcement learning can all be used as a machine learning method of such an artificial neural network. The controller 170 may control to update the artificial neural network structure after learning according to a setting.

Supervised learning refers to a method of training an artificial neural network when a label for training data is given, and a label indicates the correct answer (or result value) that the artificial neural network must infer when training data is input to the artificial neural network. It can mean. Unsupervised learning may mean a method of training an artificial neural network in a state in which a label for training data is not given. Reinforcement learning may mean a learning method in which an agent defined in a certain environment learns to select an action or sequence of actions that maximizes the cumulative reward in each state. In the present embodiment, unsupervised learning may be performed to diagnose a failure of the compressor 121 and predict a failure of the compressor 121, but is not limited thereto.

3 is a schematic block diagram of a diagnosis unit according to an embodiment of the air conditioner control apparatus of FIG. 2, FIGS. 4 and 5 are exemplary views showing output results of a learning unit of the diagnosis unit of FIG. 3, and FIG. 6 is 3 is an exemplary view of output data of a classification unit for determining a final state of a compressor by a determination unit of the diagnosis unit of FIG. 3. In the following description, portions overlapping with the descriptions of FIGS. 1 and 2 will be omitted.

3, the diagnosis unit 150 includes an acquisition unit 151, a preprocessor 152, a classification unit 153, a determination unit 154, a compressor control unit 155, a diagnosis result transmission unit 156, and An error code output unit 157 may be included. As an alternative embodiment, the diagnostic unit 150 may include one or more processors. As an alternative embodiment, the acquisition unit 151 to the error code output unit 157 may correspond to one or more processors. As an alternative embodiment, the acquisition unit 151 to the error code output unit 157 may correspond to software components configured to be executed by one or more processors.

The acquisition unit 151 may acquire compressor operation data in order to diagnose a failure of the compressor 121 and predict a failure of the compressor 121. The compressor operation data may be stored in a memory (160 in FIG. 2) or may be obtained from the compressor 121 in real time. Here, the compressor operation data may include data on the operation history of the compressor 121 and data on the operation state of the compressor 121. The data on the operating state of the dual compressor 121 may include, for example, the pressure (high pressure, low pressure) of the compressor 121, the discharge temperature, the evaporation temperature, the compression ratio, the compressor frequency, the input current, the phase current, etc. by factor. I can.

The preprocessor 152 may process the compressor operation data acquired by the acquisition unit 151 into a form suitable for input of a deep learning model based on unsupervised learning. Since the unit and range of compressor operation data are different for each factor, it is necessary to normalize it to a certain range. The preprocessor 152 may normalize the compressor operation data using a minimum-maximum normalization equation as shown in Equation 1 below.

Here, X may represent compressor operation data for each factor, that is, input data. Xnorm can represent normalized input data. Xmax may represent the maximum value of the input data. Xmin may represent the minimum value of the input data. (a, b) may be a normalization range, for example, 0 to 1.

If the input data deviates from the minimum and maximum values of the existing training data, when applying the minimum-maximum normalization, the preprocessing unit 152 performs the input data as shown in Equation 2 below. It can be automatically adjusted to the minimum/maximum value.

The preprocessor 152 may include a missing value deletion filter and/or a steady state detection filter. The missing value deletion filter may delete the data vector so that it cannot be used in an unsupervised learning-based deep learning model when there is a missing value in the input data. If the input data is a k-dimensional vector or the new input data is an m-dimensional vector (m<k), the new input data may be deleted as shown in Equation 3 below.

The steady state data filter adopts one or more of an exponentially weighted moving average filter and a method to check whether a target actuator value has been reached so that only operation data in which the heating and cooling cycle is in a stable state is adopted to ensure the safety of the compressor operation data. Can be judged. The exponential weighted moving average filter may be determined to be in a stable state when the filter error is less than or equal to a predetermined filter error value during a specific period T as shown in Equation 3 below.

Here, α may represent a constant (eg, 0 to 1) indicating how much the change in the filter value will be reflected. In addition, ε is a threshold and may be different for each compressor operation data factor. For example, when the compressor operation data is a frequency factor, ε may range from 0 to 100 Hz.

)이 제어 목표값(

)에 일정 값(ε)으로 근접할 경우, 안정상태라고 판단하며, 판단 조건은 하기 수학식 5와 같다.In addition, the determination of whether the target actuator value has been reached or not is determined by the actuator value (

) Is the control target value (

), when it is close to a certain value (ε), it is determined as a stable state, and the determination condition is as shown in Equation 5 below.

The classification unit 153 may classify the state information of the compressor 121 corresponding to the compressor operation data using a deep learning model based on unsupervised learning that has been trained in advance to classify the state information of the compressor by inputting the compressor operation data as input. have. Here, the deep learning model based on unsupervised learning is an auto encoder, k-means clustering, self-organizing map, deep belief network, and support vector machine. vector machine) or the like.

In addition, depending on the type of algorithm, the state information of the compressor that is finally output from the deep learning model based on unsupervised learning can be classified into normal, faulty, and faulty clinical states, and the root mean square error of how far out of the normal state is ( It can also be expressed in the form of MSE: mean squared error.

4 is an exemplary diagram showing the output result of the classifier 153 according to an embodiment, an example of the output result of the k-means clustering model, which is one of the unsupervised learning-based deep learning models, and data as one of three groups. This is an example of categorizing (k=3) and displaying it. The k-means clustering model of FIG. 4 performs a task of classifying new compressor operation data into the closest group in terms of distance among the three central points (red triangles), and if the new compressor operation data is closest to the cluster classified as a failure It is classified as failure data, which may indicate compressor failure.

5 is an exemplary diagram showing an output result of the classification unit 153 according to another embodiment, and is an output example of an algorithm in the form of outputting the similarity with normal data as a mean squared error, unsupervised learning The output result of the noise reduction auto encoder, one of the base deep learning models, is shown. The auto encoder is learned with the aim of reproducing the compressor operation data in which the state of the compressor 121 is normal as it is, and the reproduction error can be calculated as the root of the mean square. If the inputted compressor operation data is normal, the root mean square error will be less than a certain threshold, and if fault data is entered as input data, the root mean square error, that is, the reproduction error will increase beyond a certain threshold. Accordingly, as shown in FIG. 5, data below the threshold of the root mean square error are normal data, and data that deviate from the threshold for each step can be classified into suspected failure and impending failure, respectively.

These models based on unsupervised learning can compare the distribution between the input data and the normal data based on the distribution of the data in the normal state to determine whether there is an abnormality in the data, even if the failure data is extremely insufficient. It is suitable for the machine 100. In the case of insufficient failure data when verifying a deep learning model based on unsupervised learning for diagnosing the condition of the compressor 121, it may be difficult to verify the model.

In this embodiment, the entire data set is divided into k partial sets using k-fold cross validation, and then the first fold is used as a test data set to be verified, and the remaining data sets are used as training data. Can be used in three. And you can use the second fold as a test data set and use the rest of the data set for training. By repeating this process, the final performance of the model can be verified by calculating the average of the total k test accuracy.

In the present embodiment, the classification unit 153 may repeatedly perform the classification operation of the state information of the compressor 121 at regular intervals, and the determination unit 154 calculates the classification result of the state information of the compressor 121 in a batch unit. Obtained as

The determination unit 154 is one of the normal operation of the compressor 121, the failure of the compressor 121, and the impending failure of the compressor 121 based on the statistical information on the classification result of the status information of the compressor acquired in batch units. It is possible to determine the final state of the compressor including.

The determination unit 154 can determine the final state of the compressor 121 by statistically calculating the output results for a specific batch size (unit) without determining each output result of the classification unit 153 that individually enters. have. For example, when the determination unit 154 obtains the output result of the unsupervised learning-based deep learning model for n number of compressor operation data through the classification unit 153, the determination unit 154 determines the final state of the compressor using the mode. can do. If one or more states are tied, the determination result may be output as N/A (not available), and the mode may be compared from the placement result of the next round.

6 is an exemplary view of output data of a classification unit for determining a final state of a compressor by a determination unit of the diagnosis unit of FIG. Referring to FIG. 6A, in the case of batch 1, since the mode is <failure>, the determination unit 154 may diagnose the final state of the compressor 121 as a malfunction of the compressor 121 for the batch 1. Referring to FIG. 6B, in the case of batch 2, since the mode values are <Fault> and <Normal>, the determination unit 154 outputs the final state of the compressor 121 as N/A (not available) for the batch 1 However, the judgment may be withheld until the next batch. Although not shown, in the case of batch 3, when the mode is <imminent failure>, the determination unit 154 may diagnose the final state of the compressor 121 as an impending failure of the compressor 121 for the batch 3.

The compressor control unit 155 may receive the final state of the compressor 121 as a feedback from the determination unit 154 and actively control the operation of the compressor 121.

As a result of the determination of the determination unit 154, when the final state of the compressor 121 is determined to be a failure of the compressor 121, the compressor control unit 155 takes preliminary measures to prevent the compressor 121 from being checked by a service engineer. It is possible to prevent further damage and control the air conditioner 100 to be continuously operated.

As an example, as a result of the determination of the determination unit 154, when it is determined that the final state of the compressor 121 is impending failure of the compressor 121, the compressor control unit 155 operates the compressor 121 under the phase current limiting control condition. By doing so, it is possible to prevent the overcurrent phenomenon of the compressor 121 and prevent further burnout of the compressor 121.

The diagnosis result transmission unit 156 may transmit the final state of the compressor 121 according to the determination result of the determination unit 154 to the server 300. In particular, when the determination result of the determination unit 154 determines that the final state of the compressor 121 is a failure or an impending failure of the compressor 121, the diagnosis result transmission unit 156 transmits the final state of the compressor 121 to the server. By sending it to 300, the service engineer can take action to check the air conditioner 100.

The error code output unit 157 serves to output an error code corresponding to the determination result of the determination unit 154, when it is determined that the final state of the compressor 121 is a failure or an impending failure of the compressor 121, and , The information of the outdoor unit 120 may be transferred to a printed circuit board (PCB). The received error code may be output through a 7-segment display device (not shown) on the main PCB inside the outdoor unit 120.

7 is a flowchart illustrating a method of controlling an air conditioner according to an embodiment of the present invention. In the following description, portions overlapping with the descriptions of FIGS. 1 to 6 will be omitted.

Referring to FIG. 7, in step S710, the air conditioner 100 acquires compressor operation data in order to diagnose a failure of the compressor 121 and predict a failure of the compressor 121. The compressor operation data may be stored in the memory 160 or may be acquired from the compressor 121 in real time.

In step S720, the air conditioner 100 normalizes the compressor operation data to a predetermined range. The air conditioner 100 may filter the compressor operation data including the missing values to be excluded from the input of the unsupervised learning-based deep learning model. The air conditioner 100 may determine a stable state of the compressor operation data by using at least one of an exponentially weighted moving average filter and a method of checking whether a target actuator value has been reached.

In step S730, the air conditioner 100 classifies the state information of the compressor corresponding to the operation data of the compressor using a deep learning model based on unsupervised learning previously trained to classify the state information of the compressor by inputting the compressor operation data as input. do. Here, the classification of the state information of the compressor may be repeated every predetermined period, and the air conditioner 100 may obtain a result of classifying the state information of the compressor in batch units.

In step S740, the air conditioner 100 performs normal operation of the compressor 121, a failure of the compressor 121, and the compressor 121 based on the statistical information on the classification result of the state information of the compressor obtained in batch units. The final state of the compressor 121 including one of the imminent failures of is determined. The air conditioner 100 is the final compressor including one of the normal operation of the compressor corresponding to the mode, the failure of the compressor, and the impending compressor failure from the classification result of the status information of the compressor obtained in batch units. You can judge the state.

In step S750, the air conditioner 100 transmits the final state of the compressor 121 to the server 300 as a diagnosis result.

In step S760, the air conditioner 100 determines the failure of the compressor 121 or the impending failure of the compressor 121 in response to the final state information of the compressor 121.

In step S770, the air conditioner 100 controls the compressor 121 in an emergency when a failure of the compressor 121 or a failure of the compressor 121 is imminent as a result of the determination. When the air conditioner 100 determines that the final state of the compressor 121 is a failure of the compressor 121, the air conditioner 100 takes preliminary measures and adds the compressor 121 until the service engineer checks it. Burning can be prevented and the air conditioner 100 can be controlled to be continuously operated. As an example, when the final state of the compressor 121 is determined to be a failure of the compressor 121, the air conditioner 100 operates the compressor 121 under the phase current limiting control condition. , It is possible to prevent an overcurrent phenomenon of the compressor 121 and prevent further burnout of the compressor 121.

In step S780, when the final state of the compressor 121 is determined to be a failure or imminent failure of the compressor 121, the air conditioner 100 outputs an error code corresponding thereto. The air conditioner 100 may output an error code through a 7-segment display device (not shown) on the main PCB inside the outdoor unit 120.

The embodiment according to the present invention described above may be implemented in the form of a computer program that can be executed through various components on a computer, and such a computer program may be recorded in a computer-readable medium. In this case, the medium is a magnetic medium such as a hard disk, a floppy disk, and a magnetic tape, an optical recording medium such as a CD-ROM and a DVD, a magnetic-optical medium such as a floptical disk, and a ROM. It may include a hardware device specially configured to store and execute program instructions, such as, RAM, flash memory, and the like.

Meanwhile, the computer program may be specially designed and configured for the present invention, or may be known and usable to a person skilled in the computer software field. Examples of the computer program may include not only machine language codes such as those produced by a compiler, but also high-level language codes that can be executed by a computer using an interpreter or the like.

In the specification of the present invention (especially in the claims), the use of the term "above" and the reference term similar thereto may correspond to both the singular and the plural. In addition, when a range is described in the present invention, the invention to which individual values falling within the range are applied (unless otherwise stated), and each individual value constituting the range is described in the detailed description of the invention. Same as.

If there is no explicit order or contradictory description of the steps constituting the method according to the present invention, the steps may be performed in a suitable order. The present invention is not necessarily limited according to the order of description of the steps. The use of all examples or illustrative terms (for example, etc.) in the present invention is merely for describing the present invention in detail, and the scope of the present invention is limited by the above examples or illustrative terms unless limited by the claims. It does not become. In addition, those skilled in the art can recognize that various modifications, combinations, and changes may be configured according to design conditions and factors within the scope of the appended claims or their equivalents.

Accordingly, the spirit of the present invention is limited to the above-described embodiments and should not be defined, and all ranges equivalent to or equivalently changed from the claims to be described later as well as the claims to be described later are the scope of the spirit of the present invention. It will be said to belong to.

100: air conditioner
200: user terminal
300: server
400: network

Claims (20)

As an air conditioner control method,
Acquiring compressor operation data from a compressor included in the air conditioner;
Classifying state information of a compressor corresponding to the compressor operation data using an unsupervised learning-based deep learning model trained in advance to classify state information of a compressor by inputting compressor operation data as input;
Repeating the classifying step every predetermined period;
Obtaining a classification result of the state information of the compressor in a batch unit;
Determining a final state of a compressor including one of a normal operation of the compressor, a failure of the compressor, and an impending failure of the compressor, based on statistical information on a classification result of the condition information of the compressor obtained in units of batches; And
Including the step of actively controlling the operation of the compressor by receiving the final state of the compressor as a feedback,
Air conditioner control method. The method of claim 1,
After the step of obtaining the compressor operation data,
Further comprising the step of normalizing the compressor operation data to a predetermined range and inputting it to the unsupervised learning-based deep learning model,
Air conditioner control method. The method of claim 2,
Filtering the compressor operation data including missing values to be excluded from the input of the unsupervised learning-based deep learning model,
Air conditioner control method. The method of claim 2,
Determining a stable state of the compressor operation data using at least one of an exponentially weighted moving average filter and a method of confirming whether a target actuator value has been reached,
Air conditioner control method. The method of claim 1,
The step of determining the final state of the compressor,
The final state of the compressor including one of a normal operation of the compressor, a failure of the compressor, and an impending failure of the compressor corresponding to a mode from the classification result of the status information of the compressor obtained in the batch unit. Including the step of determining,
Air conditioner control method. The method of claim 1,
After the step of determining the final state of the compressor,
Generating an error code according to the final state of the compressor, which is one of a failure of the compressor or an impending failure of the compressor, and displaying the error code on the air conditioner.
Air conditioner control method. The method of claim 1,
The step of actively controlling the operation of the compressor,
Comprising the step of controlling the compressor to a phase current limiting condition in response to the final state of the compressor, including the impending failure of the compressor,
Air conditioner control method. A computer-readable recording medium storing a computer program for executing the method of claim 1 using a computer. As an air conditioner control device,
An acquisition unit acquiring compressor operation data from a compressor included in the air conditioner;
A classification unit for classifying state information of the compressor corresponding to the compressor operation data using a deep learning model based on unsupervised learning trained in advance to classify state information of the compressor by inputting the compressor operation data;
The classification operation of the classification unit is repeated at regular intervals, and the classification result of the state information of the compressor is obtained in batch units, and based on the statistical information on the classification result of the state information of the compressor obtained in the batch unit, the A determination unit for determining a final state of the compressor including one of a normal operation of the compressor, a failure of the compressor, and an impending failure of the compressor; And
Comprising a compressor control unit for actively controlling the operation of the compressor by receiving the final state of the compressor as a feedback,
Air conditioner control unit. The method of claim 9,
After acquiring the compressor operation data, further comprising a preprocessor for normalizing the compressor operation data to a predetermined range and inputting it to the unsupervised learning-based deep learning model,
Air conditioner control unit. The method of claim 10,
The pretreatment unit,
Configured to filter the compressor operation data including missing values to be excluded from the input of the unsupervised learning-based deep learning model,
Air conditioner control unit. The method of claim 10,
The pretreatment unit,
Configured to determine a stable state of the compressor operation data by using at least one of an exponentially weighted moving average filter and a method of confirming whether a target actuator value has been reached,
Air conditioner control unit. The method of claim 9,
The determination unit,
The final state of the compressor including one of a normal operation of the compressor, a failure of the compressor, and an impending failure of the compressor corresponding to a mode from the classification result of the status information of the compressor obtained in the batch unit. Configured to judge,
Air conditioner control unit. The method of claim 9,
Further comprising an error code output unit for generating an error code and displaying it on the air conditioner according to the final state of the compressor, which is one of a failure of the compressor or an imminent failure of the compressor,
Air conditioner control unit. The method of claim 9,
The compressor control unit,
Configured to control the compressor under a phase current limiting condition in response to a final state of the compressor including the impending failure of the compressor,
Air conditioner control unit. As an air conditioner control device,
One or more processors; And
Including a memory connected to the one or more processors,
The memory,
Obtaining compressor operation data from a compressor included in the air conditioner,
Classify the state information of the compressor corresponding to the compressor operation data by using an unsupervised learning-based deep learning model trained in advance to classify the state information of the compressor by inputting the compressor operation data as input,
Iterate the classification of the state information of the compressor every predetermined period,
Obtaining the classification result of the state information of the compressor in batches,
Determine a final state of the compressor including one of a normal operation of the compressor, a failure of the compressor, and an impending failure of the compressor based on the statistical information on the classification result of the condition information of the compressor obtained in the batch unit,
Storing a command that causes the final state of the compressor to be input as a feedback and to actively control the operation of the compressor,
Air conditioner control unit. The method of claim 16,
The above command is:
After acquiring the compressor operation data, further causing the compressor operation data to be normalized to a certain range and input into the unsupervised learning-based deep learning model,
Air conditioner control unit. The method of claim 16,
The above command is:
When determining the final state of the compressor, normal operation of the compressor corresponding to a mode from the classification result of the state information of the compressor obtained in the batch unit, a failure of the compressor, and a failure of the compressor are imminent. Causing to determine the final state of the compressor comprising one of,
Air conditioner control unit. The method of claim 16,
The above command is:
After determining the final state of the compressor, according to the final state of the compressor, which is one of a failure of the compressor or an impending failure of the compressor, further causing an error code to be generated and displayed on the air conditioner,
Air conditioner control unit. The method of claim 16,
The above command is:
Causing the compressor to be controlled under a phase current limiting condition in response to a final state of the compressor including the impending failure of the compressor,
Air conditioner control unit.

Images