Multi-split air conditioning system and control method thereof

Abstract

A multi-split air conditioning system (10) comprises an outdoor unit (13), at least one indoor unit (12), an expansion valve (111) and a controller (50). The outdoor unit (13) comprises a compressor (131) and an outdoor heat exchanger (132). And the indoor unit (12) comprises an indoor heat exchanger (121). The indoor heat exchanger (121) is connected with the outdoor heat exchanger (132) through the expansion valve (111). The controller (50) is configured to determine a plurality of fault recognition results according to feature data of the multi-split air conditioning system (10) and a plurality of fault recognition models based on a support vector machine under the condition that it is determined that the multi-split air conditioning system (10) breaks down; and the fault type corresponding to the fault recognition result with the maximum probability in the multiple fault recognition results serves as the target fault type of the multi-split air conditioning system (10).

Description

Multi-split air conditioning system and control method thereof

[0001] This application claims priority to Chinese patent application No. 202211542019.X filed on December 2, 2022, and priority to Chinese patent application No. 202310432299.7 filed on April 20, 2023, the entire contents of which are incorporated by reference into this disclosure. Technical Field

[0002] The present disclosure relates to the technical field of air conditioning equipment, and in particular to an air conditioning system. Background Art

[0003] With the continuous advancement of technology and the improvement of people's living standards, air conditioners have become an indispensable electrical appliance in daily life.

[0004] A multi-split air conditioning system is used to provide heating and cooling to a building or room. It typically consists of an outdoor unit and multiple indoor units. Each indoor unit can independently control the temperature, allowing multiple units to be installed in different rooms or areas to meet the needs of different rooms.

[0005] Summary of the Invention

[0006] In one aspect, a multi-split air conditioning system is provided, comprising an outdoor unit, at least one indoor unit, an expansion valve, and a controller. The outdoor unit comprises a compressor and an outdoor heat exchanger; the compressor is connected to the outdoor heat exchanger. Any one of the at least one indoor unit comprises an indoor heat exchanger. The indoor heat exchanger is connected to the outdoor heat exchanger via the expansion valve. Refrigerant circulates through the compressor, the outdoor heat exchanger, the expansion valve, and the indoor heat exchanger to form a refrigerant circuit. The controller is configured to: upon determining that a fault has occurred in the multi-split air conditioning system, determine multiple fault identification results based on characteristic data of the multi-split air conditioning system and multiple fault identification models based on support vector machines; any one of the multiple fault identification models is configured to identify a fault type, and the fault identification result output by any one of the fault identification models is configured to indicate the probability of the multi-split air conditioning system experiencing the fault type corresponding to the fault identification model; and use the fault type corresponding to the fault identification result with the highest probability among the multiple fault identification results as the target fault type of the multi-split air conditioning system.

[0007] On the other hand, a control method for a multi-split air-conditioning system is provided, wherein the multi-split air-conditioning system includes an outdoor unit, at least one indoor unit, and an expansion valve. The outdoor unit includes a compressor and an outdoor heat exchanger. Any one of the at least one indoor unit includes an indoor heat exchanger. Refrigerant circulates in the compressor, the outdoor heat exchanger, the expansion valve, and the indoor heat exchanger to form a refrigerant circuit. The method includes: when it is determined that a fault has occurred in the multi-split air-conditioning system, determining multiple fault identification results based on characteristic data of the multi-split air-conditioning system and multiple fault identification models based on support vector machines; any one of the multiple fault identification models is configured to identify a fault type, and the fault identification result output by any one of the fault identification models is configured to indicate the probability of the multi-split air-conditioning system having the fault type corresponding to any one of the fault identification models; and using the fault type corresponding to the fault identification result with the maximum probability among the multiple fault identification results as the target fault type of the multi-split air-conditioning system.

[0008] On the other hand, a multi-split air-conditioning system is provided, comprising: an outdoor unit, at least one indoor unit, at least one expansion valve, and a controller. The at least one indoor unit is connected to the outdoor unit. The at least one expansion valve corresponds to the at least one indoor unit. Any one of the at least one expansion valve is arranged on a pipeline between the corresponding indoor unit and the outdoor unit. The controller is configured to: obtain real-time operating data of the multi-split air-conditioning system; perform fault identification on the real-time operating data through a preset graph attention model to obtain a fault identification result; the fault identification result includes whether any one of the expansion valves in the multi-split air-conditioning system is in a faulty operating state or a normal operating state; and when any one of the expansion valves is in a faulty operating state, perform fault location on the any one of the expansion valves through the preset graph attention model to determine the indoor unit serial number of the any one of the expansion valves.

[0009] On the other hand, a control method for a multi-split air-conditioning system is provided, wherein the multi-split air-conditioning system includes an outdoor unit, at least one indoor unit, and at least one expansion valve. The at least one expansion valve corresponds to the at least one indoor unit. Any one of the at least one expansion valve is arranged on a pipeline between the corresponding indoor unit and the outdoor unit. The method comprises: obtaining real-time operating data of the multi-split air-conditioning system; performing fault identification on the real-time operating data through a preset graph attention model to obtain a fault identification result; wherein the fault identification result includes whether any one of the expansion valves in the multi-split air-conditioning system is in a faulty operating state or a normal operating state; and when any one of the expansion valves is in a faulty operating state, performing fault location on the any one of the expansion valves through the preset graph attention model to determine the indoor unit serial number of the any one of the expansion valves. BRIEF DESCRIPTION OF THE DRAWINGS

[0010] FIG1A is a block diagram of a multi-split air conditioning system according to some embodiments;

[0011] FIG1B is a structural diagram of an indoor heat exchanger, gas pipes, and liquid pipes of a multi-split air conditioning system according to some embodiments;

[0012] FIG2A is another block diagram of a multi-split air conditioning system according to some embodiments;

[0013] FIG2B is a structural diagram of a multi-split air conditioning system according to some embodiments;

[0014] FIG2C is a structural diagram of an indoor heat exchanger, pipes, and sensors of a multi-split air conditioning system according to some embodiments;

[0015] FIG3 is a structural diagram of a controller and terminal devices of a multi-split air conditioning system according to some embodiments;

[0016] FIG4 is a block diagram of a controller according to some embodiments;

[0017] FIG5 is a flow chart of a method for controlling a multi-split air conditioning system according to some embodiments;

[0018] FIG6 is a flow chart of another method for controlling a multi-split air conditioning system according to some embodiments;

[0019] FIG7 is a schematic diagram showing the impact of a fault level on a multi-split air conditioning system according to some embodiments;

[0020] FIG8 is a schematic diagram of a fault level warning for a multi-split air conditioning system according to some embodiments;

[0021] FIG9 is a flow chart of another method for controlling a multi-split air conditioning system according to some embodiments;

[0022] FIG10 is a flowchart of another method for controlling a multi-split air conditioning system according to some embodiments;

[0023] FIG11 is a flow chart of another method for controlling a multi-split air conditioning system according to some embodiments;

[0024] FIG12 is a schematic diagram of a management interface of a terminal device according to some embodiments;

[0025] FIG13 is a schematic diagram of a management interface of another terminal device according to some embodiments;

[0026] FIG14 is a schematic diagram of a management interface of another terminal device according to some embodiments;

[0027] FIG15 is a flow chart of another control method for a multi-split air conditioning system according to some embodiments;

[0028] FIG16 is a flowchart of a method for training a graph attention model for a multi-split air conditioning system according to some embodiments;

[0029] FIG17 is a flowchart of another method for training a graph attention model for a multi-split air conditioning system according to some embodiments;

[0030] FIG18 is a flowchart of another control method for a multi-split air conditioning system according to some embodiments. DETAILED DESCRIPTION

[0031] The following will be combined with the accompanying drawings to clearly and completely describe the technical solutions in some embodiments of the present disclosure. Obviously, the embodiments described are only some embodiments of the present disclosure, not all embodiments. Based on the embodiments provided by the present disclosure, all other embodiments obtained by ordinary technicians in this field are within the scope of protection of the present disclosure.

[0032] Unless the context requires otherwise, throughout the specification and claims, the term "comprise" and its other forms, such as the third person singular form "comprises" and the present participle form "comprising", are to be interpreted as open and inclusive, that is, "including, but not limited to". In the description of the specification, the terms "one embodiment", "some embodiments", "exemplary embodiments", "example", "specific example" or "some examples" are intended to indicate that the particular features, structures, materials or characteristics associated with the embodiment or example are included in at least one embodiment or example of the present disclosure. The schematic representation of the above terms does not necessarily refer to the same embodiment or example. In addition, the particular features, structures, materials or characteristics may be included in any one or more embodiments or examples in any appropriate manner.

[0033] In the following, the terms "first" and "second" are used for descriptive purposes only and should not be understood as indicating or implying relative importance or implicitly specifying the number of the technical features indicated. Therefore, a feature defined as "first" or "second" may explicitly or implicitly include one or more of the features. In the description of some embodiments of the present disclosure, unless otherwise specified, "plurality" means two or more.

[0034] When describing some embodiments, the word "connected" and its derivatives may be used. The term "connected" should be understood in a broad sense. For example, "connected" can mean fixed connection, detachable connection, or integration; it can be directly connected or indirectly connected through an intermediate medium.

[0035] “At least one of A, B and C” has the same meaning as “at least one of A, B or C” and both include the following combinations of A, B and C: A only, B only, C only, the combination of A and B, the combination of A and C, the combination of B and C, and the combination of A, B and C.

[0036] As used herein, the term "if" is optionally interpreted to mean "when" or "upon" or "in response to determining" or "in response to detecting," depending on the context. Similarly, the phrases "if it is determined that" or "if [stated condition or event] is detected" are optionally interpreted to mean "upon determining" or "in response to determining" or "upon detecting [stated condition or event]" or "in response to detecting [stated condition or event]," depending on the context.

[0037] The use of "adapted to" or "configured to" herein is intended to be open and inclusive language that does not exclude devices adapted or configured to perform additional tasks or steps.

[0038] As used herein, "about," "substantially," or "approximately" includes the stated value and an average value that is within an acceptable range of deviation from the particular value as determined by one of ordinary skill in the art taking into account the measurements in question and the errors associated with the measurement of the particular quantity (i.e., the limitations of the measurement system).

[0039] With the development of the economy and society, air conditioning is becoming increasingly common in various places, including entertainment, home, and work. When multiple small areas within the same area require air conditioning, a multi-split air conditioning system consisting of a single outdoor unit and multiple indoor units is often used to control the room temperature in multiple areas, considering energy conservation.

[0040] However, after extended operation, multi-split air conditioning systems inevitably experience various malfunctions. Some of these malfunctions are caused by gradual performance degradation of the system's equipment, such as heat exchanger fouling, refrigerant leaks, and compressor wear. These malfunctions are often manifested by component aging or wear, making them difficult to detect. Furthermore, when a multi-split air conditioning system malfunctions, the specific symptoms can be similar across multiple systems, making it difficult to accurately identify the component within the system at fault.

[0041] To this end, some embodiments of the present disclosure provide a multi-split air conditioning system. FIG1A is a block diagram of a multi-split air conditioning system according to some embodiments. As shown in FIG1A , the multi-split air conditioning system 10 includes a throttling device 11 , multiple indoor units 12 , and an outdoor unit 13 .

[0042] Multiple indoor units 12 are connected to an outdoor unit 13 via pipelines, allowing refrigerant to circulate within the pipelines. For example, the indoor units 12 may be wall-mounted or cabinet-mounted units, and are configured to exchange heat with indoor air. A throttling device 11 is disposed on the pipelines between the multiple indoor units 12 and the outdoor unit 13, and is configured to regulate the flow rate of the fluid within the pipelines between the multiple indoor units 12 and the outdoor unit 13, as well as the refrigerant flow rate. The outdoor unit 13 is configured to exchange heat with outdoor air.

[0043] FIG1B is a structural diagram of an indoor heat exchanger, gas pipes, and liquid pipes of a multi-split air conditioning system according to some embodiments.

[0044] 1B , the pipes connecting the outdoor unit 13 and the plurality of indoor units 12 include a gas pipe 15 and a liquid pipe 16. The gas pipe 15 is configured to transmit gaseous refrigerant, and the liquid pipe 16 is configured to transmit liquid refrigerant.

[0045] For example, the throttling device 11 is further configured to regulate the flow rate of the fluid within the gas pipe 15 and the liquid pipe 16. For example, the throttling device 11 includes a plurality of expansion valves 111, each corresponding to and connected to the plurality of indoor units 12. The expansion valves 111 are disposed on the pipelines between the corresponding indoor units 12 and the outdoor units 13. For example, referring to FIG. 1B , the expansion valves 111 are disposed on the liquid pipe 16 between the corresponding indoor units 12 (e.g., the indoor heat exchanger 121) and the outdoor units 13.

[0046] In some embodiments, the expansion valve 111 is configured to expand and reduce the pressure of the refrigerant flowing through the expansion valve 111 to regulate the refrigerant supply within the pipeline. For example, when the opening of the expansion valve 111 decreases, the flow resistance of the refrigerant flowing through the expansion valve 111 increases, while when the opening of the expansion valve 111 increases, the flow resistance of the refrigerant flowing through the expansion valve 111 decreases. In this way, by controlling the opening of the expansion valve 111, the flow rate of the refrigerant to the indoor unit 12 can be adjusted while maintaining the states of other components in the refrigerant circuit.

[0047] In some embodiments, the expansion valve 111 may be an electronic expansion valve. In some embodiments, as shown in Figures 2A and 2B , the indoor unit 12 includes an indoor heat exchanger 121. The indoor heat exchanger 121 includes a first port and a second port.

[0048] The first port is configured to be connected to the corresponding expansion valve 111 through the liquid pipe 16, so that the liquid refrigerant can flow between the indoor heat exchanger 121 and the corresponding expansion valve 111. The second port is configured to be connected to the outdoor unit 13 through the gas pipe 15, so that the gaseous refrigerant can flow between the indoor heat exchanger 121 and the outdoor unit 13.

[0049] In some embodiments, the indoor heat exchanger 121 includes a heat transfer pipe connected between the first port and the second port and configured to allow the refrigerant flowing in the heat transfer pipe to exchange heat with the indoor air.

[0050] In some embodiments, as shown in Figures 2A and 2B, the outdoor unit 13 includes a compressor 131 and a liquid accumulator 133. The compressor 131 is disposed between the throttling device 11 and the liquid accumulator 133 and is configured to compress a low-temperature, low-pressure gaseous refrigerant into a high-temperature, high-pressure gaseous refrigerant and discharge it to a condenser (i.e., an indoor heat exchanger or an outdoor heat exchanger). The compressor 131 may be a variable-capacity inverter compressor that performs inverter-based speed control.

[0051] For example, the second port of the indoor heat exchanger 121 is connected to the outdoor unit 13 via the gas pipe 15 , so that the gaseous refrigerant flows between the indoor heat exchanger 121 and the discharge port of the compressor 131 .

[0052] In some embodiments, as shown in FIG. 2A and FIG. 2B , the outdoor unit 13 further includes an outdoor heat exchanger 132 and a four-way valve 134 .

[0053] One end of the outdoor heat exchanger 132 is connected to the accumulator 133 via a four-way valve 134, and the other end is connected to the throttling device 11. The outdoor heat exchanger 132 includes a third port and a fourth port. The third port is configured to allow refrigerant to flow between the outdoor heat exchanger 132 and the suction port of the compressor 131 via the accumulator 133, and the fourth port is configured to allow refrigerant to flow between the outdoor heat exchanger 132 and the throttling device 11.

[0054] For example, the outdoor heat exchanger 132 further includes a heat transfer pipe connected between the third port and the fourth port and configured to perform heat exchange between the refrigerant flowing in the heat transfer pipe and the outdoor air.

[0055] In some embodiments, one end of the liquid accumulator 133 is connected to the compressor 131, and the other end is connected to the outdoor heat exchanger 132 via the four-way valve 134. The liquid accumulator 133 is configured to separate the gas-liquid two-phase refrigerant from the outdoor heat exchanger 132 (e.g., the refrigerant flowing from the outdoor heat exchanger 132 to the compressor 131 via the four-way valve 134) into gaseous refrigerant and liquid refrigerant, and to supply the gaseous refrigerant to the suction port of the compressor 131.

[0056] In some embodiments, the four ports of the four-way valve 134 are respectively connected to the compressor 131, the outdoor heat exchanger 132, the accumulator 133, and the plurality of expansion valves 111. The four-way valve 134 is configured to switch states to change the flow direction of the refrigerant in the refrigerant circuit, thereby switching between cooling mode and heating mode of the multi-split air conditioning system 10.

[0057] In some embodiments, as shown in FIG2B , the outdoor unit 13 further includes an outdoor fan 135. The outdoor fan 135 is disposed near the outdoor heat exchanger 132 and is configured to deliver an airflow of outdoor air to the outdoor heat exchanger 132 to promote heat exchange between the refrigerant flowing in the heat transfer pipe between the third port and the fourth port and the outdoor air.

[0058] In some embodiments, as shown in FIG. 2B , the indoor unit 12 further includes a display 122 , which is configured to display the indoor temperature or the current operating mode of the indoor unit 12 .

[0059] In some embodiments, as shown in FIG2B , the indoor unit 12 further includes an indoor fan 123. The indoor fan 123 is disposed near the indoor heat exchanger 121 and is configured to deliver an airflow of indoor air to the indoor heat exchanger 121 to promote heat exchange between the refrigerant flowing in the heat transfer pipe between the first port and the second port and the outdoor air.

[0060] For example, the indoor unit 12 further includes an indoor fan motor, which is configured to drive the indoor fan 123 to rotate and adjust the rotation speed of the indoor fan 123 .

[0061] In some embodiments, the refrigerant circulates in a refrigerant circuit of the multi-split air conditioning system to enable the multi-split air conditioning system to achieve cooling or heating mode. The refrigerant circuit includes a compressor 131, a condenser, an expansion valve 111, and an evaporator connected in sequence.

[0062] For example, when the multi-split air conditioning system operates in heating mode, the process of refrigerant circulating in the refrigerant circuit includes: the compressor 131 draws the low-temperature, low-pressure gaseous refrigerant after evaporation in the evaporator into the compressor chamber, compresses it into a high-temperature, high-pressure gaseous refrigerant, and transports it to the condenser. The high-temperature, high-pressure gaseous refrigerant is condensed into a high-temperature, high-pressure liquid refrigerant in the condenser, and then, after passing through the throttling device 11 (such as the expansion valve 111), it becomes a low-temperature, low-pressure liquid refrigerant. Next, the low-temperature, low-pressure liquid refrigerant enters the evaporator, is evaporated, and then returns to the compressor 131 to enter the next round of heating cycle.

[0063] It is understood that in the heating mode, the outdoor heat exchanger 132 is used as an evaporator, and the indoor heat exchanger 121 is used as a condenser. In the cooling mode, the outdoor heat exchanger 132 is used as a condenser, and the indoor heat exchanger 121 is used as an evaporator.

[0064] 2C is a structural diagram of an indoor heat exchanger, piping, and sensors of a multi-split air conditioning system according to some embodiments.

[0065] 2A and 2C , the multi-split air conditioning system 10 further includes a plurality of first temperature sensors 101 and a controller 50 . The plurality of first temperature sensors 101 are in communication with the controller 50 .

[0066] Multiple first temperature sensors 101 correspond to multiple indoor units 12. Any one of the multiple first temperature sensors 101 is arranged on the air pipe 15 between the corresponding indoor unit 12 and the outdoor unit 13, and is configured to detect the temperature value of the air pipe 15 and send the detected temperature value of the air pipe 15 to the controller 50.

[0067] For example, the plurality of first temperature sensors 101 may detect temperatures of the plurality of gas pipes 15 between the plurality of indoor units 12 and the outdoor units 13 , and transmit the temperature values ​​of the plurality of gas pipes 15 to the controller 50 .

[0068] In some embodiments, the controller 50 is configured to control the operation of the compressor 131 and the expansion valve 111 so that the multi-split air conditioning system 10 operates to achieve the various predetermined functions of the multi-split air conditioning system. In some embodiments, the controller 50 is further configured to obtain the operating frequency and operating current value of the compressor 131 at each moment.

[0069] 2A and 2C , the multi-split air conditioning system 10 further includes a plurality of second temperature sensors 102 . The plurality of second temperature sensors 102 are communicatively connected to the controller 50 .

[0070] Multiple second temperature sensors 102 correspond to multiple indoor units 12. Any one of the multiple second temperature sensors 102 is arranged on the liquid pipe 16 between the corresponding indoor unit 12 and the outdoor unit 13, and is configured to detect the temperature value of the liquid pipe and send the detected temperature value of the liquid pipe to the controller 50.

[0071] In some embodiments, as shown in FIG2A , the multi-split air conditioning system 10 further includes a plurality of third temperature sensors 103, each of which is communicatively connected to the controller 50. Each of the plurality of third temperature sensors 103 is disposed at an air outlet of the indoor unit and is configured to detect the outlet air temperature of the indoor unit and transmit the detected temperature to the controller 50.

[0072] For example, the plurality of third temperature sensors 103 correspond to the plurality of indoor units 12 , and any third temperature sensor is configured to detect the outlet air temperature of the corresponding indoor unit 12 and send the detected air temperature to the controller 50 .

[0073] In some embodiments, as shown in FIG2A , the multi-split air conditioning system 10 further includes a plurality of fourth temperature sensors 104, each of which is communicatively connected to the controller 50. Any of the plurality of fourth temperature sensors 104 is disposed at an air inlet (i.e., a return air inlet) of the indoor unit and is configured to detect the air inlet temperature (i.e., the return air temperature) of the indoor unit and transmit the detected temperature to the controller 50.

[0074] For example, the plurality of fourth temperature sensors 104 correspond to the plurality of indoor units 12 , and any fourth temperature sensor 104 is configured to detect the inlet air temperature of the corresponding indoor unit 12 and send the detected temperature to the controller 50 .

[0075] In some embodiments, as shown in FIG2A , the multi-split air conditioning system 10 further includes a fifth temperature sensor 105, which is communicatively connected to the controller 50. The fifth temperature sensor 105 is disposed at the air intake of the compressor and is configured to detect the air intake temperature of the compressor and transmit the detected air intake temperature to the controller 50.

[0076] In some embodiments, as shown in FIG2A , the multi-split air conditioning system 10 further includes a sixth temperature sensor 106, which is communicatively connected to the controller 50. The sixth temperature sensor 106 is disposed at the exhaust port of the compressor and is configured to detect the exhaust temperature of the compressor and transmit the detected exhaust temperature to the controller 50.

[0077] In some embodiments, as shown in FIG2A , the multi-split air conditioning system 10 further includes a first pressure sensor 107. The first pressure sensor 107 is communicatively coupled to the controller 50. The first pressure sensor 107 is disposed at the exhaust port of the compressor and is configured to detect the exhaust pressure of the compressor and transmit the detected exhaust pressure to the controller 50.

[0078] In some embodiments, as shown in FIG2A , the multi-split air conditioning system 10 further includes a second pressure sensor 108, which is communicatively connected to the controller 50. The second pressure sensor 108 is disposed at the suction port of the compressor and is configured to detect the suction pressure value of the compressor and transmit the detected suction pressure value to the controller 50.

[0079] It should be noted that the controller 50 in some embodiments of the present disclosure refers to a device that can generate an operation control signal according to an instruction operation code and a timing signal to instruct the multi-split air-conditioning system 10 to execute a control instruction.

[0080] For example, the controller 50 includes at least one of a central processing unit (CPU), a general-purpose processor, a network processor (NP), a digital signal processor (DSP), a microprocessor, a microcontroller, or a programmable logic device (PLD).

[0081] In some embodiments, the controller may also be other devices with processing functions, such as circuits, devices, or software modules, etc., which is not limited in this disclosure.

[0082] In some embodiments, the multi-split air conditioning system 10 further includes a remote controller configured to communicate with the controller 50 via infrared or other communication methods. Thus, the user can control the multi-split air conditioning system (e.g., control the multi-split air conditioning system 10 to execute a heating mode, etc.) via the remote controller, thereby enabling interaction between the user and the multi-split air conditioning system 10.

[0083] In some embodiments, the controller 50 includes a communicator configured to establish a communication connection with other network entities (such as a remote controller or a terminal device).

[0084] For example, the communicator includes a radio-frequency (RF) device configured to receive and transmit signals. For example, the RF device can transmit received information to the controller 50 for processing, or transmit signals generated by the controller.

[0085] For example, the RF device may include an antenna, at least one amplifier, a transceiver, a coupler, a low noise amplifier (LNA), a duplexer, and the like.

[0086] It can be understood that the controller 50 can receive control instructions issued by the user's terminal device through the communicator, and control the multi-split air-conditioning system 10 to perform corresponding processing according to the control instructions, thereby realizing interaction between the user and the multi-split air-conditioning system 10.

[0087] FIG3 is a structural diagram of a controller and terminal devices of a multi-split air conditioning system according to some embodiments.

[0088] 3 , the terminal device 300 can establish a communication connection with the controller 50. For example, a variety of network communication protocols can be used to establish a communication connection between the terminal device 300 and the controller 50.

[0089] For example, the various network communication protocols may include Ethernet, Universal Serial Bus (USB), FireWire, cellular network communication protocols (such as 3G / 4G / 5G), Bluetooth, Wireless Fidelity (Wi-Fi), NFC, etc.

[0090] It should be noted that the terminal device 300 shown in FIG3 is only an example of a terminal device. The terminal device 300 in the present disclosure can be a remote control, a mobile phone, a tablet computer, etc.

[0091] FIG4 is a block diagram of a controller according to some embodiments.

[0092] In some embodiments, as shown in FIG4 , the controller 50 includes an outdoor control device 501 and an indoor control device 502 . The outdoor control device 501 includes a first memory 5011 , and the indoor control device 502 includes a second memory 5021 .

[0093] The indoor control device 502 is connected to the outdoor control device 501 via wired or wireless communication. The outdoor control device 501 can be installed in the outdoor unit 13 or independently of the outdoor unit 13. The outdoor control device 501 is configured to control the outdoor unit 13 to perform relevant operations. The indoor control device 502 can be installed in the indoor unit 12 or independently of the indoor unit 12. The indoor control device 502 is configured to control the components of the indoor unit 12 and the throttling device 11 to perform relevant operations.

[0094] It should be noted that the above division of devices is only functional. For example, the outdoor control device 501 and the indoor control device 502 may also be integrated into one device. For example, the first memory 5011 and the second memory 5021 may also be integrated into one memory.

[0095] In some embodiments, the first memory 5011 is configured to store applications and data related to the outdoor unit 13. The outdoor control device 501 executes various functions and data processing of the multi-split air conditioning system by running the applications and data stored in the first memory 5011.

[0096] The first memory 5011 includes a program storage area and a data storage area. The program storage area can store an operating system and at least one application required for a function (such as an outdoor unit fan start function and an outdoor temperature measurement function). The data storage area can store data generated based on the use of the multi-split air conditioning system (such as outdoor temperature, the opening degree of each expansion valve, etc.). In addition, the first memory 5011 also includes a high-speed random access memory or a non-volatile memory, such as a disk storage device, a flash memory device, or other volatile solid-state memory device.

[0097] In some embodiments, the second memory 5021 is configured to store applications and data related to multiple indoor units 12 and multiple expansion valves 111. The indoor control device 502 performs various functions and data processing of the multi-split air-conditioning system by running the applications and data stored in the memory 5021.

[0098] The second memory 5021 includes a program storage area and a data storage area. The program storage area can store an operating system and at least one application required for a function (such as indoor temperature measurement). The data storage area can store data generated by using the multi-split air conditioning system (such as indoor temperature).

[0099] In some embodiments, the second memory 5021 is further configured to store a correspondence between the address of the indoor unit 12 and the address of the expansion valve 111 .

[0100] In some embodiments, the outdoor control device 501 is communicatively connected to the outdoor unit 13 and is configured to control the outdoor unit 13 to perform relevant operations according to user instructions or system default instructions.

[0101] For example, the outdoor control device 501 can control the speed of the outdoor fan according to the air conditioning operation mode selected by the user. For example, the outdoor control device 501 can also obtain the outdoor temperature according to a user instruction or a system instruction, and store the obtained outdoor temperature in the first memory 5011. For example, the outdoor control device 501 can also control the rotation of the four-way valve 134 in the outdoor unit 13 according to the air conditioning operation mode selected by the user to achieve the selection of cooling or heating mode. For example, the outdoor control device 501 can also control the operation mode, compressor frequency, etc. of the outdoor unit 13 during the address correction process.

[0102] In some embodiments, the indoor control device 502 is in communication with the indoor unit 12 and is configured to control the indoor unit 12 to perform related operations according to user instructions or system default instructions. For example, the indoor control device 502 can also control the indoor unit to turn on the indoor temperature sensor to detect the indoor temperature according to user instructions.

[0103] In some embodiments, the indoor control device 502 is communicatively coupled to the plurality of expansion valves 111 and configured to control the plurality of expansion valves 111 to perform related operations based on user instructions or system default instructions. For example, the indoor control device 502 can control the opening degree of each expansion valve 111 based on user instructions or system instructions.

[0104] It is understood that the hardware structure of the multi-split air conditioning system shown in Figure 2A does not constitute a limitation on the multi-split air conditioning system. In some embodiments, the multi-split air conditioning system may include more or fewer components than shown, or combine certain components, or arrange the components differently.

[0105] Some embodiments of the present disclosure also provide a control method for a multi-split air conditioning system. This control method can input characteristic data of the multi-split air conditioning system into multiple fault identification models to obtain multiple identification results, and can exclude irrelevant data from the operating data, thereby helping to improve the accuracy of fault identification in the multi-split air conditioning system. The following will describe the control method for the multi-split air conditioning system in detail with reference to the accompanying drawings.

[0106] It should be noted that the control method of the multi-split air conditioning system is executed by a controller. The controller can be the controller 50 described in any of the above embodiments, or other controllers, which are not limited in this disclosure.

[0107] FIG5 is a flow chart of a method for controlling a multi-split air conditioning system according to some embodiments.

[0108] In some embodiments, as shown in FIG5 , the control method of the multi-split air conditioning system includes step S101 and step S102 .

[0109] In step S101, when it is determined that a fault occurs in the multi-split air-conditioning system, multiple fault identification results are determined according to feature data of the multi-split air-conditioning system and multiple fault identification models based on support vector machines.

[0110] For example, the memory of the multi-split air conditioning system may pre-store multiple trained fault identification models based on support vector machines. When a fault is determined in the multi-split air conditioning system, the controller 50 may input characteristic data of the multi-split air conditioning system into the multiple fault identification models based on support vector machines.

[0111] It should be noted that the characteristic data of the multi-split air-conditioning system is obtained after correlation analysis of the operating data of the multi-split air-conditioning system. The characteristic data is the data related to the fault that is retained after removing data unrelated to the fault from the operating data of the multi-split air-conditioning system.

[0112] It can be understood that by performing correlation analysis on the operating data of the multi-split air-conditioning system and removing data that is not related to the fault, the impact of irrelevant data on the accuracy and efficiency of fault identification can be eliminated, thereby improving the accuracy and efficiency of fault identification of the multi-split air-conditioning system.

[0113] Regarding the method of performing correlation analysis on the operating data of the multi-split air-conditioning system to obtain characteristic data of the multi-split air-conditioning system, reference may be made to the following steps S301 to S302 , which will not be described in detail here.

[0114] The characteristic data of the multi-split air-conditioning system is input into multiple fault identification models, one fault identification model corresponds to one fault type, and the fault identification result output by one fault identification model is configured to indicate the probability of the multi-split air-conditioning system having the fault type corresponding to the fault identification model.

[0115] It is understood that a fault identification model is configured to identify a single fault type. When performing fault identification on a multi-split air conditioning system, the fault identification model only identifies the probability of the multi-split air conditioning system experiencing the fault type corresponding to that fault identification model, and is not affected by other fault types during fault identification. Based on the operating data of the multi-split air conditioning system and multiple support vector machine-based fault identification models, multiple fault identification results are determined, i.e., the probability of the multi-split air conditioning system experiencing the fault type corresponding to each fault identification model is obtained.

[0116] For example, the failure types of a multi-split air conditioning system include: abnormal refrigerant charge, dirty and blocked outdoor heat exchanger, dirty and blocked indoor heat exchanger, compressor wear, and expansion valve failure.

[0117] In some embodiments, the characteristic data of the multi-split air-conditioning system include at least one of the operating frequency of the compressor, the exhaust pressure value of the compressor, the suction pressure value of the compressor, the suction temperature value of the compressor, the exhaust temperature value of the compressor, the exhaust superheat of the compressor, the suction superheat of the compressor, the opening of the outdoor fan gear expansion valve, the outlet air temperature of each indoor unit, the return air temperature of each indoor unit, the temperature value of the gas pipe 15 and the temperature value of the liquid pipe 16.

[0118] In some embodiments, the support vector machine is a binary classification model, and the basic model of the support vector machine is a linear classifier with the largest interval defined in the feature space. The support vector machine also includes a kernel technique, which makes the support vector machine a substantially nonlinear classifier. The support vector machine is based on the VC dimension theory (Vapnik-Chervonenkis Dimension) of statistical learning theory and the principle of structural risk minimization. Based on limited sample information, the model seeks the best compromise between complexity (i.e., learning accuracy for specific training samples) and learning ability (i.e., the ability to identify any sample without error) of the model in order to obtain better generalization ability.

[0119] In some embodiments, the controller 50 may train multiple fault identification models based on a support vector machine based on the historical operating data of the multi-split air conditioning system, and store the trained multiple fault identification models in a memory, so that when performing a fault identification function, the controller 50 can promptly identify the type of fault occurring in the multi-split air conditioning system based on the trained fault identification models. For example, the historical operating data of the multi-split air conditioning system includes normal operating data of the multi-split air conditioning system during normal operation and abnormal operating data during abnormal operation (also referred to as historical fault data).

[0120] In some embodiments, the training process of the fault identification model includes: establishing a data regression line, a regression plane, and a hyperplane for fitting, and adjusting the accuracy and range of the fitting through a loss function, setting a deviation value, and a relaxation variable.

[0121] In some embodiments, the controller 50 is further configured to perform model testing after the fault identification model training is completed. The test results of the fault identification model can be visualized through the confusion matrix and the diagnostic timing diagram to obtain the diagnostic results, and three model evaluation indicators are introduced: geometric mean accuracy (GMA), false alarm rate (FAR), and missed alarm rate (MAR) for evaluation. For example, the GMA indicator represents the geometric mean of the accuracy of each classification category, and after obtaining the actual results, a judgment is made. If the accuracy does not meet the requirements, the parameter update and tuning device will be activated for self-optimization.

[0122] In step S102, the fault type corresponding to the fault identification result with the highest probability among the multiple fault identification results is used as the target fault type of the multi-split air-conditioning system.

[0123] It can be understood that one fault identification model corresponds to one fault type. The greater the fault probability indicated by a fault identification result, the higher the probability that the multi-split air-conditioning system will have the fault type corresponding to the fault identification result. Therefore, the fault type corresponding to the fault identification result with the highest probability among multiple fault identification results can be used as the target fault type of the multi-split air-conditioning system.

[0124] Exemplarily, the fault identification models include A1, B1, and C1, with A1 corresponding to fault type A, B1 corresponding to fault type B, and C1 corresponding to fault type C. When a multi-split air conditioning system is determined to have a fault, the characteristic data of the multi-split air conditioning system is input into the three fault identification models A1, B1, and C1 to obtain fault identification results from the three fault identification models. In this case, if the fault identification result output by A1 indicates a 50% probability of fault type A occurring in the multi-split air conditioning system, the fault identification result output by B1 indicates an 80% probability of fault type B occurring in the multi-split air conditioning system, and the fault identification result output by C1 indicates a 30% probability of fault type C occurring in the multi-split air conditioning system, then fault type B can be selected as the target fault type for the multi-split air conditioning system.

[0125] It is understood that in the control method for a multi-split air conditioning system provided in some embodiments of the present disclosure, after determining that a fault has occurred in the multi-split air conditioning system, the characteristic data of the multi-split air conditioning system is input into multiple fault identification models, thereby obtaining multiple identification results. Because the characteristic data of the multi-split air conditioning system is obtained through correlation analysis based on the operating data of the multi-split air conditioning system, the multiple identification results eliminate the impact of irrelevant data in the operating data on the accuracy of fault identification. In this case, the fault type corresponding to the fault identification result with the highest probability among the multiple fault identification results is used as the target fault type for the multi-split air conditioning system, which is conducive to improving the accuracy of fault identification in the multi-split air conditioning system.

[0126] FIG6 is a flowchart of another method for controlling a multi-split air conditioning system according to some embodiments.

[0127] In some embodiments, as shown in FIG6 , after confirming the target fault type of the multi-split air conditioning system (ie, after step S102 ), the method further includes steps S201 to S202 .

[0128] In step S201, a fault level identification result is determined based on the characteristic data of the multi-split air-conditioning system and a fault level identification model corresponding to the target fault type.

[0129] It's understandable that when a multi-split air conditioning system fails, different fault levels will have varying impacts on its operation. If the fault level is minor, it can be understood that the current fault will likely not significantly impact the system's operation, and repairs can be temporarily deferred. However, if the fault level is severe, the system may become inoperable, requiring repairs. Therefore, after determining the target fault type for the multi-split air conditioning system, it is necessary to derive a fault level identification result based on the system's characteristic data to determine whether repairs are necessary.

[0130] In some embodiments, a plurality of fault level identification models are pre-stored in the memory of the multi-split air conditioning system, and one fault level identification model is configured to identify the fault level of a fault type.

[0131] Therefore, after determining the target fault type of the multi-split air-conditioning system, the characteristic data of the multi-split air-conditioning system can be input into the fault level identification model corresponding to the target fault type to obtain the fault level identification result. The fault level identification result represents the degree of impact of the target fault type on the multi-split air-conditioning system.

[0132] In some embodiments, the fault level identification model may be a fault level identification model based on a machine learning algorithm.

[0133] For example, the fault level identification results include: Fault Level 1, Fault Level 2, Fault Level 3, Fault Level 4, and Fault Level 5. If the fault level identification result is Fault Level 1 or Fault Level 2, it means that the user needs to pay attention and is recommended to pay attention to the inspection. If the fault level identification result is Fault Level 3, it means that the user needs to pay attention and is recommended to repair in the near future. If the fault level identification result is Fault Level 4, it means that the user needs to pay sufficient attention and is recommended to repair in a timely manner. If the fault level identification result is Fault Level 5, it means that the user needs to pay great attention and is recommended to repair immediately.

[0134] In step S202, when the fault level indicated by the fault level identification result is above a preset fault level, an alarm message is issued.

[0135] It is understood that when the fault level indicated by the fault level identification result is above the preset fault level, it means that the fault level corresponding to the current target fault type is high and may cause the multi-split air conditioning system to not operate normally. To ensure the normal operation of the multi-split air conditioning system, an alarm message can be issued to prompt maintenance personnel to inspect the multi-split air conditioning system.

[0136] In some embodiments, the alarm information includes a target fault type so that maintenance personnel can conduct targeted repairs on the multi-split air-conditioning system based on the target fault type, which helps to improve the fault repair efficiency of the multi-split air-conditioning system.

[0137] For example, the preset fault level may be preset when the multi-split air conditioning system leaves the factory.

[0138] FIG7 is a schematic diagram illustrating the impact of a fault level on a multi-split air conditioning system according to some embodiments.

[0139] For example, as shown in FIG7 , if the multi-split air conditioning system is at fault level 5 or below, it means that the fault in the multi-split air conditioning system is minor, and the impact on the multi-split air conditioning system is small when the fault level is below 5. If the multi-split air conditioning system is at fault level 5 or above, it means that the fault in the multi-split air conditioning system is serious, that is, the impact on the multi-split air conditioning system is large when the fault level is above 5.

[0140] FIG8 is a schematic diagram of a fault level warning for a multi-split air conditioning system according to some embodiments.

[0141] For example, as shown in FIG8 , to prevent a minor fault in a multi-split air conditioning system from developing into a major fault, the preset fault level can be set to Fault Level 2. That is, when the fault level of the multi-split air conditioning system is determined to be above Fault Level 2, the controller issues an alarm message to alert maintenance personnel to perform an inspection, thereby preventing a minor fault in the multi-split air conditioning system from developing into a major fault. For example, the controller can prevent the fault level of the multi-split air conditioning system from developing from Fault Level 2 to Fault Level 5.

[0142] In some embodiments, the controller 50 may issue the warning information in a variety of ways, for example, the ways include way 1, way 2, and way 3.

[0143] Method 1 includes: the controller controls the display of the indoor unit to display the alarm information.

[0144] For example, when the target fault type of a multi-split air conditioning system is a dirty or blocked outdoor heat exchanger, the alarm message may read "The outdoor heat exchanger is seriously dirty or blocked. Immediate repair is recommended!".

[0145] In some embodiments, in order to facilitate users to promptly learn that a serious fault has occurred in the multi-split air-conditioning system, the controller can control the display of each indoor unit in the multi-split air-conditioning system to display the above-mentioned alarm information.

[0146] Method 2 includes: the controller sends alarm information to the terminal device through the communicator.

[0147] For example, when the target fault type of the multi-split air-conditioning system is the outdoor heat exchanger being dirty and blocked, the content of the alarm information received by the terminal device from the controller 50 via the Wi-Fi network or Bluetooth may be "The outdoor heat exchanger is seriously dirty and blocked, and immediate maintenance is recommended!"

[0148] Method 3 includes: the controller sends an alarm message to the terminal device through a voice prompt device.

[0149] In some embodiments, the indoor unit further includes a voice device, which may be a speaker, etc. The controller may control the voice device to broadcast an alarm message to attract the user's attention and remind the user to perform maintenance.

[0150] FIG9 is a flowchart of another method for controlling a multi-split air conditioning system according to some embodiments.

[0151] In some embodiments, as shown in FIG9 , before step 101 , the method further includes steps S301 to S303 .

[0152] In step S301, before determining that a failure occurs in the multi-split air-conditioning system, operation data of the multi-split air-conditioning system is obtained.

[0153] It is understandable that the fault type can only be further identified after it is determined that the multi-split air conditioning system has failed. Therefore, before determining that the multi-split air conditioning system has failed, the operating data of the multi-split air conditioning system can be obtained to determine whether the multi-split air conditioning system has failed.

[0154] Fig. 10 is a flow chart of another method for controlling a multi-split air conditioning system according to some embodiments. In some embodiments, as shown in Fig. 10 , step S301 includes steps S3011 to S3012.

[0155] In step S3011, original operation data of the multi-split air conditioning system is obtained.

[0156] For example, when the multi-split air-conditioning system is in operation, the controller can obtain original operation data generated by various components of the multi-split air-conditioning system during operation.

[0157] In step S3012, the original operation data of the multi-split air-conditioning system is pre-processed to obtain the operation data of the multi-split air-conditioning system.

[0158] In some embodiments, the preprocessing includes outlier removal and smoothing.

[0159] For example, the abnormal value elimination process includes: in the process of the controller acquiring the original operating data, eliminating abnormal operating data with a higher degree of discreteness compared with other operating data.

[0160] It is understandable that if the raw operating data of the multi-split air conditioning system is not filtered for outliers, the abnormal operating data may affect the accuracy of the fault diagnosis results during subsequent fault diagnosis. Therefore, after obtaining the raw operating data of the multi-split air conditioning system, it can be processed to remove outliers, thereby improving the accuracy of fault diagnosis.

[0161] For example, the smoothing process includes removing outliers from the raw operating data of the multi-split air conditioning system, then performing least squares curve fitting on the raw operating data after outlier removal using a smoothing algorithm, and replacing the removed data with the fitted data. Least squares fitting is a mathematical approximation and optimization method that, based on known data, generates a straight line or curve in a coordinate system such that the sum of the squares of the distances between the fitted points and the known data is minimized.

[0162] For example, the smoothing algorithm includes the Savitzky-Golay algorithm.

[0163] In some embodiments, performing data smoothing on the original operating data of the multi-split air-conditioning system may further include: inputting the original operating data after outlier removal processing into a generative network adversarial model to obtain the operating data of the multi-split air-conditioning system.

[0164] It should be noted that the generative adversarial network model is a deep learning model that, through its framework, includes a generative model and a discriminative model. The generative model is configured to generate fitted data, while the discriminative model is configured to judge the fitted data. For example, the discriminative model can determine whether the data is smooth. The generative and discriminative models interact to produce smoothed data.

[0165] In some embodiments, the operating data of the multi-split air-conditioning system may include: the compressor current value of the outdoor unit in the multi-split air-conditioning system during operation, the operating frequency of the compressor, the exhaust pressure value of the compressor, the suction pressure value of the compressor, the suction temperature value of the compressor, the exhaust temperature value of the compressor, the exhaust superheat of the compressor, the suction superheat of the compressor, the gear position of the outdoor fan, the opening of the expansion valve, the outlet air temperature of each indoor unit, the return air temperature of each indoor unit, the temperature value of the gas pipe, the temperature value of the liquid pipe, and at least one of the discharge pressure value and discharge temperature value at the refrigerant discharge pipe of the outdoor unit.

[0166] It should be noted that the operating data of the multi-split air-conditioning system shown above is merely exemplary, and the operating data of the multi-split air-conditioning system may also include other data, which will not be described in detail here.

[0167] In step S302, a correlation analysis is performed on the operating data of the multi-split air-conditioning system based on the maximum information coefficient method, and characteristic data of the multi-split air-conditioning system is extracted from the operating data of the multi-split air-conditioning system.

[0168] Understandably, the operating data of multi-split air conditioning systems contains a large amount of redundant data that is useless for fault diagnosis, and this redundant data can affect the efficiency and accuracy of fault diagnosis. Therefore, to prevent the redundant data in the operating data of multi-split air conditioning systems from affecting the efficiency and accuracy of fault diagnosis, a correlation analysis of the operating data of multi-split air conditioning systems can be performed based on the maximum information coefficient method to extract characteristic data of the multi-split air conditioning systems from the operating data.

[0169] The maximum information coefficient method is a feature selection algorithm used to measure the degree of association between two variables. It can measure the correlation between various fault types and the operating data of a multi-split air conditioning system, and analyze and filter the operating data based on the correlation between the operating data and the fault type.

[0170] Operation data with a low degree of correlation with the fault type means that the correlation between the operation data and the fault type is small. Therefore, the operation data can be used as redundant data. On the contrary, operation data with a high degree of correlation with the fault means that the correlation between the operation data and the fault type is large. Therefore, the operation data can be used as feature data.

[0171] In some embodiments, before performing correlation analysis on the operating data of the multi-split air-conditioning system based on the maximum information coefficient method and extracting the characteristic data of the multi-split air-conditioning system from the operating data of the multi-split air-conditioning system (i.e., before step S302), the importance of different operating data for different fault types can also be calculated based on the Gini variable importance and the association rule algorithm, and the operating data can be sorted according to the importance.

[0172] The Gini variable importance measure is based on the Gini index, which measures the importance of variables. The Gini index indicates the probability that a randomly selected sample from a sample set will be misclassified. A smaller Gini index indicates a lower probability of misclassification, meaning the set is more pure. The smaller the Gini index for a feature, the more important it is.

[0173] In step S303, based on the characteristic data of the multi-split air-conditioning system, it is determined whether a fault occurs in the multi-split air-conditioning system.

[0174] As can be seen from step S302, the characteristic data of the multi-split air conditioning system is extracted from the operating data of the multi-split air conditioning system through correlation analysis. The characteristic data of the multi-split air conditioning system can reflect the operating status of the multi-split air conditioning system. If a malfunction occurs in the multi-split air conditioning system, the characteristic data of the multi-split air conditioning system may also fluctuate.

[0175] For example, after the refrigerant passes through the throttle valve of the indoor unit, it begins to evaporate and absorb heat. When the expansion valve connected to the indoor unit is functioning normally (i.e., not faulty), the temperature of the gas pipe and the temperature of the liquid pipe connected to the indoor unit should be equal. In other words, the temperature difference between the temperature of the liquid pipe and the temperature of the gas pipe is approximately zero. However, when the expansion valve fails, for example, if the expansion valve is opened too small due to a faulty expansion valve, insufficient refrigerant flow will result, causing the refrigerant to overheat during evaporation and heat absorption, which in turn increases the temperature difference between the temperature of the liquid pipe and the temperature of the gas pipe connected to the indoor unit. Therefore, whether the expansion valve is faulty can be determined based on characteristic data (such as the temperature of the gas pipe and the temperature of the liquid pipe).

[0176] FIG11 is a flowchart of another method for controlling a multi-split air conditioning system according to some embodiments.

[0177] In some embodiments, as shown in FIG11 , step S303 includes steps S3031 to S3032 .

[0178] In step S3031, a fault diagnosis result is determined based on the characteristic data and the fault diagnosis model of the multi-split air conditioning system.

[0179] In some embodiments, a trained fault diagnosis model is pre-stored in the memory of the multi-split air-conditioning system. When performing fault diagnosis on the multi-split air-conditioning system, characteristic data of the multi-split air-conditioning system can be input into the trained fault diagnosis model to obtain a fault diagnosis result, which indicates whether a fault has occurred in the multi-split air-conditioning system.

[0180] In some embodiments, the fault diagnosis model may be a fault diagnosis model based on a support vector machine.

[0181] In some embodiments, the fault diagnosis model training process includes: setting up a gradient multi-split air-conditioning system simulation experiment for each fault type, collecting experimental data for each fault type and classifying the experimental data according to normal and abnormal multi-split air-conditioning systems for training the fault diagnosis model.

[0182] For example, taking the simulation experiment of refrigerant filling amount as an example, multiple simulation experiments are set up with a refrigerant filling amount decreasing by 10% from 120% (overcharge) to 50% (undercharge) and the experimental data of each simulation experiment are collected.

[0183] It is understood that in the simulation experiment, when the performance of the multi-split air conditioning system degrades and reaches a first preset threshold, it can be considered that the multi-split air conditioning system is abnormal. In this case, the experimental data of this simulation experiment is marked as abnormal. If the performance degradation of the multi-split air conditioning system does not reach the first preset threshold, the multi-split air conditioning system is considered to be normal, and the experimental data of this simulation experiment is marked as normal.

[0184] In step S3032, if the fault diagnosis result is yes, it is determined that a fault occurs in the multi-split air conditioning system.

[0185] In some embodiments, when the fault diagnosis result is negative, it is determined that no fault occurs in the multi-split air conditioning system.

[0186] It can be understood that the control method provided by some embodiments of the present disclosure can perform correlation analysis on the operating data of the multi-split air-conditioning system based on the maximum information coefficient method, and extract feature data from the operating data of the multi-split air-conditioning system, so that fault diagnosis, fault identification and fault level identification can be performed based on the feature data with small data volume and strong representativeness, thereby improving the efficiency and accuracy of fault diagnosis, fault identification and fault level identification.

[0187] In some embodiments, the control method of a multi-split air conditioning system provided by the embodiments of the present disclosure further includes a training process for a fault level identification model. The training process of the fault level identification model includes steps A1 to A4.

[0188] In step A1, data acquisition is performed.

[0189] In some embodiments, the data collection includes collecting data for training a fault level identification model. For example, a gradient multi-split air conditioning system simulation experiment can be conducted for each fault type, and experimental data for each fault type can be collected. The experimental data can then be classified according to the fault level of the multi-split air conditioning system for use in training the fault diagnosis model.

[0190] It should be noted that the specific steps of the simulation experiment and data collection have been introduced in detail above and will not be repeated here.

[0191] In step A2, data pre-processing is performed.

[0192] In some embodiments, after collecting experimental data configured for fault level identification model training, the collected experimental data needs to be preprocessed to eliminate abnormal experimental data and perform data smoothing on the experimental data. It should be noted that the specific steps of data preprocessing have been described in detail above and will not be repeated here.

[0193] In step A3, important feature data selection is performed.

[0194] In some embodiments, after the experimental data is preprocessed, the importance of different operating data for different fault types is calculated using the Gini variable importance and association rule algorithm, and the operating data is sorted according to the importance. The experimental data is then subjected to correlation analysis using the maximum information coefficient method to remove redundant data that is not related to the fault.

[0195] It should be noted that the Gini variable importance and maximum information coefficient method have been introduced in detail above and will not be repeated here.

[0196] In some embodiments, selecting important feature data further includes: decoupling faults.

[0197] Decoupling involves converting mathematical equations containing multiple variables into a system of equations that can be expressed using a single variable. This means that the variables no longer directly influence the outcome of an equation, thereby simplifying analysis and calculations. In multi-split air conditioning systems, faults can be correlated, causing a single fault in the system to trigger other faults. Fault decoupling involves eliminating the influence of other faults on a specific fault when analyzing the correlation between experimental data and that fault.

[0198] In some embodiments, selecting important feature data further includes analyzing existing experimental data in conjunction with an expert knowledge system. After sorting the operational data by importance and performing correlation analysis using the maximum information coefficient method, the expert knowledge system can be combined to select important feature data. Based on the operating principles of multi-split air conditioning systems, an expert knowledge system is constructed to comprehensively analyze the experimental data, which has been sorted by importance and analyzed for correlation using the maximum information coefficient method, to select important feature data.

[0199] In step A4, model training is performed.

[0200] In some embodiments, after determining the important feature data, a fault level identification model is trained based on a support vector machine model. After the training is completed, the fault level identification model is evaluated. If the accuracy of the fault level identification model does not meet the requirements, the parameter update and tuning model will be started for self-optimization.

[0201] It should be noted that the model training and optimization methods have been introduced in detail above and will not be repeated here.

[0202] FIG12 is a schematic diagram of a management interface of a terminal device according to some embodiments.

[0203] In some embodiments, as shown in FIG12 , the terminal device 300 is further configured to display a management interface 301 of the multi-split air conditioning system, so that the user can set the operating mode of the multi-split air conditioning system through the terminal device 300 .

[0204] For example, the management interface 301 includes a first button 302. The terminal device 300 is configured to, upon detecting that a user has clicked the first button 302 in the management interface 301, pop up a drop-down selection box 303 in the management interface 301. The terminal device 300 is further configured to, upon detecting that a user has selected a command in the drop-down selection box 303 and confirmed the command, send the user-confirmed command to the multi-split air conditioning system to complete the operating mode setting.

[0205] For example, the user may select a cooling mode instruction in the drop-down selection box 303 .

[0206] FIG13 is a schematic diagram of a management interface of another terminal device according to some embodiments.

[0207] In some embodiments, as shown in Figure 13, the management interface 301 further includes a second button 304. The second button 304 is configured to switch between an off state and an on state when clicked by a user. The terminal device 300 is further configured to transmit a fault detection activation instruction to the controller of the multi-split air conditioning system when the second button is switched to the on state, causing the multi-split air conditioning system to enter a fault detection mode. This allows users to activate the fault detection function through the management interface of the terminal device 300, thereby improving the efficiency of interaction between the user and the multi-split air conditioning system.

[0208] FIG14 is a schematic diagram of a management interface of another terminal device according to some embodiments.

[0209] In some embodiments, the controller 50 is further configured to: after detecting an expansion valve failure in the multi-split air-conditioning system, send an instruction indicating that a failure exists in the multi-split air-conditioning system to the terminal device 300 through the communicator.

[0210] As shown in Figure 14, the management interface 301 is also configured to: after the terminal device 300 receives an instruction indicating that there is a fault in the multi-split air conditioning system, it displays a prompt message saying "An expansion valve fault has been detected in the multi-split air conditioning system. Do you want to locate the fault immediately?" to prompt the user to choose whether to turn on the fault location function. In this case, if the user chooses to click the "OK" button, it means that the user has chosen to locate the fault immediately. In response to the user's confirmation instruction, the terminal device 300 sends the confirmation instruction (i.e., the fault location instruction) to the controller 50. The controller 50 is also configured to turn on the fault location mode after receiving the fault location instruction. If the user chooses to click the "Cancel" button, it means that the user has chosen not to locate the fault for the time being.

[0211] In some embodiments, as shown in FIG14 , the management interface 301 further includes a third button 305. The third button 305 is configured to switch between an off state and an on state when clicked by a user. The terminal device 300 is further configured to send the fault location instruction to the controller 50 when the third button 305 switches to the on state.

[0212] FIG15 is a flowchart of another control method for a multi-split air conditioning system according to some embodiments.

[0213] Some embodiments of the present disclosure also provide another method for controlling a multi-split air conditioning system. The method is executed by a controller. The controller may be the controller 50 described in any of the above embodiments, or another controller.

[0214] As shown in FIG. 15 , the method includes steps S401 to S404 .

[0215] In step S401, real-time operating data of the multi-split air-conditioning system is obtained.

[0216] It is understandable that when an expansion valve in a multi-split air conditioning system fails, it is necessary to perform fault detection on all expansion valves in the multi-split air conditioning system to locate the faulty expansion valve. Therefore, by acquiring real-time operating data of the multi-split air conditioning system, it is possible to detect and locate the faulty expansion valve.

[0217] In some embodiments, the controller is further configured to obtain real-time operating data of the multi-split air conditioning system at a first predetermined time interval after the multi-split air conditioning system is started. In this way, the controller can periodically perform fault detection on the multi-split air conditioning system, thereby improving the stability and reliability of the multi-split air conditioning system.

[0218] It should be noted that the first preset time length is data that has been preset in the memory of the multi-split air-conditioning system when it leaves the factory. The value of the first preset time length can be set according to actual usage requirements, and this disclosure does not limit this.

[0219] In some embodiments, the real-time operating data of the multi-split air conditioning system includes operating data of the outdoor unit and operating data of each indoor unit.

[0220] For example, the operating data of the outdoor unit includes: compressor frequency, pressure value, pressure top temperature value, exhaust gas superheat value, defrost temperature value and ambient temperature value.

[0221] For example, the operating data of the indoor unit includes: variable frequency heat dissipation value, liquid pipe temperature value, gas pipe temperature value, return air temperature value and outlet air temperature value.

[0222] In step S402, fault identification is performed on the real-time operation data through a preset graph attention model to obtain a fault identification result.

[0223] It should be noted that the Graph Attention Networks (GAT) is a graph neural network model that uses the self-attention mechanism. The network uses a method similar to the self-attention in the neural network model to calculate the attention of any node in the graph relative to each adjacent node, and combines the characteristics of the node itself and the attention characteristics as the characteristics of the node, and performs tasks such as node classification on this basis.

[0224] In some embodiments, the graph attention model includes a fault detection model. The fault detection model is configured to detect whether any one of the plurality of expansion valves is in a faulty operating state. The fault identification result includes whether the expansion valve is in a faulty operating state or a normal operating state.

[0225] Therefore, after the controller obtains the real-time operating data of the multi-split air-conditioning system, it inputs the real-time operating data into the preset graph attention model, thereby being able to perform fault identification on multiple expansion valves and obtain fault identification results.

[0226] In some embodiments, the preset graph attention model can be a fault detection model in the GAT model.

[0227] In step S403, it is determined whether the fault identification result is a faulty operating state. If so, step S404 is executed; if not, step S401 is executed again.

[0228] It is understood that if the controller identifies that any one of the multiple expansion valves is in a faulty operating state, it is necessary to continue executing step S404 to locate the location of the expansion valve. If all the expansion valves are in a normal operating state, it is not necessary to perform fault location. In this case, the controller re-executes step S401.

[0229] In some embodiments, step S403 includes: when the fault identification result is that the expansion valve in the multi-split air-conditioning system is in a normal operating state, re-execute step S401 after a second preset time period.

[0230] It should be noted that the second preset time length is data that has been preset in the memory of the multi-split air-conditioning system when it leaves the factory. The value of the second preset time length can be set according to actual usage requirements, and this disclosure does not limit this.

[0231] In step S404, the expansion valve fault is located using a preset graph attention model to determine the internal unit serial number of the expansion valve.

[0232] In some embodiments, the graph attention model further includes a fault location model, which is configured to locate the internal unit serial number of the expansion valve in a faulty operating state.

[0233] For example, when the fault identification result is that the expansion valve in the multi-split air-conditioning system is in a faulty operating state, each original node feature vector (which will be introduced in detail below) is input into the fault location model in the GAT model, so that the internal unit serial number of the expansion valve in the faulty operating state can be located.

[0234] In some embodiments, the preset graph attention model may be a fault location model in the graph attention model.

[0235] It will be appreciated that some embodiments of the present disclosure provide another control method for a multi-split air conditioning system, which inputs acquired real-time operating data of the multi-split air conditioning system into a preset graph attention model to obtain a fault identification result for the expansion valve in the multi-split air conditioning system, thereby improving the efficiency of expansion valve fault detection. Furthermore, when an expansion valve is determined to be in a faulty operating state, the preset graph attention model is used to locate the expansion valve and obtain the internal unit serial number of the first expansion valve, thereby reducing the data stream preprocessing process and improving the accuracy of expansion valve fault location.

[0236] FIG16 is a flowchart of a method for training a graph attention model for a multi-split air conditioning system according to some embodiments.

[0237] In some embodiments, the GAT model can be trained by the controller. The training method of the GAT model is shown in FIG16 , which is executed by the controller and includes steps S11 to S12.

[0238] In step S11, normal operation data and fault operation data of the multi-split air-conditioning system are obtained.

[0239] It should be noted that the normal operating data refers to the operating data of the multi-split air conditioning system when the expansion valve is in a normal operating state, and the fault operating data refers to the operating data of the multi-split air conditioning system when the expansion valve is in a faulty operating state.

[0240] For example, the controller can input the normal operating data and faulty operating data of a multi-split air conditioning system into the GAT model for training. This allows the GAT model to capture the operating data characteristics of the expansion valve when it is in a normal state, as well as the operating data characteristics of the expansion valve when it is in a faulty state. This allows the GAT model to more accurately identify whether there is faulty operating data in the real-time operating data of the multi-split air conditioning system.

[0241] In some embodiments, the training principle of the GAT model can be based on a graph neural network (GNN). A graph neural network is a connection model that obtains dependencies in a graph by passing information between nodes in the network. GNN updates the node state from its neighbors at any depth, and this state can represent state information.

[0242] In step S12, the initial graph attention model is trained based on the normal operation data and the fault operation data to obtain a preset graph attention model after training.

[0243] In some embodiments, before the controller inputs the normal operation data and the fault operation data into the initial graph attention model to obtain the preset graph attention model after training, the controller can also convert the normal operation data and the fault operation data into a graph structure through the GNN model, perform node aggregation operations on the graph structure, iteratively update the GAT model parameters, and finally output the fault detection and positioning results of the expansion valve.

[0244] FIG17 is a flowchart of another method for training a graph attention model for a multi-split air conditioning system according to some embodiments. As shown in FIG17 , the method includes step S21 .

[0245] In step S21, the normal operation data and the fault operation data are converted into a graph structure through a preset graph neural network model.

[0246] In some embodiments, before the controller trains the initial graph attention model based on the normal operation data and the fault operation data, the controller converts the normal operation data and the fault operation data into a graph structure through a preset graph neural network model.

[0247] For example, the graph structure includes original node feature vectors and connection relationships between original node feature vectors. The original feature vectors are vector representations of normal operation data and fault operation data.

[0248] In some embodiments, the connection relationship between the original node feature vectors can be presented in the form of edges, and the edges are represented in the form of an adjacency matrix.

[0249] In some embodiments, a Pearson correlation coefficient is calculated for different data information between two original node feature vectors. If the Pearson correlation coefficient is greater than a second preset threshold, the two original node feature vectors are connected by an edge, i.e., the two original node feature vectors are correlated; if the Pearson correlation coefficient is less than the second preset threshold, the two original node feature vectors are not connected by an edge, i.e., the two original node feature vectors are uncorrelated.

[0250] In some embodiments, the second preset threshold is 0.5.

[0251] For example, if the Pearson correlation coefficient A is greater than a second preset threshold, ie, A>0.5, the two original node feature vectors are connected by an edge. If A<0.5, the two original node feature vectors are not connected by an edge.

[0252] For example, the graph structure can be expressed as shown in formula (1): AM = [S1, T1], [S2, T2], [S3, T3], ..., [S n , T n ]] Formula (1)

[0253] In formula (1), AM is the adjacency matrix, S and T are different running data, and n is the location of the running data.

[0254] In some embodiments, after the controller converts the normal operation data and the fault operation data into a graph structure, the controller converts the original node feature vector into a target node feature vector according to the graph structure, and allocates attention to the target node feature vector.

[0255] In some embodiments, after the controller converts the normal operation data and the fault operation data into a graph structure, the controller inputs the original node feature vector into the attention layer through linear transformation to obtain the target node feature vector.

[0256] In some embodiments, the attention coefficient can be assigned to the target node feature vector through the self-attention mechanism. As shown in formula (2) and formula (3): x = {x1, x1, ..., x n Formula (2) x'={x'1,x'2,...,x' n} Formula (3)

[0257] In formula (2) and formula (3), x is the original node feature vector, and x' is the target node feature vector.

[0258] It should be noted that self-attention is an important mechanism in the neural network model, which is used to perform self-attention calculation on each element in the input sequence and obtain the self-attention representation of each element.

[0259] In some embodiments, the range matrix W of the target node feature vector satisfies the relationship and the attention coefficient e ij The linear change is shown in formula (4) and formula (5): W∈R F′*F Formula (4)

[0260] In formula (4) and formula (5), R is the feature space of the target node feature vector, F' is the dimension of the target node feature vector, F is the dimension of the original node feature vector, that is, the influence coefficient of the target node feature vector i on the target node feature vector j, a represents a single-layer feedforward neural network, and the output is a numerical value.

[0261] In some embodiments, after the attention coefficient is determined, an aggregation operation is performed on the target node feature vector according to the attention coefficient.

[0262] For example, the target node feature vector V i The surrounding nodes {V a , V b,...,V n}Weighted to V i , as an aggregation of the target node feature vector.

[0263] It should be noted that, in the embodiment of the present disclosure, all target node feature vectors may be aggregated, and the aggregation method is as shown in formula (6): M i =G({w j *x j , j∈N i}) Formula (6)

[0264] In formula (6), G() is the aggregation function, w j is the weight, x j is the target node feature vector, N i is the set of node feature vectors connected to the target node feature vector.

[0265] In some embodiments, after the aggregation operation of the target node feature vector is completed, the preset graph attention model is updated through the iteration of the preset graph neural network model, and the controller determines whether the original node feature vector reaches the preset convergence condition.

[0266] In some embodiments, when the controller determines that the original node feature vectors do not meet the preset convergence conditions, the attention coefficient is updated to obtain the influence relationship between the original node feature vectors.

[0267] In some embodiments, the attention coefficient is updated as shown in formula (7): L t+1 =F w (L t ,x) Formula (7)

[0268] In formula (7), L is the original node feature vector, F w For compression mapping.

[0269] In some embodiments, when the controller determines that the original node feature vector does not meet the preset convergence condition, the training of the initial graph attention model is stopped to obtain the preset graph attention model after the training is completed.

[0270] In some embodiments, the method for determining whether the state vector of the original node feature vector x meets the convergence condition is as shown in formula (8): t -x t-1 |<ε F Formula (8)

[0271] In some embodiments, after the GAT model stops updating, the controller inputs real-time operating data of the multi-split air conditioning system into the GAT model and outputs a real-time result. The real-time result is compared with the expansion valve operating status and indoor unit serial number in the model result to confirm whether the GAT model has been successfully trained.

[0272] For example, if the operating status of the expansion valve in the real-time result matches the operating status of the expansion valve in the model result, and the indoor unit serial number in the real-time result matches the indoor unit serial number in the model result, it is confirmed that the GAT model training is successful.

[0273] For example, if the operating status of the expansion valve in the real-time result matches the operating status of the expansion valve in the model result, but the indoor unit serial number in the real-time result does not match the indoor unit serial number in the model result, the parameters of the GAT model are updated.

[0274] For example, if the operating status of the expansion valve in the real-time result does not match the operating status of the expansion valve in the model result, and the indoor unit serial number in the real-time result does not match the indoor unit serial number in the model result, the parameters of the GAT model are updated.

[0275] In some embodiments, when the GAT model stops updating, the real-time operating data of the multi-split air-conditioning system is input into the fault detection model in the GAT model, and the fault identification result is obtained after the back-propagation parameters are updated.

[0276] In some embodiments, the controller constructs maintenance information of the multi-split air-conditioning system through actual operation data and status of the multi-split air-conditioning system in multiple stages and actual operation data and status of other units, thereby continuously updating the GAT model.

[0277] FIG18 is a flowchart of another control method for a multi-split air conditioning system according to some embodiments.

[0278] In some embodiments, the control method includes steps S1 to S8.

[0279] In step S1, normal operation data and fault operation data of the multi-split air-conditioning system are obtained.

[0280] It should be noted that the contents of the normal operation data and fault operation data of the multi-split air-conditioning system are consistent with the above-mentioned real-time operation data and will not be repeated here.

[0281] For example, the normal operation data is the operation data of the multi-split air-conditioning system when the expansion valve is in a normal operation state; the fault operation data is the operation data of the multi-split air-conditioning system when the expansion valve is in a fault operation state.

[0282] In step S2 , the normal operation data and the fault operation data are input into the initial GAT model.

[0283] In some embodiments, the initial GAT model is trained by the GNN model. The training method of the GNN model for the initial GAT model is as described in step S21 above and will not be repeated here.

[0284] In some embodiments, the controller first inputs the normal operation data and the fault operation data into the fault detection model of the initial GAT model, converts the normal operation data and the fault operation data into a graph structure through the GNN model, and determines whether the expansion valve of the multi-split air-conditioning system is in a faulty operation state based on the graph structure.

[0285] For example, when the controller determines that the expansion valve of the multi-split air conditioning system is in a faulty operating state, the controller inputs the normal operating data and the faulty operating data into the fault location model of the initial GAT model to locate the internal unit serial number of the expansion valve in the faulty operating state.

[0286] In step S3, it is determined whether the result output by the initial GAT model meets the preset conditions. If so, step S5 is executed; if not, step S4 is executed.

[0287] For example, the preset condition includes: the expansion valve fault state output by the initial GAT model and the internal unit serial number of the expansion valve in the faulty operating state are both correct.

[0288] In step S4, the parameters of the initial GAT model are updated.

[0289] For example, when the expansion valve fault state output by the initial GAT model is incorrect, or the expansion valve fault state output by the GAT model is correct but the internal unit number of the expansion valve in the faulty operating state is incorrect, the controller updates the parameters of the initial GAT model.

[0290] In step S5 , it is determined that the initial GAT model is successfully trained as a preset GAT model.

[0291] In step S6, the real-time operating data of the multi-split air conditioner is obtained and input into a preset GAT model.

[0292] In some embodiments, the controller first inputs the real-time operation data into the fault detection model of the preset GAT model, converts the real-time operation data into a graph structure through the GNN model, and determines whether the expansion valve of the multi-split air-conditioning system is in a faulty operation state based on the graph structure.

[0293] For example, when the controller determines that the expansion valve of the multi-split air conditioning system is in a faulty operating state, the controller inputs real-time operating data into the fault location model of the preset GAT model to locate the internal unit serial number of the expansion valve in the faulty operating state.

[0294] In step S7, it is determined whether the result output by the initial GAT model meets the preset conditions. If so, step S8 is executed; if not, step S4 is executed.

[0295] In step S8, the preset GAT model is controlled to learn and improve according to the real-time operating data of the multi-split air conditioner.

[0296] For example, when the expansion valve fault status output by the initial GAT model is incorrect, or the expansion valve fault status output by the GAT model is correct but the indoor unit serial numbers of the expansion valves in the faulty operating state are incorrect, the controller controls the preset GAT model to learn and improve based on the real-time operating data of the multi-split air conditioner.

[0297] Those skilled in the art will understand that the scope of the present invention is not limited to the above specific embodiments, and that certain elements of the embodiments may be modified and replaced without departing from the spirit of the present invention. The scope of the present invention is limited by the appended claims.

Claims

1. A multi-split air conditioning system, comprising: outdoor unit, including the compressor and outdoor heat exchanger; The compressor is connected to the outdoor heat exchanger; at least one indoor unit, any one of the at least one indoor unit comprising an indoor heat exchanger; an expansion valve, wherein the indoor heat exchanger is connected to the outdoor heat exchanger via the expansion valve; wherein the refrigerant circulates through the compressor, the outdoor heat exchanger, the expansion valve, and the indoor heat exchanger to form a refrigerant circuit; and The controller is configured as: In the case where it is determined that a fault occurs in the multi-split air-conditioning system, a plurality of fault identification results are determined based on characteristic data of the multi-split air-conditioning system and a plurality of fault identification models based on support vector machines; wherein any one of the plurality of fault identification models is configured to identify a type of fault, and the fault identification result output by any one of the fault identification models is configured to indicate a probability of the multi-split air-conditioning system having the type of fault corresponding to any one of the fault identification models; and The fault type corresponding to the fault identification result with the highest probability among the multiple fault identification results is used as the target fault type of the multi-split air-conditioning system.

2. The multi-split air conditioning system according to claim 1, wherein: The controller is further configured to: After taking the fault type corresponding to the fault identification result with the highest probability among the multiple fault identification results as the target fault type of the multi-split air-conditioning system, determining a fault level identification result based on characteristic data of the multi-split air-conditioning system and a fault level identification model corresponding to the target fault type; and When the fault level indicated by the fault level identification result is above a preset fault level, an alarm message is issued; wherein the alarm message includes the target fault type, and the alarm message is configured to prompt the multi-split air-conditioning system to be repaired.

3. The multi-split air conditioning system according to claim 2, wherein: The controller is further configured to: Before determining that the multi-split air-conditioning system has failed, obtaining operating data of the multi-split air-conditioning system; performing a correlation analysis on the operating data of the multi-split air-conditioning system based on a maximum information coefficient method, and extracting characteristic data of the multi-split air-conditioning system from the operating data of the multi-split air-conditioning system; as well as Based on the characteristic data of the multi-split air-conditioning system, it is determined whether a fault occurs in the multi-split air-conditioning system.

4. The multi-split air conditioning system according to claim 3, wherein: The determining whether a fault occurs in the multi-split air conditioning system based on the characteristic data of the multi-split air conditioning system includes: Determining a fault diagnosis result based on the characteristic data and the fault diagnosis model of the multi-split air conditioning system; wherein the fault diagnosis result is configured to indicate whether a fault occurs in the multi-split air conditioning system; and If the fault diagnosis result is yes, it is determined that a fault occurs in the multi-split air conditioning system.

5. The multi-split air conditioning system according to claim 3, wherein: The obtaining of the operating data of the multi-split air conditioning system includes: Obtaining original operating data of the multi-split air conditioning system; and The original operating data of the multi-split air-conditioning system is preprocessed to obtain the operating data of the multi-split air-conditioning system; wherein the preprocessing includes outlier elimination processing and smoothing processing.

6. The multi-split air conditioning system according to any one of claims 1 to 5, further comprising a liquid pipe and a gas pipe, wherein the outdoor unit is connected to the at least one indoor unit via the liquid pipe and the gas pipe; in, The characteristic data of the multi-split air-conditioning system include: at least one of the operating frequency of the compressor, the exhaust pressure value of the compressor, the suction pressure value of the compressor, the suction temperature value of the compressor, the exhaust temperature value of the compressor, the exhaust superheat of the compressor, the suction superheat of the compressor, the gear position of the outdoor fan, the opening of the expansion valve, the outlet air temperature of the indoor unit, the return air temperature of the indoor unit, the temperature value of the gas pipe and the temperature value of the liquid pipe.

7. A control method for a multi-split air conditioning system, wherein: The multi-split air conditioning system includes an outdoor unit, at least one indoor unit, and an expansion valve; the outdoor unit includes a compressor and an outdoor heat exchanger; any one of the at least one indoor unit includes an indoor heat exchanger; wherein a refrigerant circulates through the compressor, the outdoor heat exchanger, the expansion valve, and the indoor heat exchanger to form a refrigerant circuit; The method comprises: In a case where it is determined that a fault occurs in the multi-split air-conditioning system, a plurality of fault identification results are determined based on characteristic data of the multi-split air-conditioning system and a plurality of fault identification models based on support vector machines; wherein any one of the plurality of fault identification models is configured to identify a fault type, and the fault identification result output by any one of the fault identification models is configured to indicate a probability of the multi-split air-conditioning system having the fault type corresponding to any one of the fault identification models; and The fault type corresponding to the fault identification result with the highest probability among the multiple fault identification results is used as the target fault type of the multi-split air-conditioning system.

8. The method according to claim 7, wherein: The method further comprises: After taking the fault type corresponding to the fault identification result with the highest probability among the multiple fault identification results as the target fault type of the multi-split air-conditioning system, determining a fault level identification result based on characteristic data of the multi-split air-conditioning system and a fault level identification model corresponding to the target fault type; and When the fault level indicated by the fault level identification result is above a preset fault level, an alarm message is issued; wherein the alarm message includes the target fault type, and the alarm message is configured to prompt the multi-split air-conditioning system to be repaired.

9. The method according to claim 8, wherein The method further comprises: Before determining that the multi-split air-conditioning system has failed, obtaining operating data of the multi-split air-conditioning system; performing a correlation analysis on the operating data of the multi-split air-conditioning system based on a maximum information coefficient method, and extracting characteristic data of the multi-split air-conditioning system from the operating data of the multi-split air-conditioning system; and Based on the characteristic data of the multi-split air-conditioning system, it is determined whether a fault occurs in the multi-split air-conditioning system.

10. The method according to claim 9, wherein: The determining whether a fault occurs in the multi-split air conditioning system based on the characteristic data of the multi-split air conditioning system includes: Determining a fault diagnosis result based on the characteristic data and the fault diagnosis model of the multi-split air conditioning system; wherein the fault diagnosis result is configured to indicate whether a fault occurs in the multi-split air conditioning system; and If the fault diagnosis result is yes, it is determined that a fault occurs in the multi-split air conditioning system.

11. The method according to claim 9, wherein The obtaining of the operating data of the multi-split air conditioning system includes: Obtaining original operating data of the multi-split air conditioning system; and The original operating data of the multi-split air-conditioning system is preprocessed to obtain the operating data of the multi-split air-conditioning system; wherein the preprocessing includes outlier elimination processing and smoothing processing.

12. A multi-split air conditioning system comprising: Outdoor unit; at least one indoor unit connected to the outdoor unit; at least one expansion valve corresponding to the at least one indoor unit; Any one of the at least one expansion valve is disposed on a pipeline between the corresponding indoor unit and the outdoor unit; as well as The controller is configured as: Acquiring real-time operating data of the multi-split air conditioning system; Performing fault identification on the real-time operation data using a preset graph attention model to obtain a fault identification result; wherein the fault identification result includes whether any expansion valve in the multi-split air-conditioning system is in a faulty operation state or a normal operation state; as well as In the case that any one of the expansion valves is in a faulty operating state, the fault location of the any one of the expansion valves is performed using the preset graph attention model to determine the internal unit serial number of the any one of the expansion valves.

13. The multi-split air conditioning system according to claim 12, wherein: The controller is further configured to: Acquiring normal operation data and fault operation data of the multi-split air-conditioning system; wherein the normal operation data is the operation data of the multi-split air-conditioning system when the expansion valve is in a normal operation state; the fault operation data is the operation data of the multi-split air-conditioning system when the expansion valve is in a fault operation state; and The initial graph attention model is trained according to the normal operation data and the fault operation data to obtain a preset graph attention model after training.

14. The multi-split air conditioning system according to claim 13, wherein: The controller is further configured to: Before training the initial graph attention model based on the normal operation data and the fault operation data, the normal operation data and the fault operation data are converted into a graph structure through a preset graph neural network model; wherein, the graph structure includes multiple original node feature vectors and the connection relationship between the multiple original node feature vectors; the original node feature vector is a vector representation of the normal operation data and the fault operation data.

15. The multi-split air conditioning system according to claim 14, wherein: The controller is further configured to: calculate a Pearson correlation coefficient for different data information between two original node feature vectors among the plurality of original node feature vectors; Wherein, when the Pearson correlation coefficient is greater than the preset threshold, the two original node feature vectors are connected by an edge and the two original node feature vectors are correlated; when the Pearson correlation coefficient is less than the preset threshold, the two The original node feature vectors are not connected by edges, and the two original node feature vectors are not correlated.

16. The multi-split air conditioning system according to claim 14, wherein: The training of the initial graph attention model according to the normal operation data and the fault operation data includes: Converting the original node feature vector into a target node feature vector and assigning an attention coefficient to the target node feature vector; Performing an aggregation operation on the target node feature vector according to the attention coefficient; By iterating the preset graph neural network model, the preset graph attention model is updated to determine whether the original node feature vector meets the preset convergence condition; If the original node feature vector does not meet the preset convergence condition, updating the attention coefficient; and When the original node feature vector reaches the preset convergence condition, the training of the initial graph attention model is stopped to obtain the preset graph attention model after training.

17. The multi-split air conditioning system according to any one of claims 12 to 16, wherein: The real-time operation data includes the real-time operation data of the outdoor unit and the real-time operation data of the indoor unit; The real-time operating data of the outdoor unit includes compressor frequency, pressure value, top pressure temperature value, exhaust superheat value, defrost temperature value and ambient temperature value; The real-time operation data of the indoor unit includes a variable frequency heat dissipation value, a liquid pipe temperature value, a gas pipe temperature value, a return air temperature value, and an outlet air temperature value.

18. A control method for a multi-split air conditioning system, wherein: The multi-split air conditioning system includes an outdoor unit, at least one indoor unit and at least one expansion valve; the at least one expansion valve corresponds to the at least one indoor unit; Any one of the at least one expansion valve is disposed on a pipeline between the corresponding indoor unit and the outdoor unit; The method comprises: Acquiring real-time operating data of the multi-split air conditioning system; Performing fault identification on the real-time operation data through a preset graph attention model to obtain a fault identification result; wherein the fault identification result includes whether any expansion valve in the multi-split air-conditioning system is in a faulty operation state or a normal operation state; and In the case that any one of the expansion valves is in a faulty operating state, the fault location of the any one of the expansion valves is performed using the preset graph attention model to determine the internal unit serial number of the any one of the expansion valves.

19. The method according to claim 18, further comprising: Acquiring normal operation data and fault operation data of the multi-split air-conditioning system; wherein the normal operation data is the operation data of the multi-split air-conditioning system when the expansion valve is in a normal operation state; the fault operation data is the operation data of the multi-split air-conditioning system when the expansion valve is in a fault operation state; and The initial graph attention model is trained according to the normal operation data and the fault operation data to obtain a preset graph attention model after training.

20. The method according to claim 19, wherein The training of the initial graph attention model according to the normal operation data and the fault operation data includes: Converting the original node feature vector into a target node feature vector and assigning an attention coefficient to the target node feature vector; Performing an aggregation operation on the target node feature vector according to the attention coefficient; By iterating the preset graph neural network model, the preset graph attention model is updated to determine whether the original node feature vector meets the preset convergence condition; If the original node feature vector does not meet the preset convergence condition, updating the attention coefficient; and When the original node feature vector reaches the preset convergence condition, the training of the initial graph attention model is stopped to obtain the preset graph attention model after training.

Images