Apparatus for controlling air conditioner

Abstract

According to an embodiment of the present disclosure, an apparatus for controlling an air conditioner includes: a driving unit for driving a compressor, a fan, and a valve of the air conditioner; a detection unit including one or more sensors so as to detect one or more pieces of state information about the driving state of the compressor and the operation state of the air conditioner; an estimation unit which uses the detection result from the detection unit as an input to a pre-trained artificial neural network and estimates a normal probability value for the state at one or more points in a pipe of the air conditioner on the basis of an output from the artificial neural network; and a determination unit for determining whether a blockage has occurred at the one or more points, on the basis of the normal probability value. The estimation unit estimates the normal probability value according to a set period. The determination unit increases an abnormality count when the normal probability value is equal to or less than a probability reference value, and when the abnormality count is equal to or greater than a count reference value, determines that a blockage has occurred at the corresponding point.

Description

Control unit of air conditioner

[0001] The present invention relates to a control device for controlling an air conditioner, and more specifically, to a control device for an air conditioner capable of diagnosing and controlling the status of the air conditioner.

[0002] Air conditioners, chillers, and HVAC systems, among others, utilize a heat exchange system based on a refrigerant cycle to control indoor air temperature. The heat exchange system consists of an indoor and outdoor unit. The indoor unit comprises an indoor heat exchanger and a fan, while the outdoor unit comprises an expansion valve (EEV), an outdoor heat exchanger, and a compressor. Piping is installed between the indoor and outdoor units to allow refrigerant to circulate. Depending on the operating environment, the piping can be as long as 100 meters, or the elevation difference can be tens of meters. To provide both cooling in summer and heating in winter, the direction of refrigerant circulation must be changed. Therefore, (system) air conditioners have a four-way valve on the outdoor unit.

[0003] Meanwhile, the compressor is a device that provides the energy (pressure, temperature) required for refrigerant circulation. It compresses low-temperature, low-pressure refrigerant and releases high-temperature, high-pressure refrigerant. While the compressor supplies energy to the cycle regardless of whether it's cooling or heating, the flow of refrigerant / energy within the cycle must vary depending on whether it's cooling or heating. To achieve this, a four-way valve changes the direction of the piping connection between the compressor's input and output terminals and the indoor and outdoor units. The low-temperature refrigerant entering the compressor is designed to cool the motor within the compressor.

[0004] At this time, the pipes can become clogged for various reasons, such as foreign substances or pipe damage. Operating with the pipes clogged is called vacuum operation. In vacuum operation, that is, when the pipes become clogged, the refrigerant input to the compressor decreases, and the phenomenon of increased heat generation and reduced compressor cooling capacity occurs due to decreased compression efficiency. The increase in heat generation and decreased heat dissipation continuously raise the temperature of the compressor, and it can exceed 200 degrees Celsius over time. Since the heat resistance limit of the components constituting the compressor is generally lower than 200 degrees, the phenomenon of reduced refrigerant volume due to pipe clogging can cause permanent damage to the compressor due to high temperature.

[0005] If the compressor's suction pressure could be measured via a sensor, such pipe blockages or vacuum operation could be detected and prevented in advance. However, due to the high cost of pressure sensors, some products use only one pressure sensor to reduce material costs. Using only one pressure sensor can result in suction pressure not being measured depending on the air conditioner's operating mode (cooling or heating).

[0006] Monitoring the internal temperature of the compressor would allow for detection of high-temperature rises. However, compressors are structurally sealed, making it difficult to incorporate temperature sensors internally. Inserting temperature sensors is costly, and even if installed, their lifespan is limited due to damage caused by internal temperatures. Furthermore, it is difficult to accurately detect pipe blockages or vacuum operation.

[0007] Even if a pressure sensor or temperature sensor is installed, if a blockage occurs at a point in the pipe where no sensor is installed, such as between the temperature sensor installed in the suction section and the compressor, the blockage cannot be detected based on the sensing results alone, making it difficult to deal with the blockage, which is the fundamental cause of the rise in temperature.

[0008] Meanwhile, as a prior art related to such pipe blockage, vacuum operation, and high temperature rise, prior art document 1 (Korean Patent Registration No. 10-2317340 B1) discloses a technology for detecting poor circulation (leakage, blockage) of refrigerant based on temperature of a heat exchanger and driving current information of a compressor, and prior art document 2 (Korean Patent Publication No. 10-2014-0018782 A) discloses a technology for determining poor circulation of refrigerant and high-pressure blockage based on temperature information through a test run mode.

[0009] While prior art document 1 can detect poor refrigerant circulation, it has limitations in that it cannot detect clogging in pipes or vacuum operation. Prior art document 2 also cannot detect clogging in pipes, limiting the ability to take appropriate action based on the location of the clog.

[0010] In other words, structural and design constraints of the compressor have made it difficult to prevent pipe blockages and the resulting vacuum operation and high-temperature damage. Furthermore, no technology has been proposed to accurately detect these pipe blockages, vacuum operation, and high-temperature damage. Consequently, the lifespan, safety, and reliability of air conditioners have been limited, and a technology capable of improving these issues has been required.

[0011] Furthermore, model-based diagnosis can be challenging when diagnosing abnormalities such as blockages in the air conditioning cycle, as modeling the complex physical phenomena of the cycle can be challenging. In such cases, data-driven artificial neural network algorithms can be utilized, particularly those that utilize probabilistic neural networks to output probability values ​​for normal or abnormal conditions. However, relying solely on probability values ​​can lead to misdiagnoses, or the normal / fault judgment can be reversed when the probability values ​​fluctuate beyond the reference value.

[0012] The problem that the present disclosure seeks to solve is to provide a control device for an air conditioner that can further improve the accuracy of diagnosis of the condition of the air conditioner by post-processing the diagnosis results of an artificial neural network.

[0013] The problem that the present disclosure seeks to solve is to provide a control device for an air conditioner that can accurately detect the occurrence of a blockage in a pipe of the air conditioner and prevent vacuum operation.

[0014] The problem that the present disclosure seeks to solve is to provide a control device for an air conditioner that can detect vacuum operation of the air conditioner and control the air conditioner in response to the vacuum operation.

[0015] The problem that the present disclosure seeks to solve is to provide a control device for an air conditioner capable of determining the point at which a blockage in a pipe occurs.

[0016] The problem to be solved by the present disclosure is to provide a control device for an air conditioner capable of determining the occurrence of a blockage in a pipe without having a sensor for detecting the suction pressure or internal temperature of the compressor.

[0017] The tasks of the present disclosure are not limited to the tasks mentioned above, and other tasks not mentioned will be clearly understood by those skilled in the art from the description below.

[0018] In order to achieve the above object, a control device of an air conditioner according to an embodiment of the present disclosure includes a driving unit that drives a compressor, a fan, and a valve of the air conditioner, a detection unit that detects one or more status information on an operating state of the compressor and an operating state of the air conditioner, including one or more sensors, an estimation unit that uses a detection result of the detection unit as an input of a pre-trained artificial neural network and estimates a normal probability value for a state of one or more points in a pipe of the air conditioner based on an output of the artificial neural network, and a determination unit that determines whether a blockage has occurred at the one or more points based on the normal probability value, wherein the estimation unit estimates the normal probability value according to a set cycle, and the determination unit increases an abnormality count when the normal probability value is less than or equal to a probability reference value, and determines that a blockage has occurred at the corresponding point when the abnormality count is greater than or equal to the count reference value.

[0019] The above count reference value may be set such that the product of the count reference value and the cycle is greater than the detection time of a failure other than the blockage failure and is less than the high temperature burnout time of the compressor.

[0020] The above abnormal count can be set based on the above probability criterion.

[0021] When the above probability criterion is set to a first value, the above abnormal count may be set to a second value, and when the above probability criterion is set to a third value smaller than the first value, the above abnormal count may be set to a fourth value smaller than the second value.

[0022] According to one embodiment of the present disclosure, a control device for an air conditioner includes a driving unit for driving a compressor, a fan, and a valve of the air conditioner, a detection unit for detecting one or more status information on an operating state of the compressor and an operating state of the air conditioner, including one or more sensors, an estimation unit for estimating a probability value for a state of one or more points in a pipe of the air conditioner based on an output of the artificial neural network by using a detection result of the detection unit as an input of a pre-trained artificial neural network, and a determination unit for determining whether an abnormality has occurred based on the probability value, wherein the determination unit determines that an abnormality has occurred at the point if a state in which the probability value satisfies an abnormality occurrence criterion is maintained for a predetermined period of time or longer.

[0023] The above-mentioned period of time may be set to be greater than the time required to detect multiple abnormal conditions and to be less than the high-temperature burnout time of the compressor.

[0024] The above probability value may be a normal probability value indicating a probability that the one or more points are in a normal state, or an abnormal probability value indicating a probability that the one or more points are in an abnormal state.

[0025] When using the above normal probability value, the probability criterion included in the above abnormal occurrence criterion may be proportional to the above schedule time.

[0026] Meanwhile, the estimation unit can estimate the normal probability value when the operating frequency of the compressor is greater than or equal to the minimum frequency.

[0027] In addition, the estimation unit estimates the probability value according to a set period, and the period can correspond to the operation speed of the artificial neural network.

[0028] The control device of an air conditioner according to one embodiment of the present disclosure may further include a control unit that stops the compressor in an emergency when the abnormal count is greater than a count reference value.

[0029] The above driving state may include one or more of the driving speed, driving frequency, torque, driving current, driving voltage, ambient pressure, suction temperature, discharge temperature, internal temperature, external temperature, and control command of the compressor.

[0030] The above estimation unit can estimate a probability value of a point corresponding to at least one of the points between a sensor provided around the compressor and the compressor.

[0031] The above estimation unit can further estimate one or more pieces of undetected information that the detection unit does not detect during the driving state.

[0032] According to at least one of the embodiments of the present disclosure, the accuracy of diagnosing the condition of an air conditioner can be further improved by post-processing the diagnosis results of an artificial neural network.

[0033] According to at least one of the embodiments of the present disclosure, it is possible to accurately detect the occurrence of a blockage in a pipe of an air conditioner, thereby preventing vacuum operation.

[0034] According to at least one of the embodiments of the present disclosure, the vacuum operation of the air conditioner can be detected and the air conditioner can be controlled in response to the vacuum operation.

[0035] According to at least one of the embodiments of the present disclosure, it is possible to determine the point at which a blockage in a pipe occurs.

[0036] According to at least one of the embodiments of the present disclosure, it is possible to determine whether a blockage occurs in a pipe without having a sensor for detecting suction pressure or internal temperature of the compressor.

[0037] Meanwhile, various other effects will be disclosed directly or implicitly in the detailed description according to the embodiments of the present disclosure to be described later.

[0038] FIG. 1 is a block diagram showing the configuration of a control device of an air conditioner according to one embodiment of the present disclosure.

[0039] Figures 2 and 3 are schematic diagrams showing the configuration of an air conditioner according to one embodiment of the present disclosure.

[0040] FIG. 4 is a block diagram showing a specific configuration of a control device of an air conditioner according to an embodiment of the present disclosure.

[0041] FIGS. 5A to 5C are diagrams illustrating an artificial neural network of an estimation unit according to an embodiment of the present disclosure.

[0042] FIG. 6 is a diagram showing an example of an artificial neural network implementation of an estimation unit according to an embodiment of the present disclosure.

[0043] Figures 7a to 7d are diagrams illustrating waveforms of status information in a normal driving state.

[0044] Figures 8a to 8d are diagrams illustrating waveforms of state information when a blockage occurs.

[0045] Figures 9a to 9d are diagrams illustrating waveforms of state information when a blockage occurs.

[0046] FIG. 10 is a drawing for reference in an explanation of determining whether a blockage has occurred and determining the location of a blockage in a control device of an air conditioner according to one embodiment of the present disclosure.

[0047] FIG. 11 is a block diagram showing a process of determining whether a blockage has occurred and performing control in a control device of an air conditioner according to one embodiment of the present disclosure.

[0048] FIG. 12 is a flowchart illustrating a method for controlling an air conditioner according to an embodiment of the present disclosure.

[0049] FIGS. 13 to 16 are drawings for reference in the description of a control method for an air conditioner according to an embodiment of the present disclosure.

[0050] Hereinafter, embodiments disclosed in this specification will be described in detail with reference to the attached drawings. Regardless of the drawing numbers, identical or similar components are assigned the same reference numbers, and redundant descriptions thereof will be omitted.

[0051] The suffixes “module” and “part” used for components in the following description are given or used interchangeably only for the convenience of writing specifications, and do not have distinct meanings or roles in themselves.

[0052] Terms that include ordinal numbers, such as first, second, etc., may be used to describe various components, but the components are not limited by these terms. These terms are used solely to distinguish one component from another.

[0053] When a component is referred to as being "connected" or "connected" to another component, it should be understood that it may be directly connected or connected to that other component, but that there may be other components intervening. Conversely, when a component is referred to as being "directly connected" or "connected" to another component, it should be understood that there are no other components intervening.

[0054] In addition, the attached drawings are only intended to facilitate easy understanding of the embodiments disclosed in this specification, and the technical ideas disclosed in this specification are not limited by the attached drawings, and should be understood to include all modifications, equivalents, or substitutes included in the spirit and technical scope of the present invention.

[0055] First, a control device (hereinafter, “control device”) of an air conditioner according to an embodiment of the present disclosure is described.

[0056] The above control device may be a control device that controls an air conditioner that performs an air conditioning function.

[0057] The above control device can control the operation of the air conditioner by controlling one or more devices included in the air conditioner according to the operation mode of the air conditioner.

[0058] The above control device can control the operation of the air conditioner by controlling one or more of the compressor, fan, and valve included in the air conditioner.

[0059] Referring to FIG. 1, a control device (200) controls the air conditioner (100) including an outdoor unit (10) and an indoor unit (20) equipped with a compressor (11) for compressing refrigerant.

[0060] The above control device (200) includes a driving unit (210) that drives the compressor (11), fan (15) and valve (16) of the air conditioner (100), and a detection unit (220) that detects one or more status information about the driving status of the compressor (11) and the operating status of the air conditioner (100), including one or more sensors.

[0061] The above control device (200) includes an estimation unit (230) that estimates a probability value for the state of one or more points among the pipes of the air conditioner (100) based on the output of the artificial neural network by using the detection result of the detection unit (220) as an input of a pre-trained artificial neural network (NN), and a judgment unit (240) that determines whether an abnormality has occurred based on the probability value.

[0062] According to one embodiment of the present disclosure, the estimation unit (230) estimates a probability value for a normal state probability or a blockage occurrence probability of one or more points among the pipes of the air conditioner (100) based on the output of the artificial neural network by using the detection result of the detection unit (220) as an input of the pre-trained artificial neural network (NN). In addition, the judgment unit (240) determines whether a blockage has occurred at one or more points based on the estimation result of the estimation unit (230).

[0063] Meanwhile, the control device (200) further includes a control unit (250) that controls the driving unit (210) according to the judgment result of the judgment unit (240).

[0064] That is, the control device (200) drives the compressor (11), the fan (15), and the valve (16) through the drive unit (210), detects at least one piece of state information including the drive state and the operation state through the detection unit (220), estimates a state such as a normal state probability value or a probability value of blockage occurrence at at least one point of the pipe through the estimation unit (230), determines whether blockage has occurred at at least one point through the determination unit (240), and controls the drive unit (210) through the control unit (250) to control the operation of the air conditioner (100).

[0065] Here, the status information means information including the driving status and the operation status, and can be detected through one or more sensors included in the detection unit (220).

[0066] The above driving state may mean a state related to the driving of the compressor (11).

[0067] Additionally, the driving state may mean a control state in which the driving unit (210) controls the driving of the compressor (11).

[0068] The above driving state may include one or more of the driving speed, driving frequency, torque, driving current, driving voltage, ambient pressure, suction temperature, discharge temperature, internal temperature, external temperature, and control command of the compressor (11).

[0069] Here, the ambient pressure may include at least one of the pressure of the suction portion through which the refrigerant is introduced from the compressor (11), the pressure of the discharge portion through which the refrigerant is discharged from the compressor (11), and the pressure of the refrigerant pipe.

[0070] That is, the ambient pressure may include one or more of the suction pressure and discharge pressure of the compressor (11).

[0071] The above driving state may mean a state related to the operation of the air conditioner (100).

[0072] For example, it may be a state related to the temperature of the air conditioner (100).

[0073] The above driving condition may include one or more of the heat exchange cycle temperature and the ambient temperature of the air conditioner (100).

[0074] That is, the control device (200) can detect one or more of the operating speed, operating frequency, torque, driving current, driving voltage, ambient pressure, suction temperature, discharge temperature, internal temperature, external temperature, and control command of the compressor (11), the heat exchange cycle temperature of the air conditioner (100), and the ambient temperature through the detection unit (220) of the one or more sensors, and control the operation of the air conditioner (100) based on the detection result.

[0075] The air conditioner (100) whose operation is controlled by the above control device (200) may be as shown in FIGS. 2 and 3.

[0076] The above air conditioner (100), as shown in FIG. 2, includes an outdoor unit (10) and an indoor unit (20) connected to the outdoor unit (10) to perform air conditioning, and the outdoor unit (10) and the indoor unit (20) can be connected through a (refrigerant) pipe (30).

[0077] Fig. 2 shows an example of a stand-alone air conditioner, and the air conditioner (100) may be a wall-mounted, ceiling-mounted, or multi-air conditioner in addition to the stand-alone air conditioner shown in Fig. 2.

[0078] The air conditioner (100) including the outdoor unit (10), the indoor unit (20) and the refrigerant pipe (30) is composed of the compressor (11), the outdoor heat exchanger (12), the expansion valve (13) and the indoor heat exchanger (14) in the outdoor unit (10), as shown in FIG. 3.

[0079] A refrigeration cycle is provided, and a plurality of fans (15) are installed on the upper surface or side of the outdoor unit (10) to suck in outside air and exchange heat with the outdoor heat exchanger (12), and the indoor heat exchanger (14) may be connected to the refrigerant pipe (30) for supplying cold water or hot water to the indoor unit (20), and a refrigerant switching valve (16) may be provided at the outlet of the compressor (11) to switch the compressed refrigerant toward the outdoor heat exchanger (12) or the indoor heat exchanger (14) depending on the operating conditions.

[0080] The air conditioner (100) having the above configuration can be operated as a cooling device in the summer and as a heating device in the winter.

[0081] The above air conditioner (100) can be operated in a cooling operation mode in which it operates as an air conditioner or a heating operation mode in which it operates as a heater, depending on the setting of the operation mode.

[0082] When the above air conditioner (100) is operated in a cooling operation mode, the refrigerant compressed to a high temperature and high pressure in the compressor (11) is guided to the outdoor heat exchanger (12) by the refrigerant switching valve (16), where it exchanges heat with air in the outdoor heat exchanger (12) to radiate heat, and then is made low temperature and low pressure in the expansion valve (13), where it exchanges heat with water in the indoor heat exchanger (14), and is supplied to the indoor unit (20) that uses the heat-exchanged water as a cooling heat source.

[0083] In addition, when the air conditioner (100) is operated in a heating operation mode, the refrigerant switching valve (16) guides the refrigerant toward the indoor heat exchanger (14), so that the high temperature, high pressure refrigerant exchanges heat with water in the indoor heat exchanger (14), and the heat-exchanged water is supplied to the indoor unit (20) that uses it as a heating source.

[0084] The specific configuration of the control device (200) that controls the air conditioner (100) having such a configuration may be as shown in FIG. 4.

[0085] FIG. 4 illustrates an example of the control device (200), and the control device (200) may further include a configuration other than the configuration illustrated in FIG. 4, or may include a configuration other than the configuration illustrated in FIG. 4.

[0086] The above control device (200), as shown in FIG. 4, includes the driving unit (210), the detection unit (220), the estimation unit (230), the judgment unit (240), and the control unit (250), and controls the air conditioner (100).

[0087] Here, the estimation unit (230) may be included in the control unit (250).

[0088] Additionally, the judgment unit (240) may also be included in the control unit (250).

[0089] That is, one or more of the estimation unit (230) and the judgment unit (240) may be included in the control unit (250).

[0090] In this case, each of the estimation unit (230) and the judgment unit (240) may be formed as a calculation device that performs a dedicated calculation in the control unit (250).

[0091] The control device (200) may also further include one or more of an input unit (260) having one or more input means for receiving a command for operating the air conditioner (100), an output unit (270) having one or more output means for outputting information on the operating status of the air conditioner (100), a storage unit (280) for storing data of the air conditioner (100), and a communication unit (290) for transmitting and receiving data with a communication target device of the air conditioner (100).

[0092] The above input unit (260) is equipped with one or more buttons, switches, etc., and can receive setting commands, driving commands, etc. from a user, etc.

[0093] The above output unit (270) may be equipped with a display means for displaying various data, information, etc. in the form of characters or images, an audio output means, etc.

[0094] The above storage unit (280) is equipped with a memory such as an EEPROM, and can store data necessary for operating the outdoor unit and indoor unit, controlling the indoor fan, controlling the valve, etc., and can store operating information and status information.

[0095] The above communication unit (290) can transmit and receive data to and from the communication target device via wired or wireless communication.

[0096] Hereinafter, a specific embodiment of the control device (200) that detects the occurrence of a blockage in the pipe (30) and performs control in response to the occurrence of the blockage, including the driving unit (210), the detection unit (220), the estimation unit (230), the judgment unit (240), and the control unit (250), will be described with a focus on the following.

[0097] The above driving unit (210) can be controlled by the control unit (250).

[0098] The above driving unit (210) can be controlled by the control unit (250) to drive the compressor (11).

[0099] The above driving unit (210) can generate a driving signal according to a control command received from the control unit (250) and apply the driving signal to the compressor (11) to drive the compressor (11).

[0100] The driving unit (210) may also generate a driving signal for each of the devices constituting the outdoor unit (10) or the indoor unit (20) according to a control command authorized from the control unit (250), and apply each of the driving signals to the devices constituting the outdoor unit (10) or the indoor unit (20) to drive the devices constituting the outdoor unit (10) or the indoor unit (20).

[0101] The above driving unit (210) may include one or more of a compressor driving unit (211) that drives the compressor (11), a fan driving unit (212) that drives the fan (15), and a valve driving unit (213) that drives the valve (13, 16).

[0102] The above compressor driving unit (211) can apply power or a driving signal to drive the compressor (11) based on the input current to the compressor (11).

[0103] The compressor driving unit (211) can drive the compressor (11) by applying a driving signal to the compressor (11) in response to a control command of the control unit (250).

[0104] Accordingly, the driving unit (210) can drive the compressor (11) through the compressor driving unit (211).

[0105] The above fan driving unit (212) can drive a fan motor so that at least one fan (15) rotates in response to a control command of the control unit (250).

[0106] The above valve driving unit (213) controls the opening and closing of at least one valve (13, 16) according to a control command of the control unit (250), and can control the opening amount of the valve as needed.

[0107] The above detection unit (220) can be controlled by the control unit (250).

[0108] The above detection unit (220) is controlled by the control unit (250) and can detect the status information.

[0109] The above detection unit (220) includes one or more sensors and can detect one or more pieces of status information through the one or more sensors.

[0110] For example, one or more of the operating speed, operating frequency, torque, driving current, driving voltage, ambient pressure, suction temperature, discharge temperature, internal temperature, external temperature and control command of the compressor (11), the heat exchange cycle temperature of the air conditioner (100) and the ambient temperature can be detected.

[0111] The above detection unit (220) may include one or more temperature sensors (221) that detect the suction temperature and discharge temperature of the compressor (11), the heat exchange cycle temperature of the air conditioner (100), and the ambient temperature.

[0112] The above detection unit (220) may also further include one or more pressure sensors (222) that detect the pressure of the refrigerant flowing into the compressor (11), the pressure of the refrigerant discharged from the compressor (11), and the pressure of the refrigerant pipe (30).

[0113] The above pressure sensor (222) may be, for example, a suction sensor that detects pressure at the suction portion of the compressor (11) and a discharge sensor that detects pressure at the discharge portion of the compressor (11).

[0114] The detection unit (220) may further include one or more sensors among one or more rotation speed sensors (223) that detect the rotation speed of the motor that rotates the compressor (11) and the fan (15), one or more voltage sensors (224) that detect the voltage of the input power applied to the outdoor unit (10) or the indoor unit (20), the voltage applied to the compressor (11) and the fan (15), and one or more current sensors (225) that detect the current of the input power applied to the outdoor unit (10) or the indoor unit (20), and the current applied to the compressor (11) and the fan (15).

[0115] The above detection unit (220) can generate a signal for the result of detecting one or more of the status information through the one or more sensors and transmit the signal to the control unit (250).

[0116] The above estimation unit (230) can be controlled by the control unit (250).

[0117] The above estimation unit (230) is controlled by the control unit (250), and can estimate the output of the artificial neural network (NN) as a normal probability value for the probability of a normal state of one or more points among the pipes of the air conditioner (100) or a probability value for the probability of a blockage occurring, by using the detection result of the detection unit (220) as an input of the artificial neural network (NN).

[0118] That is, the estimation unit (230) can estimate the normal probability value or the probability value of blockage occurrence at one or more points of the pipe (30) through the artificial neural network (NN).

[0119] The above artificial neural network (NN) is an artificial neural network (ANN) learned through machine learning or deep learning, and may refer to an artificial intelligence implementation circuit or an artificial intelligence implementation algorithm.

[0120] That is, the artificial neural network (NN) can be implemented in software form or in hardware form such as a chip.

[0121] Here, the above machine learning can mean that the computer learns through data without a person directly instructing the computer on logic, and through this, the computer solves problems on its own.

[0122] The above Deep Learning is an artificial intelligence technology that allows computers to learn on their own like humans without being taught by humans, by teaching computers the way humans think based on artificial neural networks (NN) for constructing artificial intelligence.

[0123] In this way, the artificial neural network (NN) may be configured as an artificial intelligence implementation circuit or an artificial intelligence implementation algorithm learned based on the machine learning or deep learning, and may be configured as an operation device of the estimation unit (230), or as an operation algorithm or operation program of the operation device.

[0124] The above artificial neural network (NN) can be implemented as any one of a Feed-Forward Neural Network (FNN) as shown in FIG. 5a, a Convolution Pooling Neural Network (CNN) as shown in FIG. 5b, and a Recurrent Neural Network (RNN) as shown in FIG. 5c.

[0125] The above artificial neural network (NN) can be trained to calculate specific data based on input data.

[0126] The above artificial neural network (NN) can be trained to calculate the probability value of one or more points from one or more of the operating speed, operating frequency, torque, operating current, operating voltage, ambient pressure, suction temperature, discharge temperature, internal temperature, external temperature and control command of the compressor (11), the heat exchange cycle temperature of the air conditioner (100) and the ambient temperature.

[0127] Accordingly, the artificial neural network (NN) may calculate and output a probability value of one or more points by inputting one or more of the operating speed, operating frequency, torque, operating current, operating voltage, ambient pressure, suction temperature, discharge temperature, internal temperature, external temperature, and control command of the compressor (11) included in the detection result, the heat exchange cycle temperature of the air conditioner (100), and the ambient temperature.

[0128] The above estimation unit (230) inputs the detection results for one or more of the operating speed, operating frequency, torque, operating current, operating voltage, ambient pressure, suction temperature, discharge temperature, internal temperature, external temperature and control command of the compressor (11), the heat exchange cycle temperature of the air conditioner (100) and the ambient temperature into the artificial neural network (NN), and when the artificial neural network (NN) calculates and outputs the probability values ​​of the one or more points calculated based on the detection results, the output result can be estimated as the probability values ​​of the one or more points.

[0129] The above estimation unit (230) can estimate the probability value of a point corresponding to at least between a sensor provided around the compressor (11) among the pipes (30) and the compressor (11).

[0130] That is, the artificial neural network (NN) can be trained to calculate and output a probability value of a point between a sensor provided around the compressor (11) in the pipe (30) and the compressor (11), by inputting one or more of the operating speed, operating frequency, torque, operating current, operating voltage, ambient pressure, suction temperature, discharge temperature, internal temperature, external temperature, and control command of the compressor (11), the heat exchange cycle temperature of the air conditioner (100), and the ambient temperature.

[0131] Accordingly, the artificial neural network (NN) calculates the probability value of a point between a sensor provided around the compressor (11) in the pipe (30) and the compressor (11) based on one or more of the operating speed, operating frequency, torque, operating current, operating voltage, ambient pressure, suction temperature, discharge temperature, internal temperature, external temperature and control command of the compressor (11), the heat exchange cycle temperature and the ambient temperature of the air conditioner (100), and the estimation unit (230) can estimate the probability value of a point between a sensor provided around the compressor (11) in the pipe (30) and the compressor (11).

[0132] In this way, by estimating the probability value of the point corresponding to the sensor provided around the compressor (11) in the pipe (30) and the compressor (11) through the artificial neural network (NN), the estimation unit (230) can replace the sensor that senses the state of the point corresponding to the sensor provided around the compressor (11) in the pipe (30) and the compressor (11).

[0133] An example in which the estimation unit (230) is implemented as an artificial neural network (NN) may be as shown in FIG. 6.

[0134] As illustrated in FIG. 6, when one or more state information included in the detection result is input to the artificial neural network (NN) of the estimation unit (230) through the INPUT LAYER, preprocessing of the input variables is performed, and when the preprocessing result is input to the LEARNING LAYER, the probability value of each of the one or more points is calculated, and the calculation result can be output through the OUTPUT LAYER.

[0135] The principle by which the above estimation unit (230) estimates the probability value of one or more points is explained with reference to FIGS. 7 (a to d), 8 (a to d), and 9 (a to d).

[0136] Fig. 7 shows a waveform graph according to status information when the air conditioner (100) is operating normally, and Figs. 8 and 9 show waveform graphs according to status information when a blockage occurs at the first point and the second point, respectively, of the pipe.

[0137] The above artificial neural network (NN) can learn state history information as shown in FIGS. 7 to 9, and estimate the probability value of one or more points by comparing the state information with previously learned information.

[0138] For example, if the waveform of the above status information is as shown in FIG. 7, it is determined that no blockage has occurred at one or more points of the pipe, and thus the abnormal probability value may be estimated to be low and the normal probability value may be estimated to be high.

[0139] In addition, when the waveform of the above status information is as shown in FIGS. 8 and 9, it is determined that a blockage has occurred at a specific point in the pipe, and thus the abnormal probability value can be estimated as low and the normal probability value can be estimated as high.

[0140] Meanwhile, the artificial neural network (NN) learns the fluctuation waveforms (X1 and X2) due to blockage occurrence from the status history information as shown in FIGS. 8 and 9, and if the waveform of the status information appears similar to the fluctuation waveforms (X1 and X2), it determines that a blockage has occurred at the corresponding point, and can estimate the probability value of each point.

[0141] That is, the estimation unit (230) can estimate the probability value of one or more points according to the most similar waveform by comparing the waveform of the state information with the waveform shown in FIGS. 7 to 9 by having the artificial neural network (NN) learn the state history information as shown in FIGS. 7 to 9.

[0142] The above estimation unit (230) can generate a signal for the result of estimating the probability value of one or more points and transmit the signal to the judgment unit (240).

[0143] The above estimation unit (230) can also estimate one or more of the undetected information that the detection unit (220) does not detect during the driving state.

[0144] That is, the estimation unit (230) may further estimate information among the status information of the driving state that is not detected by the detection unit (220) due to the lack of the sensor.

[0145] Here, the undetected information may be one or more of the pressure and temperature of one or more points of the compressor (11).

[0146] For example, it may be one of the internal temperature, external temperature, suction pressure, and discharge pressure of the compressor (11).

[0147] Accordingly, the estimation unit (230) can further estimate one or more undetected information among the internal temperature, external temperature, suction pressure, and discharge pressure of the compressor (11).

[0148] In this case, the estimation unit (230) can estimate the output of the artificial neural network (NN) as one or more of the internal temperature, external temperature, suction pressure, and discharge pressure of the compressor (11) by using the detection result of the detection unit (220) as the input of the artificial neural network (NN).

[0149] That is, the estimation unit (230) can estimate one or more of the pressure and temperature of one or more points of the compressor (11) through the artificial neural network (NN).

[0150] The above estimation unit (230) inputs the detection results for one or more of the operating speed, operating frequency, torque, driving current, driving voltage, suction temperature, discharge temperature, ambient pressure, internal temperature, external temperature, control command, heat exchange cycle temperature of the air conditioner (100), and ambient temperature of the compressor (11) into the artificial neural network (NN), and when the artificial neural network (NN) calculates and outputs the pressure of one or more points of the compressor (11) based on the detection results, the output result can be estimated as the pressure of one or more points of the compressor (11).

[0151] In this case, the artificial neural network (NN) can be trained to calculate the pressure of one or more points of the compressor (11) from one or more of the operating speed, operating frequency, torque, driving current, driving voltage, suction temperature, discharge temperature, ambient pressure, internal temperature, external temperature, control command, heat exchange cycle temperature of the air conditioner (100), and ambient temperature of the compressor (11).

[0152] Accordingly, the artificial neural network (NN) may calculate and output the pressure of one or more points of the compressor (11) by inputting one or more of the operating speed, operating frequency, torque, operating current, operating voltage, suction temperature, discharge temperature, ambient pressure, internal temperature, external temperature, control command, heat exchange cycle temperature of the air conditioner (100), and ambient temperature of the compressor (11) included in the detection result.

[0153] The above estimation unit (230) can estimate the suction pressure of the compressor (11).

[0154] That is, the artificial neural network (NN) can be trained to calculate and output the suction pressure of the compressor (11) by inputting one or more of the operating speed, operating frequency, torque, driving current, driving voltage, suction temperature, discharge temperature, ambient pressure, internal temperature, external temperature, control command, heat exchange cycle temperature of the air conditioner (100), and ambient temperature of the compressor (11).

[0155] Accordingly, the artificial neural network (NN) calculates the suction pressure of the compressor (11) based on one or more of the operating speed, operating frequency, torque, operating current, operating voltage, suction temperature, discharge temperature, ambient pressure, internal temperature, external temperature, control command, heat exchange cycle temperature of the air conditioner (100), and ambient temperature of the compressor (11), so that the estimation unit (230) can estimate the suction pressure of the compressor (11).

[0156] By estimating the suction pressure through the artificial neural network (NN) in this way, the estimation unit (230) can replace the sensor that senses the suction pressure.

[0157] Accordingly, the above estimation unit (230) may perform a virtual sensor function that replaces the pressure sensor.

[0158] The above estimation unit (230) can also further estimate the discharge pressure of the compressor (11).

[0159] That is, the estimation unit (230) can estimate the suction pressure and further estimate the discharge pressure.

[0160] Accordingly, the estimation unit (230) may replace one or more of the suction pressure sensor and the discharge pressure sensor of the compressor (11).

[0161] The above estimation unit (230) also inputs the detection results for one or more of the operating speed, operating frequency, torque, driving current, driving voltage, suction pressure, suction temperature, discharge pressure, discharge temperature, control command, heat exchange cycle temperature of the air conditioner (100), and ambient temperature of the compressor (11) into the artificial neural network (NN), and when the artificial neural network (NN) calculates and outputs the temperature of one or more points of the compressor (11) based on the detection results, the output result can be estimated as the temperature of one or more points of the compressor (11).

[0162] In this case, the artificial neural network (NN) can be trained to calculate the pressure of one or more points of the compressor (11) from one or more of the operating speed, operating frequency, torque, driving current, driving voltage, suction pressure, suction temperature, discharge pressure, discharge temperature, control command, heat exchange cycle temperature of the air conditioner (100), and ambient temperature of the compressor (11).

[0163] Accordingly, the artificial neural network (NN) may calculate and output the pressure of one or more points of the compressor (11) by inputting one or more of the operating speed, operating frequency, torque, driving current, driving voltage, suction pressure, suction temperature, discharge pressure, discharge temperature, control command, heat exchange cycle temperature of the air conditioner (100), and ambient temperature of the compressor (11) included in the detection result.

[0164] The above estimation unit (230) can estimate, for example, one or more of the internal temperature and external temperature of the compressor (11).

[0165] That is, the artificial neural network (NN) can be trained to calculate and output at least one of the internal temperature and external temperature of the compressor (11) by inputting at least one of the operating speed, operating frequency, torque, driving current, driving voltage, suction pressure, suction temperature, discharge pressure, discharge temperature, control command, heat exchange cycle temperature of the air conditioner (100), and ambient temperature of the compressor (11).

[0166] Accordingly, the artificial neural network (NN) calculates at least one of the internal temperature and the external temperature of the compressor (11) based on at least one of the operating speed, operating frequency, torque, driving current, driving voltage, suction pressure, suction temperature, discharge pressure, discharge temperature, control command, heat exchange cycle temperature of the air conditioner (100), and ambient temperature of the compressor (11), so that the estimation unit (230) can estimate at least one of the internal temperature and the external temperature of the compressor (11).

[0167] In this way, by estimating at least one of the internal temperature and the external temperature through the artificial neural network (NN), the estimation unit (230) may replace a sensor that senses the internal temperature and the external temperature.

[0168] Accordingly, the above estimation unit (230) may perform a virtual sensor function that replaces the temperature sensor.

[0169] The above estimation unit (230) can also estimate one or more of the suction temperature and discharge temperature of the compressor (11).

[0170] That is, the estimation unit (230) can estimate at least one of the internal temperature and the external temperature, and further estimate at least one of the suction temperature and the discharge temperature.

[0171] Accordingly, the estimation unit (230) may replace one or more of the internal temperature sensor of the compressor (11), the external temperature sensor of the compressor (11), the suction temperature sensor of the compressor (11), and the discharge temperature sensor of the compressor (11).

[0172] In this way, the estimation unit (230) performs a virtual sensor function by estimating at least one of the internal pressure and temperature through the artificial neural network (NN), thereby solving the problem of difficulty in internal sensing due to the structural limitations of the compressor (11) that prevents a sensor from being installed inside the compressor (11).

[0173] The above estimation unit (230) can generate a signal for the result of further estimating the pressure and temperature of one or more points of the compressor (11) and transmit the signal to the judgment unit (240).

[0174] The above judgment unit (240) can be controlled by the control unit (250).

[0175] The above judgment unit (240) is controlled by the control unit (250) and can determine whether a blockage has occurred at one or more points based on the estimation result of the estimation unit (230). In addition, the judgment unit (240) is controlled by the control unit (250) and can determine whether an abnormality has occurred at one or more of the compressor (11) and the detection unit (220) based on the estimation result of the estimation unit (230).

[0176] The above judgment unit (240) may also determine whether a blockage has occurred at one or more of the above points through an artificial neural network (NN).

[0177] In this case, the artificial neural network (NN) of the judgment unit (240) can be trained to determine whether a blockage has occurred at one or more points based on the estimation result.

[0178] The above judgment unit (240) can determine whether a blockage has occurred at each of the one or more points based on the probability value of each of the one or more points.

[0179] The above judgment unit (240) can determine whether a blockage has occurred at each of the one or more points based on the result of comparing the probability value of each of the one or more points with one or more preset reference values.

[0180] That is, the judgment unit (240) may determine whether a point corresponding to the probability value is in a normal / abnormal state and whether a blockage has occurred, based on the result of comparing each of the probability values ​​with each of the one or more reference values.

[0181] Here, the above reference value may be a probability value that the point is in a normal state.

[0182] Alternatively, the reference value may be a probability value of a high temperature phenomenon due to blockage occurrence. Accordingly, the judgment unit (240) may determine whether blockage has occurred by detecting whether the temperature rises at a point corresponding to the probability value through a comparison between the probability value and the reference value.

[0183] The above reference value may be set to multiple reference values, and in this case, the multiple reference values ​​may be different values.

[0184] In this way, when the above reference value is set to multiple values ​​of different values, by comparing the probability value with each of the multiple reference values, the degree or possibility of blockage occurring at the point corresponding to the probability value can be further determined.

[0185] In this way, the judgment unit (240) that determines whether a blockage has occurred at one or more points based on the above estimation results can generate a signal for the judgment result and transmit the signal to the control unit (250).

[0186] The above control unit (250) can control the operation of the outdoor unit (10) and the indoor unit (20) by controlling the driving unit (210).

[0187] The above control unit (250) can control the operation of the air conditioner (100) by controlling the operation of the outdoor unit (10) and the indoor unit (20).

[0188] That is, the control unit (250) can control the operation of the outdoor unit (10) and the indoor unit (20) by controlling the driving unit (210), thereby controlling the air conditioning function of the air conditioner (100) to be performed.

[0189] The above control unit (250) controls the driving unit (10) according to one or more of the detection result of the detection unit (220), the estimation result of the estimation unit (230), and the judgment result of the judgment unit (240).

[0190] That is, the control unit (250) can control the operation of the outdoor unit (10) and the indoor unit (20) by controlling the driving unit (10) based on one or more of the detection result, the estimation result, and the judgment result.

[0191] The control unit (250) may control the driving unit (210) according to the operation performance of the air conditioner (100), or may control the operation performance of the air conditioner (100) according to one or more of the detection result, the estimation result, the judgment result, and the control result of the driving unit (210).

[0192] The control unit (250) controls the driving unit (210), the detection unit (220), the estimation unit (230), and the judgment unit (240), and can control the driving unit (210) so that the internal temperature of the compressor (11) is maintained below a preset reference temperature according to the judgment result of the judgment unit (240).

[0193] That is, the control unit (250) can control the driving unit (210) based on the result of determining whether or not a blockage has occurred at one or more of the points. Accordingly, the occurrence of blockage and the resulting rise in the temperature of the compressor (11) due to vacuum operation can be prevented.

[0194] The above control unit (250) can control the driving unit (210) according to a preset first control mode when the judgment result corresponds to one or more of the one or more points being either likely to be blocked or having been blocked.

[0195] That is, the control unit (250) may control the driving unit (210) according to the first control mode when there is at least one point among the one or more points that corresponds to either a possibility of blockage or a blockage occurrence.

[0196] The above first control mode may be a mode that controls one or more of the compressor (11), the fan (15) and the valve (13, 16) with a driving pattern to re-detect whether a blockage has occurred at one or more points.

[0197] That is, if, as a result of the determination, there is at least one point among the one or more points that corresponds to either a possibility of blockage or a blockage occurring, the control unit (250) may control the driving unit (210) according to the first control mode to re-detect whether a blockage has occurred at the one or more points.

[0198] Accordingly, the driving unit (210) may drive one or more of the compressor (11), the fan (15), and the valve (13, 16) according to the first control mode.

[0199] The above control unit (250) can control the driving unit (210) according to the first control mode, and then re-determine whether a blockage has occurred at the corresponding point based on the change in the status information according to the first control mode.

[0200] In this case, the control unit (250) can determine a change in the status information based on the detection result of the detection unit (220) while one or more of the compressor (11), the fan (15), and the valves (13, 16) are driven according to the first control mode, and can re-determine whether a blockage has occurred at the corresponding point based on the determined result.

[0201] In addition, in this case, the estimation unit (230) may estimate the probability value of the corresponding point based on the detection result of the detection unit (220) while driving according to the first control mode, and the judgment unit (240) may determine whether a blockage has occurred at the corresponding point based on the re-estimated probability value.

[0202] When the control unit (250) controls the driving unit (210) according to the first control mode, the control unit (250) can control the driving unit (210) by changing the control command for one or more of the compressor (11), the fan (15), and the valve (13, 16).

[0203] That is, when the control unit (250) controls the driving unit (210) according to the first control mode, the control unit (250) changes the control command to control the driving unit (210), so that one or more of the compressor (11), the fan (15), and the valve (13, 16) can be controlled to change their operating state and operate.

[0204] In this case, the control unit (250) may control the driving unit (210) by changing the control command to drive at least one of the compressor (11), the fan (15) and the valve (13, 16) in a detection pattern to re-detect whether a blockage has occurred at one or more points.

[0205] Accordingly, the operating status of one or more of the compressor (11), the fan (15) and the valve (13, 16) may be changed to a detection pattern according to the control command and driven.

[0206] That is, the control unit (250) may control the driving unit (210) according to the first control mode for the predetermined period of time. Here, the predetermined period of time may be a time for re-determining whether a blockage has occurred at one or more points.

[0207] The control unit (250) may also control the driving unit (210) according to a preset second control mode so that the internal temperature becomes lower than the reference temperature when the judgment result corresponds to the occurrence of blockage at one or more points.

[0208] That is, the control unit (250) may control the driving unit (210) according to the second control mode when there is at least one point corresponding to a blockage among the one or more points.

[0209] The second control mode may be a mode that controls one or more of the compressor (11), the fan (15), and the valve (13, 16) in a driving pattern in which the internal temperature is reduced below the reference temperature.

[0210] That is, if, as a result of the determination, at least one of the one or more points corresponds to a blockage occurrence, the control unit (250) may control the driving unit (210) according to the second control mode so that the internal temperature is reduced below the reference temperature.

[0211] Accordingly, the driving unit (210) may drive one or more of the compressor (11), the fan (15), and the valve (13, 16) according to the second control mode.

[0212] When the control unit (250) controls the driving unit (210) according to the second control mode, the control unit (250) can control the driving unit (210) by changing the control command for one or more of the compressor (11), the fan (15), and the valve (13, 16).

[0213] That is, when the control unit (250) controls the driving unit (210) according to the second control mode, the control unit (250) changes the control command to control the driving unit (210), so that one or more of the compressor (11), the fan (15), and the valve (13, 16) can be controlled to change their operating state and operate.

[0214] In this case, the control unit (250) may control the driving unit (210) by changing the control command so that one or more of the compressor (11), the fan (15), and the valve (13, 16) is driven in an emergency pattern in which the internal temperature is reduced below the reference temperature.

[0215] Accordingly, the operating status of one or more of the compressor (11), the fan (15), and the valve (13, 16) may be changed to an emergency pattern according to the control command and driven.

[0216] When the control unit (250) controls the driving unit (210) according to the second control mode, when changing the control command of the compressor (11), the control command can be changed so that at least one of the operating speed and power consumption of the compressor (11) is reduced to a certain reduction standard and driven.

[0217] Accordingly, the compressor (11) may have at least one of the operating speed and power consumption reduced to a certain reduction standard while operating in accordance with the second control mode.

[0218] When the control unit (250) controls the driving unit (210) according to the second control mode, when changing the control command of the fan (15), the control command can be changed so that at least one of the operating speed and power consumption of the fan (15) is reduced to a certain reduction standard.

[0219] Accordingly, the fan (15) may have at least one of the driving speed and power consumption reduced to a certain reduction standard while being driven according to the second control mode.

[0220] When the control unit (250) controls the driving unit (210) according to the second control mode, if the control command of the valve (13, 16) is changed, the control command can be changed so that the direction of the valve (13, 16) is switched.

[0221] Accordingly, the valve (13, 16) can change the opening and closing direction while driving according to the second control mode.

[0222] The control unit (250) can control the driving unit (210) according to the second control mode, and then generate result information regarding the judgment result and the control result according to the second control mode. The generated information can be stored or transmitted to an external communication target device.

[0223] In this case, the control unit (250) may store the result information in the storage unit (280) or transmit the result information to an external communication target device through the communication unit (290), thereby recording or notifying the communication target device of whether a blockage has occurred and whether vacuum operation has occurred accordingly.

[0224] Accordingly, the user or operator of the air conditioner (100) can be made aware of whether a blockage has occurred in the pipe (30) and whether vacuum operation has occurred as a result.

[0225] In this way, the control unit (250) controls the drive unit (210) in the second control mode according to the judgment result so that the internal temperature is lower than the reference temperature, thereby preventing vacuum operation of the air conditioner (100) and a resulting increase in high temperature.

[0226] In addition, by controlling the internal temperature to be lower than the reference temperature in response to the occurrence of blockage at one or more of the above points, emergency control in response to the occurrence of blockage and the resulting vacuum operation can be appropriately performed.

[0227] On the other hand, if the cycle is diagnosed only by comparing the size of temperature / current information, there is a high possibility of misdiagnosis, and since the diagnosis starts immediately when the reference frequency is exceeded, there is a high possibility of misdiagnosis if the operating frequency temporarily increases before the cycle is stabilized.

[0228] In the case of air conditioners, the amount of refrigerant, the length of the pipe, and the indoor / outdoor environmental temperatures can vary greatly depending on the installation environment. Therefore, simply comparing the instantaneous measured value with the reference value increases the probability of misdiagnosis. Therefore, it is necessary to consider the tendency of each value to change when diagnosing. To perform a model-based diagnosis that considers the amount of change in each variable, various physical phenomena such as fluid, heat, and electricity must be modeled. Therefore, it is necessary to include complex physical equations rather than simply comparing measured values ​​with reference values, and modeling the interactions between each physical phenomenon is necessary.

[0229] On the other hand, using an artificial neural network based on supervised learning creates diagnostic logic that takes into account these complex multi-physics phenomena, enabling relatively accurate diagnosis without complex modeling.

[0230] Meanwhile, according to an embodiment of the present disclosure, cycle blockage can be diagnosed using internal air conditioner data without using separate additional sensors. The diagnostic method utilizes an artificial neural network (NN) that outputs probability values. If the normal probability output by the algorithm falls below a threshold or the failure probability exceeds a threshold, a failure is determined. If the probability remains within the suspicion threshold, a failure detection pattern is applied.

[0231] However, if the result of the artificial neural network (NN) algorithm exceeds a certain value, it may be misdiagnosed as suspected or suspected of failure, or if the result of the algorithm fluctuates within the reference value boundary, frequent changes in the judgment of normal / failure may occur.

[0232] In particular, when the normal probability value output by the artificial neural network (NN) temporarily drops below the reference value due to a user changing the setting conditions of the indoor unit or a temporary phase change in the refrigerant, the product can be stopped and a failure notification can be sent to the customer by misdiagnosing it as a failure.

[0233] If a malfunction occurs in an air conditioning product, rendering it inoperable, customers may experience significant inconvenience and businesses may incur unnecessary service costs until a service technician visits in the middle of a hot summer or cold winter.

[0234] An embodiment of the present disclosure proposes a post-processing algorithm for reducing misdiagnosis in an artificial neural network diagnosis algorithm that outputs probability values.

[0235] According to one embodiment of the present disclosure, when the normal probability value output by the artificial neural network (NN) is lower than or equal to a reference value and the operating speed of the compressor (11) is continuously operated at or above the reference speed for a certain period of time, the possibility of misdiagnosis can be reduced by determining that it is abnormal.

[0236] The estimation unit (230) uses the detection result of the detection unit (220) as input to a pre-trained artificial neural network (NN) and estimates the state of the air conditioner (100) based on the output of the artificial neural network (NN).

[0237] For example, the estimation unit (230) estimates a probability value for the state of one or more points in the piping of the air conditioner (100) based on the output of the artificial neural network (NN).

[0238] Meanwhile, the judgment unit (240) determines whether an abnormality has occurred based on the probability value estimated by the estimation unit (230). At this time, the judgment unit (240) determines that an abnormality has occurred at the corresponding point if the probability value satisfies the abnormality occurrence criterion and is maintained for a certain period of time or longer.

[0239] The above-described predetermined time period may be set to be greater than the time required to detect multiple abnormal conditions. The air conditioner (100) comprises multiple components and has various failure modes. Therefore, the determination unit (240) checks multiple items and detects multiple abnormal conditions to confirm the normal state of the cycle and the air conditioner (100). For example, the determination unit (240) may detect abnormal conditions such as blockage of pipes at one or more points, abnormality of the compressor (11), or refrigerant leakage.

[0240] The required detection time required to determine each abnormal condition may vary. In this case, the determination unit (240) may continue the abnormal condition determination logic to determine whether any remaining critical abnormal conditions exist, even if the time required to determine a specific abnormal condition has elapsed. In some cases, the determination unit (240) may set the predetermined time based on the longest required detection time among the times required to detect multiple abnormal conditions.

[0241] In addition, the above-mentioned predetermined time may be set to be less than the high-temperature burnout time of the compressor (11). Even if other abnormalities are checked, the compressor (11) must be prevented from being damaged. Therefore, the maximum value of the above-mentioned predetermined time may be set to be less than the high-temperature burnout time of the compressor (11).

[0242] Meanwhile, the above probability value may be a normal probability value indicating the probability that one or more points are in a normal state.

[0243] Alternatively, the probability value may be an abnormal probability value indicating the probability that the one or more points are in an abnormal state.

[0244] According to one embodiment of the present disclosure, when using the above normal probability value, the probability threshold included in the above abnormality occurrence criterion may be proportional to the above predetermined time. That is, the longer the detection time required to determine the abnormality, the higher the probability threshold may be set.

[0245] Meanwhile, the estimation unit (230) can estimate the normal probability value when the operating frequency of the compressor (11) becomes higher than the minimum frequency. Since the compressor (11) is operated at low speed immediately after startup, there is almost no cycle circulation, making it difficult to determine a cycle failure, diagnosis begins when the operating speed of the compressor (11) becomes higher than the reference speed.

[0246] In addition, the estimation unit (230) can estimate the probability value according to a set cycle. At this time, the cycle can correspond to the operation speed of the artificial neural network (NN).

[0247] Meanwhile, if the control unit (250) determines that an abnormality has occurred, it can cause an emergency stop of the compressor (11).

[0248] Meanwhile, the estimation unit (230) may further estimate one or more pieces of undetected information that the detection unit (220) does not detect during the driving state. The estimation unit (230) may perform a virtual sensor function that replaces a predetermined sensor. For example, the estimation unit (230) may perform a virtual sensor function that replaces a suction pressure sensor.

[0249] In order to reduce material costs, there are cases where the compressor (11) is controlled only with a discharge pressure sensor without using a suction pressure sensor. In such cases, since the suction pressure cannot be accurately measured, the product may not detect vacuum operation due to pipe blockage.

[0250] If the compressor (11) is operated continuously at high speed in a vacuum state, there is no flow of refrigerant inside the compressor, so the motor, compression chamber, and bearing parts are not cooled, which may cause various high-temperature failures such as melting of the motor insulation, bearing fusion due to high temperature, and poor lubrication due to a decrease in oil viscosity.

[0251] In general, there is a temperature sensor at the discharge end of the compressor (11), and when the flow of refrigerant is smooth, the heat inside the compressor (11) is transferred to the discharge end, so that a high-temperature failure can be prevented by the temperature sensor at the discharge end. However, when vacuum operation occurs, the refrigerant flowing out from the inside of the compressor (11) to the outside is thin, so heat transfer does not occur, making it difficult to detect a high-temperature failure inside the compressor (11) by the temperature sensor at the discharge end.

[0252] In addition, even if there is a suction pressure sensor, if a blockage occurs between the pressure sensor and the suction portion of the compressor (11), the compressor blockage cannot be detected using information from the pressure sensor.

[0253] To solve these problems, an embodiment of the present disclosure is an abnormality classification technology based on an artificial neural network (NN) algorithm, which can estimate an abnormal state even if the operating environment changes from air conditioner operation data.

[0254] For example, by reproducing high-temperature damage caused by vacuum operation due to pipe blockage in various environments, time-series data is collected and an artificial neural network (NN) trained on this data is constructed to detect vacuum operation due to pipe blockage in real time in various usage environments. Real-time detection includes detection on the product's MCU (edge) or in a cloud environment.

[0255] Referring to Fig. 10, data such as refrigerant amount, pipe length, indoor / outdoor temperature, and initial temperature can be collected under various environmental conditions and according to various blockage locations, and an artificial neural network (NN) learning algorithm can be configured based on the collected data.

[0256] Determining whether a blockage exists and the location of the blockage can be accomplished using a classification method. The artificial neural network (NN) learning algorithm can be configured in the FNN / CNN / RNN illustrated in Figure 5, as well as other forms.

[0257] Meanwhile, the input data of the artificial neural network (NN) is classified according to the data collection location, such as within the compressor (11), within the outdoor unit (10), and all cycle information, and the performance may vary depending on the data location and number.

[0258] The presence and location of blockages can be predicted based on the changes and interrelationships of multiple variables over time. Each collected data point contains some information about the presence and location of blockages. Because waveforms vary depending on the environment, environmental information must be inferred from multiple input variables. By inferring and eliminating environmental information, blockage information can be determined even in diverse environments. Insufficient input variables or sampling can make determination impossible.

[0259] When vacuum operation occurs due to pipe blockage, changes occur in various data, as shown in the waveform in Figure 11. Information exists in various variables, such as target high pressure, current high pressure, pipe in / out temperature, discharge temperature, suction temperature, and power consumption.

[0260] Determining the type of fault and the location of the blockage requires comparing waveforms for normal, other faults, and blockages. Furthermore, in addition to determining abnormal operation, information regarding the type of fault and the location of the blockage must be determined based on various variables.

[0261] Artificial neural network (NN) algorithms can use probabilistic neural network algorithms that output normal probabilities. However, if vacuum operation is determined solely based on probability values, misdiagnosis may occur, or if probability values ​​fluctuate outside the reference range, the normal / failure judgment may be reversed.

[0262] Therefore, in order to reduce misdiagnosis, when the operating speed of the compressor (11) is higher than the reference speed and the normal probability value is lower than the reference value and the compressor (11) is operated continuously for a certain period of time, it is judged as abnormal, thereby reducing the possibility of misdiagnosis.

[0263] The estimation unit (230) can estimate a normal probability value for the state of one or more points in a pipe based on an artificial neural network (NN). The estimation unit (230) can estimate the normal probability value according to a set cycle.

[0264] The judgment unit (240) can determine whether a blockage has occurred at one or more of the points based on the normal probability value. If the normal probability value is below a probability reference value, the judgment unit (240) increases an abnormality count, and if the abnormality count is above the count reference value, it determines that a blockage has occurred at the point.

[0265] The above count reference value may be set so that the required detection time, which is the product of the count reference value and the cycle, is greater than the detection time of a failure other than the blockage failure, and may be set so as to be less than the high temperature burnout time of the compressor (11).

[0266] The above abnormal count can be set based on the above probability criterion.

[0267] For example, when the probability criterion is set to a first value, the abnormal count may be set to a second value, and when the probability criterion is set to a third value that is less than the first value, the abnormal count may be set to a fourth value that is less than the second value.

[0268] That is, if the probability criterion is small, the abnormal count can also be set small, and if the probability criterion is large, the abnormal count can also be set large.

[0269] According to the present disclosure, the accuracy of a diagnostic algorithm for an air conditioning cycle utilizing an artificial neural network (NN) can be improved. The artificial neural network (NN) outputs the probability of a state, such as normal or faulty, as a probability value, and the judgment unit (240) performs an appropriate post-processing algorithm to accurately diagnose the cycle based on the output probability.

[0270] The judgment unit (240) selects a judgment criterion probability value (reference value) based on the performance of an artificial neural network (NN), and diagnoses a failure when the normal probability remains below the reference value for a certain period of time. This is to prevent misdiagnosis due to fluctuations in the normal probability value caused by temporary phenomena.

[0271] FIG. 11 is a block diagram showing a process of determining whether a blockage has occurred and performing control in a control device of an air conditioner according to one embodiment of the present disclosure.

[0272] Referring to FIG. 11, the control device (200) can detect (B1) the status information by the detection unit (220) when the driving unit (210) drives at least one of the compressor (11), the fan (15), and the valve (13, 16).

[0273] When the above detection unit (220) detects the state information (B1) and transmits the detection result to the estimation unit (230), the estimation unit (230) can preprocess the detection result into an input variable (B2) and then input it into the artificial neural network (NN) (B3).

[0274] In addition, the estimation unit (230) can estimate (B4) the output of the artificial neural network (NN) as a normal probability value for one or more points.

[0275] When the above estimation unit (B4) estimates (B4) the probability value of each of the one or more points and transmits the estimation result to the judgment unit (240), the judgment unit (240) does not immediately determine whether it is normal or abnormal, but performs a post-processing algorithm (B5).

[0276] In the post-processing algorithm, the judgment unit (240) can determine whether a state in which the probability value is judged to be abnormal is maintained for a certain period of time.

[0277] Alternatively, in the post-processing algorithm, the judgment unit (240) may compare each of the probability values ​​with one or more of the reference values ​​to increase or maintain the abnormality count by one.

[0278] The above judgment unit (240) can determine that an abnormality, such as a pipe blockage, has occurred if a state in which an abnormality is determined to have occurred is maintained for a certain period of time or if the abnormality count is greater than a set count reference value.

[0279] The above judgment unit (240) can transmit the judgment result to the control unit (250) (B6). For example, if a pipe blockage occurs and high-temperature damage is predicted, the judgment unit (240) can transmit the judgment result to the control unit (250) (B6).

[0280] The control unit (250) may perform high-temperature damage prevention control (B8) based on the judgment result. For example, the control unit (250) may perform control (B7) so that the internal temperature is maintained below the reference temperature based on the judgment result.

[0281] In the above judgment, the control unit (250) can forcibly stop the compressor (11) and inform the customer of relevant information. In addition, the control unit (250) can minimize customer inconvenience by operating the compressor (11) at a speed that does not cause high temperature damage through compressor (11) limiting operation when necessary, thereby enabling minimal cooling / heating operation.

[0282] In some embodiments, when the product is judged to be abnormal, the diagnostic results can be transmitted to a separate service server without stopping the product, thereby providing relevant information to the service engineer so that the service engineer can perform a proactive inspection.

[0283] According to an embodiment of the present disclosure, when a failure occurs due to cycle blockage, the cause of the failure is accurately diagnosed and provided to a service technician, thereby reducing the number of revisits by the service technician.

[0284] Meanwhile, during vacuum operation, the compressor's internal temperature can be high, which can lead to malfunctions such as melting of motor insulation, magnet demagnetization, and damage to bearings and thrust components due to reduced oil viscosity. According to embodiments of the present disclosure, product life and reliability can be improved by avoiding failure of additional components due to vacuum operation.

[0285] In addition, according to an embodiment of the present disclosure, in an air conditioner (100) using only one pressure sensor, it is possible to accurately diagnose blockage of the cycle, etc., without an additional sensor.

[0286] FIG. 12 is a flowchart illustrating a method for controlling an air conditioner according to an embodiment of the present disclosure.

[0287] First, when the compressor operating frequency exceeds the minimum frequency (A) (S1210), the normal probability calculation using an artificial neural network (NN) algorithm begins (S1220). The minimum frequency (A) is the minimum frequency at which the cycle stabilizes and may vary depending on the specifications of the air conditioner (200) and compressor (11). Generally, the minimum frequency (A) can be selected based on the dwell frequency when the product starts.

[0288] If the minimum frequency (A) is too low, the cycle may not be stabilized, which may lead to a misdiagnosis of an abnormality during normal operation. If the minimum frequency (A) is too high, the probability cannot be calculated during abnormal operation, which may lead to a failure to diagnose an abnormality.

[0289] When the air conditioner (100) starts operating and the compressor operating frequency is low, the diagnosis accuracy is low because the cycle has not yet stabilized. Therefore, when the compressor operating frequency is higher than the minimum frequency (A) (S1210), the estimation unit (230) estimates the normal probability using an artificial neural network (NN) algorithm, thereby improving the diagnosis accuracy. In addition, when the frequency is lower than the minimum frequency (A), the diagnosis algorithm is not operated, thereby reducing the amount of computation.

[0290] The judgment unit (240) can increase the abnormal count (count) if the normal probability value is lower than the probability reference value (B) (S1230). The judgment unit (240) can initialize the abnormal count (count) if the normal probability value is higher than the probability reference value (B) (S1230) (S1245). Alternatively, the judgment unit (240) can maintain the abnormal count (count) if the normal probability value is higher than the probability reference value (B) (S1230) (S1245).

[0291] Additionally, the judgment unit (240) can initialize the abnormal count (count) (S1250) even when the compressor operating frequency is less than the minimum frequency (A) (S1210). Alternatively, the judgment unit (240) can maintain the abnormal count (count) (S1250) even when the compressor operating frequency is less than the minimum frequency (A) (S1210).

[0292] Meanwhile, the probability threshold (B) can be selected based on factors such as the performance of the probabilistic neural network and the evaluation results of existing data. A larger probability threshold (B) increases the likelihood of normal cases being diagnosed as abnormal, while a smaller probability threshold (B) increases the likelihood of abnormal cases being diagnosed as normal.

[0293] Meanwhile, if the above abnormal count is greater than the count reference value (C) (S1260), the judgment unit (240) can determine that a blockage, etc. has occurred at the point and that there is an abnormality (S1270).

[0294] The count threshold (C) can be selected based on the performance (diagnosis speed) and computation cycle of the artificial neural network (NN). A larger count threshold (C) can slow down the diagnosis of abnormalities, while a smaller one can increase the likelihood of normal conditions being diagnosed as abnormal.

[0295] The above-described diagnostic process can be repeated at a predetermined interval after the compressor operating frequency reaches the minimum frequency (A) (S1210), and the process from probability calculation (S1220) to abnormality determination (S1270) can be repeated.

[0296] Figure 13 is a drawing showing the results of an experiment in which an abnormal diagnosis post-processing algorithm was applied, and shows the results of confirming the operation of the algorithm when a blockage occurs during normal operation.

[0297] Referring to Fig. 13, the diagnosis starts when the compressor speed is higher than the minimum frequency (A), and the result illustrated in Fig. 13 is the result when the compressor speed at 0s is higher than the minimum frequency (A).

[0298] Additionally, in the example of Figure 13, a normal probability (B) of 20% or less and an abnormal count (C) of 5 or more were considered abnormal. The execution cycle of the diagnostic algorithm was set to 6 seconds, so that a final blockage was determined when the normal probability was 20% or less for 30 seconds.

[0299] Referring to Figure 13, it can be seen that when a cycle blockage occurs, the normal probability quickly drops to close to 0%.

[0300] Since there is a probability fluctuation of up to 10%p in a blocked state, the standard probability (B) was selected as 20%.

[0301] In addition, mechanical phenomena such as refrigerant flow change at a relatively slow rate of several seconds, unlike electrical phenomena that change rapidly in units of several milliseconds. Therefore, to prevent false detection due to temporary phenomena, the abnormality count (C) was set to 5, and if the same phenomenon continues for 30 seconds, it is determined to be a final blockage.

[0302] Typically, algorithms that detect component failures or other anomalies within a cycle make judgments after detecting phenomena for several seconds. Therefore, if a similar driving pattern to a cycle blockage occurs due to a failure of another component, if the Abnormal Count (C) value is too small, it may be diagnosed as a cycle blockage before other failures are detected, resulting in a misdiagnosis. Therefore, the minimum Abnormal Count (C) value should be selected while also considering the detection time of other failures.

[0303] On the other hand, if the abnormal Count(C) value is too large, the vacuum operation time increases, causing the temperature inside the compressor to rise, which may cause the compressor to fail due to a high temperature before being diagnosed as vacuum, and thus the maximum value should be selected by considering the time it takes for the compressor to be damaged by high temperature during vacuum operation.

[0304] Additionally, the ideal Count(C) value needs to be set considering the computation cycle. Specifically, the faster the computation time of the artificial neural network (NN), the higher the ideal Count(C) value can be.

[0305] Figures 14 and 15 illustrate blockage judgment when the output characteristics of an artificial neural network (NN) algorithm are different.

[0306] Compared to the example of Fig. 13, in cases where the response of the normal probability is relatively slow but the fluctuation is small, as shown in Fig. 14, the reference probability (B11)(B) can be set to a relatively high value to enable a relatively quick fault determination. At this time, the product of the abnormality count (C11)(C) and the artificial neural network (NN) is the time to determine an abnormal state such as a blockage.

[0307] Meanwhile, when the reference probability (B12)(B) is set low in the same output characteristics as in Fig. 15, the abnormal Count (C12)(C) value is set small, and as a result, the time for determining blockage is set to the same value as in Fig. 11, thereby preventing high-temperature failure of the compressor (11).

[0308] In this way, the standard probability (B) and abnormal count (C) values ​​need to be set by considering the output characteristics of the probability neural network and the conditions for high-temperature failure of the compressor.

[0309] Figure 16 is a diagram illustrating three cases in which the compressor speed reaches the reference speed (V1).

[0310] The diagnostic algorithm according to the present disclosure operates after the compressor speed reaches the reference speed (V1). Furthermore, rather than immediately generating an error based on the output of an artificial neural network (NN) as a diagnostic result, an error may occur a certain amount of time after an abnormality occurs, depending on whether the abnormality is maintained for a certain period of time and the count is determined.

[0311] Figure 16 (a) illustrates a case where a pipe blockage occurs after a predetermined time has elapsed after reaching the reference speed (V1). Meanwhile, according to the post-processing algorithm of the present disclosure, a pipe blockage occurs and a blockage error occurs after a predetermined time.

[0312] Figure 16 (b) illustrates a case where a pipe blockage occurs immediately after reaching the reference speed (V1) or where a pipe blockage has already occurred before reaching the reference speed (V1). Even in this case, the diagnostic algorithm is executed after reaching the reference speed (V1), and thus generates a blockage error after a certain period of time after reaching the reference speed (V1).

[0313] Figures 16 (b) and (C) illustrate cases where a pipe blockage occurs immediately after reaching the reference speed (V1) or where a pipe blockage has already occurred before reaching the reference speed (V1). In this case, the diagnostic algorithm is executed after reaching the reference speed (V1), and therefore generates a blockage error after a certain period of time after reaching the reference speed (V1).

[0314] The control device and control method of the air conditioner as described above can be applied and implemented to all control devices of air conditioners, air conditioning systems, control devices of air conditioning systems, control methods of air conditioning systems, etc. to which the technical idea of ​​the above technology can be applied.

[0315] Although the 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 various modifications can be made by a person having ordinary skill in the art to which the present disclosure pertains without departing from the gist of the present invention as claimed in the claims. Furthermore, such modifications should not be understood individually from the technical idea or prospect of the present invention.

Claims

1. In a control device of an air conditioner that controls an air conditioner, A driving unit that drives the compressor, fan and valve of the above air conditioner; A detection unit including one or more sensors to detect one or more status information on the operating status of the compressor and the operating status of the air conditioner; An estimation unit that estimates a normal probability value for the state of one or more points in the pipes of the air conditioner based on the output of the artificial neural network by using the detection result of the above detection unit as an input to a pre-learned artificial neural network; and A judgment unit for determining whether a blockage occurs at one or more of the above points based on the above normal probability value; The above estimation part is, Estimate the above normal probability value according to the set cycle, The above judgment committee, An air conditioner control device that increases an abnormal count when the above normal probability value is lower than or equal to a probability criterion, and determines that a blockage has occurred at the corresponding point when the above abnormal count is higher than or equal to the count criterion.

2. In paragraph 1, The above estimation part is, An air conditioner control device that estimates the normal probability value when the operating frequency of the compressor becomes higher than the minimum frequency.

3. In paragraph 1, The above count criteria are, The product of the above count reference value and the above cycle, A control device of an air conditioner, wherein the detection time is set to be greater than that of a failure other than the above blockage occurrence failure and is set to be less than the high temperature burnout time of the compressor.

4. In paragraph 1, The above abnormal count is a control device of an air conditioner set based on the above probability criterion.

5. In paragraph 4, When the above probability criterion is set to the first value, the above abnormal count is set to the second value, An air conditioner control device, wherein when the above probability criterion is set to a third value less than the first value, the above abnormal count is set to a fourth value less than the second value.

6. In paragraph 1, The above cycle is, A control device for an air conditioner corresponding to the computational speed of the above artificial neural network.

7. In paragraph 1, A control device for an air conditioner, further comprising a control unit that stops the compressor in an emergency if the above abnormal count is greater than a count reference value.

8. In paragraph 1, The above driving status is, A control device of an air conditioner including at least one of the operating speed, operating frequency, torque, operating current, operating voltage, ambient pressure, suction temperature, discharge temperature, internal temperature, external temperature and control command of the compressor.

9. In paragraph 1, The above estimation part is, An air conditioner control device that estimates a probability value of a point between a sensor provided around the compressor and the compressor among one or more of the above points.

10. In paragraph 1, The above estimation part is, An air conditioner control device that further estimates one or more pieces of undetected information that are not detected by the detection unit during the above driving state.

11. In a control device of an air conditioner that controls an air conditioner, A driving unit that drives the compressor, fan and valve of the above air conditioner; A detection unit including one or more sensors to detect one or more status information on the operating status of the compressor and the operating status of the air conditioner; An estimation unit that estimates a probability value for the state of one or more points in the pipes of the air conditioner based on the output of the artificial neural network by using the detection result of the above detection unit as an input to a pre-learned artificial neural network; and Includes a judgment unit that determines whether an abnormality has occurred based on a probability value; The above judgment unit is a control device of an air conditioner that determines that an abnormality has occurred at a given point if a state in which the above probability value satisfies the abnormality occurrence criterion is maintained for a certain period of time.

12. In paragraph 11, The above estimation part is, An air conditioner control device that estimates the probability value when the operating frequency of the compressor becomes higher than the minimum frequency.

13. In paragraph 11, The above schedule time is, A control device of an air conditioner, wherein the control device is set to be greater than the time required to detect multiple abnormal conditions and less than the high temperature burnout time of the compressor.

14. In paragraph 11, The above probability values are, A control device for an air conditioner, wherein the normal probability value represents the probability that the one or more points are in a normal state, or the abnormal probability value represents the probability that the one or more points are in an abnormal state.

15. In paragraph 14, When using the above normal probability value, the probability criterion included in the above abnormal occurrence criterion is a control device of an air conditioner proportional to the above regular time.

16. In paragraph 11, The above estimation part is, The above probability value is estimated according to the set cycle, The above cycle is, A control device for an air conditioner corresponding to the computational speed of the above artificial neural network.

17. In paragraph 11, A control device of an air conditioner further comprising a control unit that stops the compressor in an emergency when it is determined that the above abnormality has occurred.

18. In paragraph 11, The above driving status is, A control device of an air conditioner including at least one of the operating speed, operating frequency, torque, operating current, operating voltage, ambient pressure, suction temperature, discharge temperature, internal temperature, external temperature and control command of the compressor.

19. In paragraph 11, The above estimation part is, A control device for an air conditioner that estimates a probability value of a point between a sensor provided around the compressor and the compressor among one or more of the above points.

20. In paragraph 11, The above estimation part is, An air conditioner control device that further estimates one or more pieces of undetected information that are not detected by the detection unit during the above driving state.

Images