First multi-split air-conditioning unit, cloud servers, and control method for cloud server

Abstract

Provided are a first multi-split air-conditioning unit, cloud servers, and a control method for a cloud server. A cloud server comprises a communicator and a processor. The processor is configured to: acquire operating data of an air conditioner; combine operating data of at least one indoor unit with operating data of at least one outdoor unit to obtain first combined data in a first dimension and second combined data in a second dimension; input the first combined data into a first fault detection model to obtain first fault information of the air conditioner in the first dimension, and input the second combined data into a second fault detection model to obtain second fault information of the air conditioner in the second dimension; and on the basis of the first fault information and the second fault information, determine a fault detection result of the air conditioner.

Description

First multi-split air conditioning unit, cloud server and control method thereof

[0001] This application claims priority to Chinese patent application No. 202410308836.1 filed on March 18, 2024; and priority to Chinese patent application No. 202410096675.4 filed on January 23, 2024, the entire contents of which are incorporated by reference into this application. Technical Field

[0002] The present disclosure relates to the technical field of air conditioning, and in particular to a first multi-split air conditioning unit, a cloud server and a control method thereof. Background Art

[0003] With the development of the economy and society, air conditioners are increasingly used in various places such as entertainment, home, and work. An air conditioner may include at least one indoor unit and at least one outdoor unit. The at least one indoor unit and at least one outdoor unit are connected to perform the air conditioning function of the air conditioner.

[0004] Summary of the Invention

[0005] In one aspect, a cloud server is provided. The cloud server includes a communicator and a processor. The communicator is configured to establish a communication connection with an air conditioner. The air conditioner includes at least one indoor unit and at least one outdoor unit. The processor is communicatively connected to the communicator and configured to: obtain operating data of the air conditioner; the operating data of the air conditioner includes operating data of the at least one indoor unit and operating data of the at least one outdoor unit; combine the operating data of the at least one indoor unit with the operating data of the at least one outdoor unit to obtain first combined data of a first dimension and second combined data of a second dimension; input the first combined data into a first fault detection model to obtain first fault information of the air conditioner in the first dimension, and input the second combined data into a second fault detection model to obtain second fault information of the air conditioner in the second dimension; and determine a fault detection result of the air conditioner based on the first fault information and the second fault information.

[0006] On the other hand, a control method for a cloud server is provided. The method is applied to the cloud server. The cloud server includes a communicator and a processor. The communicator is configured to establish a communication connection with an air conditioner. The air conditioner includes at least one indoor unit and at least one outdoor unit. The processor is communicatively connected to the communicator. The method includes: obtaining operating data of the air conditioner; the operating data of the air conditioner includes operating data of at least one indoor unit and operating data of at least one outdoor unit; combining the operating data of the at least one indoor unit with the operating data of the at least one outdoor unit to obtain first combined data of a first dimension and second combined data of a second dimension; inputting the first combined data into a first fault detection model to obtain first fault information of the air conditioner in the first dimension, and inputting the second combined data into a second fault detection model to obtain second fault information of the air conditioner in the second dimension; and determining a fault detection result of the air conditioner based on the first fault information and the second fault information.

[0007] In another aspect, a cloud server is provided. The cloud server includes a communicator and a processor. The communicator is configured to communicate with a first multi-split air conditioning unit and a second multi-split air conditioning unit. The processor is connected to the communicator and is configured to: obtain a first sample set and a second sample set; the first sample set includes multiple first samples, and the multiple first samples are generated according to the first operating data of the first multi-split air-conditioning unit; the second sample set includes multiple second samples, and the multiple second samples are generated according to the first operating data and the second operating data of the second multi-split air-conditioning unit; any second sample of the multiple second samples has a first label or a second label; the first label is configured to indicate the operating mode corresponding to the second sample; the second label is configured to indicate the fault type corresponding to the second sample; when the proportion of the target sample in the second sample set is less than or equal to a preset threshold, the second sample set is expanded to adjust the proportion of the target sample in the second sample set; the target sample is the second sample with the first label or the second label; the first sample set is adjusted according to the feature space distribution of the second sample set; the initial model is trained according to the first sample set and the second sample set to obtain a trained fault diagnosis model; and the fault diagnosis model is used to diagnose the first multi-split air-conditioning unit.

[0008] In some embodiments, the processor is further configured to: obtain at least one neighboring sample of the target sample; perform random linear interpolation processing between the target sample and any one of the at least one neighboring samples to generate a new target sample; and add the new target sample to the second sample set to adjust the proportion of the target sample in the second sample set.

[0009] In some embodiments, the processor is further configured to: input the first sample set and the second sample set into a domain adaptation model to adjust the first sample set; the domain adaptation model adopts an adversarial neural network structure.

[0010] In some embodiments, the processor is further configured to: perform an extraction operation; the extraction operation includes the processor extracting a first feature from the first sample through a feature extractor, extracting a second feature from a second sample generated based on the first operating data of the second multi-split air-conditioning unit, and extracting a third feature from a second sample generated based on the second operating data of the second multi-split air-conditioning unit; perform an input operation; the input operation includes the processor inputting the first feature and the second feature into a domain classifier to obtain a domain loss function, and inputting the third feature into a label classifier to obtain a label loss function; perform an update operation; the update operation includes the processor updating the parameters of the feature extractor, the domain classifier and the label classifier according to the domain loss function and the label loss function through a back propagation algorithm; repeatedly perform the input operation and the update operation until a convergence condition is reached; the convergence condition includes at least one of the following: the domain classification loss function value of the domain classifier is greater than or equal to a first threshold, and the label loss function value of the label classifier is less than or equal to a second threshold; or, the number of iterations reaches a preset number.

[0011] In some embodiments, the processor is further configured to: obtain first test data of the first multi-split air-conditioning unit and second test data of the second multi-split air-conditioning unit; the first test data includes first operating data of the first multi-split air-conditioning unit, and the second test data includes first operating data and second operating data of the second multi-split air-conditioning unit; preprocess the first test data and the second test data to obtain preprocessed first test data and second test data; the preprocessing includes at least one of data cleaning, data reduction, data encoding and data normalization; and generate the first sample set and the second sample set based on the preprocessed first test data and the second test data.

[0012] In some embodiments, the first test data includes at least one of the following: the operating conditions, load rate, set temperature, and air volume level of the first multi-split air-conditioning unit under different test environments; the second test data includes at least one of the following: the operating conditions, load rate, set temperature, and air volume level of the second multi-split air-conditioning unit under different test environments.

[0013] In another aspect, a first multi-split air conditioning unit is provided. The first multi-split air conditioning unit includes at least one indoor unit, at least one outdoor unit, and a controller. The controller is configured to: obtain a first sample set and a second sample set; the first sample set includes multiple first samples, the multiple first samples generated based on first operating data of the first multi-split air conditioning unit; the second sample set includes multiple second samples, the multiple second samples generated based on first operating data and second operating data of the second multi-split air conditioning unit; any second sample in the multiple second samples has a first label or a second label; the first label is configured to indicate the operating mode corresponding to the second sample; the second label is configured to indicate the fault type corresponding to the second sample; if the proportion of target samples in the second sample set is less than or equal to a preset threshold, expand the second sample set to adjust the proportion of the target samples in the second sample set; the target samples are the second samples with the first label or the second label; adjust the first sample set based on the feature space distribution of the second sample set; train an initial model based on the first and second sample sets to obtain a trained fault diagnosis model; and use the fault diagnosis model to perform fault diagnosis on the first multi-split air conditioning unit.

[0014] In some embodiments, the controller is further configured to: obtain at least one neighboring sample of the target sample; perform random linear interpolation processing between the target sample and any one of the at least one neighboring samples to generate newly added samples with the first label or the second label; and add the newly added samples with the first label or the second label to the second sample set to adjust the proportion of the target sample in the second sample set.

[0015] In some embodiments, the controller is further configured to: input the first sample set and the second sample set into a domain adaptation model to adjust the first sample set; the domain adaptation model adopts an adversarial neural network structure.

[0016] In some embodiments, the controller is further configured to: perform an extraction operation; the extraction operation includes the controller extracting a first feature from the first sample through a feature extractor, extracting a second feature from a second sample generated based on the first operating data of the second multi-split air-conditioning unit, and extracting a third feature from a second sample generated based on the second operating data of the second multi-split air-conditioning unit; perform an input operation; the input operation includes the controller inputting the first feature and the second feature into a domain classifier to obtain a domain loss function, and inputting the third feature into a label classifier to obtain a label loss function; perform an update operation: the update operation includes the controller updating the parameters of the feature extractor, the domain classifier and the label classifier according to the domain loss function and the label loss function through a back propagation algorithm; repeatedly perform the input operation and the update operation until the domain classification loss function value of the domain classifier is maximized and the label loss function value of the label classifier is minimized. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] FIG1 is a schematic diagram of an air conditioner and a cloud server according to some embodiments;

[0018] FIG2 is a schematic diagram of an air conditioner, a cloud server, and a user terminal according to some embodiments;

[0019] FIG3 is a structural diagram of an air conditioner according to some embodiments;

[0020] FIG4 is a block diagram of a cloud server according to some embodiments;

[0021] FIG5 is a flow chart of steps performed by a cloud server according to some embodiments;

[0022] FIG6 is a schematic diagram of first combined data according to some embodiments;

[0023] FIG7 is a schematic diagram of second combined data according to some embodiments;

[0024] FIG8 is another flow chart of steps performed by a cloud server according to some embodiments;

[0025] FIG9 is another flow chart of steps executed by a cloud server according to some embodiments;

[0026] FIG10 is another flow chart of steps performed by a cloud server according to some embodiments;

[0027] FIG11 is another flow chart of steps performed by a cloud server according to some embodiments;

[0028] FIG12 is a block diagram of a method for controlling a cloud server according to some embodiments;

[0029] FIG13 is a schematic diagram of a multi-split air conditioning unit and a cloud server according to some embodiments;

[0030] FIG14 is a schematic diagram of a multi-split air conditioning unit according to some embodiments;

[0031] FIG15A is a block diagram of a multi-split air conditioning unit according to some embodiments;

[0032] FIG15B is another block diagram of a multi-split air conditioning unit according to some embodiments;

[0033] FIG16 is another flow chart of steps performed by a cloud server according to some embodiments;

[0034] FIG17 is another flow chart of steps performed by a cloud server according to some embodiments;

[0035] FIG18 is another block diagram of a cloud server according to some embodiments;

[0036] FIG19 is another flow chart of steps performed by a cloud server according to some embodiments;

[0037] FIG20 is a block diagram of an adversarial neural network according to some embodiments;

[0038] FIG21 is another flow chart of steps performed by a cloud server according to some embodiments;

[0039] FIG22 is a block diagram of a fault diagnosis apparatus according to some embodiments;

[0040] FIG23 is a flow chart of a diagnostic method for a multi-split air conditioning unit according to some embodiments; DETAILED DESCRIPTION

[0041] The following will clearly and completely describe some embodiments of the present disclosure in conjunction with the accompanying drawings. However, the described embodiments 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.

[0042] 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.

[0043] In the following, the terms "first" and "second" are used for descriptive purposes only and should not be understood to indicate or imply relative importance or implicitly specify 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 the embodiments of the present disclosure, unless otherwise specified, "plurality" means two or more.

[0044] When describing some embodiments, the word "connected" and its derivatives may be used. The term "connected" should be understood broadly. For example, "connected" can mean fixed, removable, or integrated; it can be directly connected or indirectly connected through an intermediary. The embodiments disclosed herein are not necessarily limited to the contents of this document.

[0045] “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.

[0046] 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.

[0047] 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).

[0048] Generally, as air conditioner utilization increases, the probability of failure also increases. In some solutions, a fault diagnosis model equivalent to that used for the indoor unit can be used to diagnose indoor unit failures, and a fault diagnosis model equivalent to that used for the outdoor unit can be used to diagnose outdoor unit failures. However, with the advancement of air conditioning technology, air conditioners have become increasingly complex. For example, air conditioners include multiple indoor and outdoor units, and the various components interact with each other. This makes it difficult for fault diagnosis models to accurately determine the specific location and type of air conditioner failures, affecting the accuracy of fault diagnosis.

[0049] In order to solve the above problems, some embodiments of the present disclosure provide a cloud server 103. The cloud server 103 divides the operating data combination of the air conditioner 1001 into first combination data of a first dimension (indoor unit dimension) and second combination data of a second dimension (air conditioner dimension), and inputs the first combination data into a first fault diagnosis model to obtain first fault information; and inputs the second combination data into a second fault diagnosis model to obtain second fault information. The cloud server 103 obtains a diagnosis result of the air conditioner 1001 based on the first fault information and the second fault information. In this way, the fault of the air conditioner 1001 is considered from both the dimension of the indoor unit 301 and the overall dimension of the air conditioner 1001, so that the source of the fault can be accurately located and identified, thereby improving the accuracy of fault diagnosis.

[0050] As shown in Figures 1 and 2, the application scenario includes at least one air conditioner 1001 and a cloud server 103. The at least one air conditioner 1001 and the cloud server 103 can be communicatively connected.

[0051] In some embodiments, the at least one air conditioner 1001 includes one air conditioner 1001 , and the at least one air conditioner 1001 may also include multiple air conditioners 1001 .

[0052] In some embodiments, the air conditioner 1001 is a device that regulates and controls parameters such as the temperature, humidity, and flow rate of the ambient air within a building or structure. In some embodiments, the cloud server 103 can provide at least one basic cloud computing service selected from the group consisting of cloud service, cloud database, cloud computing, cloud storage, network service, cloud communications, middleware service, domain name service, security service, content delivery network, and big data server.

[0053] In some embodiments, as shown in FIG2 , the cloud server 103 may obtain operating data of the air conditioner 1001 , such as temperature data, pressure data, valve opening data, current data, frequency data, air volume level data, and the like.

[0054] In some embodiments, the cloud server 103 may store the operating data of the air conditioner 1001 collected in real time in a database.

[0055] In some embodiments, the cloud server 103 may obtain operating data of multiple groups of air conditioners 1001 from a database, and perform data processing operations on the operating data of the air conditioners 1001 .

[0056] In some embodiments, as shown in FIG2 , the application scenario may further include a client 304. The cloud server 103 and the client 304 may be in communication connection.

[0057] In some embodiments, the client (also referred to as the user end) 304 refers to a program corresponding to the server that provides local services to customers. The client is usually installed on a terminal device, such as a mobile phone, tablet computer, desktop, laptop, handheld computer, notebook computer, ultra-mobile personal computer (Ultra-Mobile Personal Computer, UMPC), netbook, as well as a cellular phone, personal digital assistant (Personal Digital Assistant, PDA), augmented reality (Augmented Reality, AR)\virtual reality (Virtual Reality, VR) device, etc. In some embodiments, after determining that the air conditioner 1001 has a fault, the cloud server 103 can send the fault result to the client 304, so that the client 304 can visualize the fault result to remind the user that the air conditioner 1001 has a fault.

[0058] In some embodiments, as shown in FIG3 , the air conditioner 1001 further includes at least one indoor unit 301 ; the air conditioner 1001 further includes at least one outdoor unit 302 .

[0059] The air conditioner 1001 can be a wall-mounted air conditioner or a cabinet-mounted air conditioner. In the case of the wall-mounted air conditioner 1001, the indoor unit 301 can be hung on the indoor wall; in the case of the cabinet-mounted air conditioner 1001, the indoor unit 301 can be installed on the ground.

[0060] The outdoor unit 302 is usually installed outdoors and is configured to exchange heat with the indoor environment. In addition, in FIG2 , since the outdoor unit 302 is located outdoors on the opposite side of the indoor unit 301 across a wall, the outdoor unit 302 can be represented by a dotted line.

[0061] In some embodiments, as shown in FIG4 , the cloud server 103 includes a processor 203. The processor 203 may be a central processing unit (CPU), an application specific integrated circuit (ASIC), or one or more integrated circuits configured to implement some embodiments of the present disclosure. For example, the processor 203 may be one or more digital signal processors (DSPs) or one or more field programmable gate arrays (FPGAs).

[0062] In some embodiments, the cloud server 103 also includes a memory 202. The memory 202 can be used to store software programs and data, such as the operating data of the air conditioner 1001. The processor 203 executes various functions and data processing of the cloud server 103 by running the software programs or data stored in the memory 202. The memory 202 may include a high-speed random access memory, and may also include a non-volatile memory, such as at least one disk storage device, a flash memory device, or other volatile solid-state storage device. The memory 202 stores an operating system that enables the cloud server 103 to run. The memory 202 can store the operating system and various application programs, and may also store code for executing the control method of the cloud server 103 provided in some embodiments of the present disclosure.

[0063] In some embodiments, the cloud server 103 further includes a communicator 201. The communicator 201 is used to establish a communication connection with other network entities, for example, to establish a communication connection with the air conditioner 1001 and the client 304. The communicator 201 includes a radio frequency (RF) device, a cellular device, a wireless fidelity (WIFI) module, and a global positioning system (GPS) device. Taking the communicator 201 as an example, the RF device can be used to receive and send signals. For example, the RF device sends the received information to the processor 203 for processing; in addition, it sends the signal generated by the processor 203. The RF circuit may include an antenna (Aerial), at least one amplifier (Amplifier), a transceiver (Transceiver), a coupler (Coupler), a low noise amplifier (LNA), a duplexer (Duplexer), etc.

[0064] It should be noted that the hardware structure shown in FIG4 does not constitute a limitation on the cloud server 103. The cloud server 103 may include more or fewer components than those shown in FIG4 (e.g., the communicator 201, the memory 202, and the processor 203), or may combine certain components or arrange the components differently. The following describes how the cloud server 103 diagnoses a fault of the air conditioner 1001 based on the operating data of the air conditioner 1001. In some embodiments, as shown in FIG5, the cloud server 103 is configured to perform steps 601 to 604.

[0065] In step 601 , the operating data of the air conditioner 1001 is acquired.

[0066] The operating data of the air conditioner 1001 includes operating data of at least one indoor unit 301 and operating data of at least one outdoor unit 302 .

[0067] In some embodiments, the cloud server 103 can collect the operating data of the air conditioner 1001 in real time, and after collecting the operating data of the air conditioner 1001, store the collected operating data of the air conditioner 1001 in a database, so that when fault detection is performed based on the operating data of the air conditioner 1001, the cloud server 103 can easily obtain the operating data of the air conditioner 1001 from the database.

[0068] It is understood that storing the operating data of the air conditioner 1001 in a database can achieve centralized management and storage, making it easier to access, query, and analyze the operating data. In addition, the database has the characteristic of persistent storage, so that the operating data of the air conditioner 1001 stored in the database can avoid being lost due to restarting or power failure of the cloud server 103. This helps the cloud server 103 perform tasks such as fault detection of the air conditioner 1001.

[0069] In some embodiments, the operating data of the air conditioner 1001 (i.e., the operating data of any one of the at least one indoor unit 301, or the operating data of any one of the at least one outdoor unit 302) may include temperature data, pressure data, valve opening data, frequency data, air volume level data, current data, etc.

[0070] Temperature data refers to the temperature data measured during the operation of the air conditioner 1001, such as the indoor temperature, the air supply temperature, etc. Abnormal temperature data may be caused by a malfunction of the air conditioner 1001 or other problems.

[0071] Pressure data is a series of static and dynamic pressures generated during the operation of the air conditioner 1001, such as condensing pressure, evaporating pressure, exhaust pressure, suction pressure, etc. Abnormal pressure data may indicate that the air conditioner 1001 has a problem of excessively high or low pressure, which may be caused by refrigerant leakage, valve failure, or other system problems.

[0072] Valve opening data refers to the opening of the valves that regulate refrigerant flow in air conditioner 1001, for example, the opening of the electronic expansion valve in indoor unit 301 and the opening of the electronic expansion valve in outdoor unit 302. By monitoring valve opening data, we can understand the refrigerant flow of air conditioner 1001 under different operating conditions and help determine whether the valves are functioning properly. Abnormal valve opening values ​​may indicate valve blockage, damage, or control signal problems, which may affect the cooling or heating performance of air conditioner 1001.

[0073] Frequency data represents the operating frequency of components such as compressor 401 and the fan in air conditioner 1001. For example, the operating frequency of compressor 401 and the fan can be monitored to understand changes in the operating status of components such as compressor 401. Abnormal frequency data may indicate a fault in a component such as compressor 401, such as abnormal voltage, excessive current, or a problem with the mechanical components of compressor 401.

[0074] Air volume level data is fundamental data reflecting the supply and return air conditions of the air supply system of air conditioner 1001. For example, it includes the supply air volume level of indoor unit 301 and the return air volume level of the indoor unit. This data can be used to determine whether the air supply system of air conditioner 1001 is functioning properly and whether indoor air circulation is adequate. Abnormal air volume level data may be caused by a fan failure, a clogged air duct, or a problem with a regulating valve, which can affect the cooling or heating performance of air conditioner 1001.

[0075] Current data refers to the current consumed by various components of air conditioner 1001 during operation, such as the current of compressor 401 and the fan. By monitoring this current data, the operating status of components in air conditioner 1001, including the power consumption of compressor 401 and fans, can be assessed. Abnormal current values ​​may indicate an electrical fault or mechanical load issue with a component in air conditioner 1001.

[0076] In step 602, the operating data of at least one indoor unit 301 is combined with the operating data of at least one outdoor unit 302 to obtain first combined data of a first dimension and second combined data of a second dimension.

[0077] The first combined data of the first dimension can reflect the operating status of the indoor unit 301, excluding the operating data of the outdoor unit and other components. Alternatively, the first combined data of the first dimension can also include the operating data of the outdoor unit and other components. The second combined data of the second dimension can reflect the overall operating status of the air conditioner 1001. For example, the second combined data can reflect the operating status of the indoor unit 301, the outdoor unit 302, and the related pipes, circuits, and control components.

[0078] By combining and distinguishing the operating data of the air conditioner 1001 into first combined data and second combined data, the specific source of a fault in the air conditioner 1001 can be accurately determined. The first combined data can help identify and detect faults related to the indoor unit 301 alone, such as a blockage in the indoor unit 301 or a fault in the expansion valve of the indoor unit 301. The second combined data can provide comprehensive fault information, helping to identify and detect faults in the entire air conditioner 1001, such as a refrigerant leak or a compressor 401 fault.

[0079] In some embodiments, the cloud server 103 may combine the operating data of any one of the at least one indoor unit 301 with the operating data of all of the at least one outdoor unit 302 to obtain first combined data. The first combined data includes data corresponding to each of the at least one indoor unit 301.

[0080] For example, as shown in Figure 6, an air conditioner 1001 includes four indoor units 301, namely a first indoor unit 301A, a second indoor unit 301B, a third indoor unit 301C, and a fourth indoor unit 301D. The air conditioner 1001 also includes two outdoor units 302, namely a first outdoor unit 302A and a second outdoor unit 302B. In this case, the cloud server 103 can obtain four first combination data, namely, first sub-combination data, second sub-combination data, third sub-combination data, and fourth sub-combination data. The first sub-combination data is composed of the first indoor unit 301A, the first outdoor unit 302A, and the second outdoor unit 302B. The second sub-combination data is composed of the second indoor unit 301B, the first outdoor unit 302A, and the second outdoor unit 302B. The third sub-combination data is composed of the third indoor unit 301C, the first outdoor unit 302A, and the second outdoor unit 302B. The fourth sub-combination data is composed of the fourth indoor unit 301D, the first outdoor unit 302A, and the second outdoor unit 302B.

[0081] In some embodiments, the cloud server 103 may combine the operating data of all indoor units 301 in the at least one indoor unit 301 with the operating data of all outdoor units 302 in the at least one outdoor unit 302 to obtain second combined data.

[0082] For example, as shown in FIG7 , when the air conditioner 1001 includes four indoor units 301 and two outdoor units 302, the cloud server 103 can also obtain a second combination data. The second combination data is the second combination data formed by combining the operating data of the first indoor unit 301A, the operating data of the second indoor unit 301B, the operating data of the third indoor unit 301C, and the operating data of the fourth indoor unit 301D with the operating data of the first outdoor unit 302A and the operating data of the second outdoor unit 302B.

[0083] In step 603, the first combined data is input into the first fault diagnosis model to obtain the first fault information of the air conditioner in the first dimension, and the second combined data is input into the second fault diagnosis model to obtain the second fault information of the air conditioner 1001 in the second dimension.

[0084] The first fault information and the second fault information are respectively used to indicate the fault condition of the air conditioner 1001. The first fault information and the second fault information may respectively include the first information (fault information) or the second information (normal information).

[0085] It is understood that a failure of the air conditioner 1001 in the second dimension refers to a failure occurring in the entire air conditioner 1001, including the indoor unit 301, the outdoor unit 302, and the associated piping, circuits, and control components. The second dimension considers the overall operation and failure of the air conditioner 1001. A failure of the air conditioner 1001 in the first dimension refers to a failure occurring only in the indoor unit 301, excluding the outdoor unit 302 and other components of the air conditioner 1001. The first dimension focuses on the operation and failure of the indoor unit 301. By performing fault diagnosis in both the first and second dimensions, the specific source of the air conditioner 1001 failure can be accurately determined, thereby improving the accuracy of fault diagnosis for the air conditioner 1001.

[0086] In step 604, a fault diagnosis result of the air conditioner 1001 is determined based on the first fault information and the second fault information.

[0087] In some embodiments, the cloud server 103 may determine the fault diagnosis result of the air conditioner 1001 based on a preset correspondence between the first fault information, the second fault information, and the fault diagnosis result of the air conditioner 1001 .

[0088] For example, for a fault of the air conditioner 1001 in the second dimension, the preset correspondence between the first fault information, the second fault information and the fault diagnosis result of the air conditioner 1001 can be as shown in Table 1.

[0089] Table 1

[0090] According to the corresponding relationship described in Table 1, for a fault of the air conditioner 1001 in the second dimension, such as a refrigerant fault, an outdoor unit blockage fault, an outdoor unit expansion valve fault, a compressor fault, etc., the outdoor unit blockage can refer to a blockage in a pipeline or an air duct. In the case where the second fault information is fault information, the cloud server 103 determines the fault result as the fault diagnosis result of the air conditioner, thereby determining that the air conditioner 1001 has a fault in the second dimension, such as a refrigerant fault, an outdoor unit blockage fault, an outdoor unit expansion valve fault, a compressor fault, etc.

[0091] For example, for a fault of the air conditioner 1001 in the first dimension, the preset correspondence between the first fault information, the second fault information and the fault diagnosis result of the air conditioner can be as shown in Table 2.

[0092] Table 2

[0093] According to the correspondence described in Table 2, for a fault of the air conditioner 1001 in the first dimension, such as an indoor unit blockage fault or an indoor unit expansion valve fault, if at least one of the first fault information and the second fault information is fault information, the cloud server 103 determines the fault result as a fault detection result of the air conditioner 1001, i.e., determines that the air conditioner 1001 has a fault in the first dimension, such as the indoor unit blockage fault or indoor unit expansion valve fault. In some embodiments, if the first fault information is the second information and the second fault information is the first information, the diagnosis result of the air conditioner 1001 is normal.

[0094] In some embodiments of the present disclosure, the cloud server 103 divides the operating data of the air conditioner 1001 into combined data of different dimensions, such as first combined data of a first dimension and second combined data of a second dimension, thereby comprehensively and accurately detecting faults in the air conditioner 1001 from different dimensions. For example, the first combined data can be input into a first fault detection model for detecting faults in the air conditioner 1001 in the first dimension to accurately detect faults in the indoor unit 301. The second combined data can also be input into detecting faults in the air conditioner 1001 in the overall dimension of the air conditioner 1001 to improve the accuracy of detecting faults in the air conditioner 1001 as a whole. Furthermore, the cloud server 103 can obtain a target diagnostic result for the air conditioner 1001 based on the first fault information and the second fault information. In this way, this approach considers faults in the air conditioner 1001 from both the indoor unit dimension and the overall dimension, allowing for precise location and identification of the fault source, thereby improving the accuracy of fault diagnosis. In some embodiments, after step 601 and before step 602, as shown in Figure 8, the cloud server 103 is further configured to perform steps 605 and 606. In step 605, the operating data of the air conditioner 1001 is processed to extract non-steady-state operating data from the operating data of the air conditioner 1001.

[0095] In some embodiments, the cloud server 103 may extract non-steady-state operation data based on a sliding window method.

[0096] In some embodiments, as shown in FIG9 , step 605 includes step 6051 and step 6052 .

[0097] In step 6051, the operating data of the air conditioner 1001 is periodically diagnosed at a preset time.

[0098] It should be noted that a diagnostic cycle includes operating data within a preset time period.

[0099] In some embodiments, the preset duration may be pre-set or determined according to a diagnosis cycle.

[0100] In some embodiments, the preset duration can be obtained by the following formula (1): D = [ec(td)] Formula (1)

[0101] D is the preset duration; c and d are constant coefficients; t is the diagnosis cycle, and e is a natural constant.

[0102] In step 6052, for any diagnosis period, when the standard deviation of the operation data in the diagnosis period is within a preset range, the operation data in the diagnosis period is determined to be non-steady-state operation data.

[0103] For example, the preset range may be a numerical range greater than a preset threshold.

[0104] In some embodiments, for any diagnostic cycle, if the standard deviation of the operating data in the diagnostic cycle is outside the preset range, the cloud server 103 determines that the operating data in the diagnostic cycle is steady-state operating data.

[0105] In step 606 , non-steady-state operation data is eliminated.

[0106] It is understandable that as the operating conditions of the air conditioner 1001 change, such as load changes, the operating data of the air conditioner 1001 may change, making the operating data of the air conditioner 1001 unstable or abnormal, thereby failing to accurately reflect the fault condition of the air conditioner 1001 and affecting the accuracy of fault diagnosis of the air conditioner 1001. Therefore, it is necessary to eliminate unstable operating data (i.e., the aforementioned non-steady-state operating data) from the operating data of the air conditioner 1001.

[0107] In some embodiments, the cloud server 103 may delete the non-steady-state operation data stored in the database.

[0108] In some embodiments, the cloud server 103 may also remove the operating data within a preset time period after the air conditioner 1001 is turned on.

[0109] It is understandable that the operating data of the air conditioner 1001 for a period of time after startup may be affected by various factors, such as temperature adjustment, wind speed changes, etc., which may cause fluctuations in the operating data of the air conditioner 1001. By eliminating the operating data during this period, stable operating data can be obtained, thereby accurately diagnosing the fault of the air conditioner 1001.

[0110] In some embodiments, the preset time period can be obtained by formula (2). N = [ea(tb)] Formula (2)

[0111] N is the preset time period; a and b are constant coefficients; and t is the diagnosis cycle.

[0112] In some embodiments, as shown in FIG. 10 , after step 601 and before step 602 , the cloud server 103 is further configured to perform step 607 .

[0113] In step 607, the operating data of the air conditioner 1001 is processed into a target data type.

[0114] The target data type is a data type that can be used for diagnosis by at least one of the first fault diagnosis model or the second fault diagnosis model.

[0115] The operating data of the air conditioner 1001 is converted into operating data that can be processed and recognized by at least one of the first fault diagnosis model and the second fault diagnosis model. In this way, the operating data can be kept consistent with the requirements of the target data type, thereby reducing the deviation of the fault diagnosis model caused by the difference in operating data of different air conditioners 1001.

[0116] In some embodiments, the target data type can indicate the data volume, and the cloud server 103 can process the operating data of the air conditioner 1001 according to the target data type so that the data volume of the processed operating data of the air conditioner 1001 matches the data volume indicated by the target data type.

[0117] It is understood that by making the operating data of different air conditioners 1001 have a consistent data volume. In this way, the inconsistency of the input features of at least one of the first diagnosis fault diagnosis model or the second diagnosis fault diagnosis model caused by the different data volumes of the operating data of different air conditioners 1001 can be reduced, and the deviation of the diagnosis fault diagnosis model can be further reduced. By maintaining the consistency of the input features, the diagnosis fault diagnosis model can accurately diagnose and diagnose the fault of the air conditioner 1001 for different air conditioners 1001.

[0118] In some embodiments, as shown in FIG11 , step 607 includes steps 6071 to 6073 .

[0119] In step 6071, it is determined whether the operating data of the air conditioner 1001 is the first data. If "yes", step 6072 is executed; if "no", step 6073 is executed.

[0120] In step 6072, the operating data of the air conditioner 1001 is weighted averaged according to the weight value corresponding to the operating data of the air conditioner 1001, so that the operating data of the air conditioner 1001 after weighted average matches the target data type.

[0121] It should be noted that the weight value represents the on or off state of the air conditioner 1001. For example, when the air conditioner 1001 is on, the weight value of the air conditioner 1001 operating data is 1. When the air conditioner 1001 is off, the weight value of the air conditioner 1001 operating data is 0. The first data includes temperature data and pressure data.

[0122] In some embodiments, when the at least one indoor unit 301 and the at least one outdoor unit 302 in the air conditioner 1001 are considered as a whole, the first data can be considered as data obtained by diagnosing the same object at different locations. For example, with respect to the outdoor temperature in the operating data of the air conditioner 1001, each outdoor unit 302 may include a separate temperature sensor. Due to differences in the location, configuration, and external conditions of different outdoor units 302, the outdoor temperatures diagnosed by different outdoor units 302 may vary slightly.

[0123] In some embodiments, the cloud server 103 may perform weighted averaging on the operating data of the air conditioner 1001 to match the weighted average operating data of the air conditioner 1001 with the target data type, thereby increasing the consistency of the input features.

[0124] Take the air conditioner 1001 as an example, which includes a first indoor unit 301A, a second indoor unit 301B, and a third indoor unit 301C. Furthermore, the operating data of the first indoor unit 301A is A1, and the weight value of the operating data of the first indoor unit 301A is status1. The operating data of the second indoor unit 301B is B1, and the weight value of the operating data of the second indoor unit 301B is status2. The operating data of the third indoor unit 301C is C1, and the weight value of the operating data of the third indoor unit 301C is status3. The operating data of the air conditioner 1001 is weighted averaged, and the weighted average value D is obtained according to formula (3). D = (status1×A1+status2×B1+status3×C1) / (status1+status2+status3) Formula (3)

[0125] In step 6073, the operating data of the air conditioner 1001 is added so that the added operating data of the air conditioner 1001 matches the target data type.

[0126] It should be noted that, when the operating data of the air conditioner 1001 is the second data, the cloud server 103 adds the operating data of the air conditioner 1001 so that the added operating data of the air conditioner 1001 matches the target data type. The second data includes other data in addition to the temperature data and the pressure data. For example, the second data includes frequency data, air volume level data, current data, and valve opening data.

[0127] In some embodiments, when the operating data of the air conditioner 1001 represents the characteristics and performance of the air conditioner 1001 itself, and at least one indoor unit 301 in the air conditioner 1001 is regarded as a whole, and at least one outdoor unit 302 is regarded as a whole, the total operating data obtained by adding up the operating data of at least one indoor unit 301 or at least one outdoor unit 302 reflects the comprehensive characteristics and performance of the entire air conditioner 1001, and the operating data of the air conditioner 1001 can be determined as the second data.

[0128] In some embodiments, the cloud server 103 may add the operating data of the air conditioner 1001 when the operating data of the air conditioner 1001 is the second data. This approach not only makes the added operating data of the air conditioner 1001 match the target data type, thereby maintaining the consistency of the input features, but also preserves the physical meaning of the operating data.

[0129] It should be noted that steps 601 to 604 , steps 605 to 606 , steps 6051 and 6052 , step 607 , and steps 6071 to 6072 may be executed by the processor 203 of the cloud server 103 .

[0130] Some embodiments of the present disclosure provide a method for controlling a cloud server 103 .

[0131] As shown in FIG. 5 , the method includes steps 601 to 604 .

[0132] In step 601, the operating data of the air conditioner is obtained.

[0133] In step 602, the operating data of at least one indoor unit is combined with the operating data of at least one outdoor unit to obtain first combined data and second combined data.

[0134] In step 603, the first combined data is input into the first fault diagnosis model to obtain first fault information of the air conditioner in the first dimension, and the second combined data is input into the second fault diagnosis model to obtain second fault information of the air conditioner in the second dimension.

[0135] In step 604, a fault diagnosis result of the air conditioner is determined based on the first fault information and the second fault information.

[0136] For example, as shown in FIG12 , the operating data of the air conditioner can be obtained from the device side through the cloud server 103, and the operating data can be stored in the database. When the cloud server 103 performs fault diagnosis on the air conditioner, the cloud server 103 can obtain the operating data of the air conditioner from the database. The cloud server 103 can perform data preprocessing on the acquired operating data of the air conditioner 1001, including non-steady-state data elimination, data unification (, data combination. After performing data preprocessing on the acquired operating data of the air conditioner 1001, the cloud server 103 can input the preprocessed operating data of the air conditioner 1001 into the first diagnostic fault diagnosis model and the second diagnostic fault diagnosis model to obtain the fault diagnosis result of the air conditioner 1001. When the cloud server 103 determines that the air conditioner 1001 has a fault, the cloud server 103 can issue a fault warning. The cloud server 103 can also send the fault diagnosis result of the air conditioner 1001 to the user side for visual display to the user at the user side.

[0137] As the structure and application scenarios of multi-split air conditioning units (MSUs) continue to evolve, the types of faults they encounter are becoming increasingly diverse. Currently, fault diagnosis methods for MSUs primarily rely on data-driven detection methods. These methods are generally applicable only to specific types of MSUs. Due to the diversity of MSUs in terms of the number of indoor and outdoor units, the capacity ratio of indoor and outdoor units, and the installation scenarios, these methods lack generalizability and portability across systems, making them difficult to effectively adapt to the diverse range of MSUs.

[0138] To address the above-mentioned issues, some embodiments of the present disclosure provide a cloud server 103. The cloud server 103 can generate a feature space distribution of a second sample set based on the normal operating data and fault operating data of the second multi-split air-conditioning unit, and adjust the first sample set generated based on the normal operating data of the first multi-split air-conditioning unit so that the feature space distribution of the first sample set is consistent with or close to the feature space distribution of the second training sample. In this way, a fault diagnosis model can be trained based on the first sample set and the second sample set with consistent or close feature space distributions. The trained fault diagnosis model can be applied to fault diagnosis of the first and second multi-split air-conditioning units, thereby realizing cross-unit fault diagnosis and improving the generalization performance of the fault diagnosis model.

[0139] In some embodiments of the present disclosure, an air conditioner utilizes a compressor, a condenser, an electronic expansion valve, an evaporator, and a four-way valve as a refrigerant circulation circuit to implement a refrigeration cycle. The refrigeration cycle includes a series of processes involving compression, condensation, expansion, and evaporation, supplying refrigerant to the conditioned and heat-exchanged air.

[0140] The compressor compresses low-temperature, low-pressure refrigerant gas and discharges the compressed, high-temperature, high-pressure refrigerant gas. The discharged refrigerant gas flows into the condenser. The condenser condenses the compressed refrigerant gas into liquid refrigerant, releasing heat into the surrounding environment through the condensation process.

[0141] The electronic expansion valve throttles the high-temperature, high-pressure liquid refrigerant condensed in the condenser to a low-pressure, two-phase refrigerant. The evaporator evaporates the refrigerant throttled by the electronic expansion valve and returns the low-temperature, low-pressure refrigerant gas to the compressor. The evaporator achieves cooling by utilizing the latent heat of evaporation to exchange heat with the medium being cooled. Throughout this cycle, the air conditioner regulates the indoor temperature.

[0142] The outdoor unit of the air conditioner includes a compressor and an outdoor heat exchanger, the indoor unit of the air conditioner includes an indoor heat exchanger, and the expansion valve can be provided in the indoor unit or the outdoor unit.

[0143] The indoor heat exchanger and the outdoor heat exchanger function as a condenser or an evaporator. When the indoor heat exchanger functions as a condenser, the air conditioner functions as a heater in a heating mode, and when the indoor heat exchanger functions as an evaporator, the air conditioner functions as a cooler in a cooling mode.

[0144] As shown in FIG13 , an application scenario of some embodiments of the present disclosure includes multiple multi-split air conditioning units 100, for example, including a first multi-split air conditioning unit 101 and a second multi-split air conditioning unit 102; the application scenario also includes a cloud server 103. Cloud server 103 is in communication with the multiple multi-split air conditioning units 100. In some embodiments, the multi-split air conditioning units 100 can send their operating data to cloud server 103, so that cloud server 103 can detect faults in the multi-split air conditioning units 100 based on the operating data.

[0145] As shown in FIG. 14 , the multi-split air conditioning unit 100 includes at least one indoor unit 301 .

[0146] In some embodiments, the number of the indoor unit 301 is at least one, and the number of the outdoor unit 302 is at least one. This application does not limit the number of the indoor unit 301 and the outdoor unit 302.

[0147] The multi-split air conditioning unit 100 is a wall-mounted air conditioner or a floor-standing air conditioner. In the case of a wall-mounted air conditioner, the indoor unit 301 can be mounted on a wall indoors; in the case of a floor-standing air conditioner, the indoor unit 301 can be mounted on the ground.

[0148] As shown in Figure 14, the multi-split air conditioning unit 100 further includes at least one outdoor unit 302. The at least one outdoor unit 302 is located outdoors and can be connected to the at least one indoor unit 101, configured to exchange heat with the indoor environment. Furthermore, the outdoor unit 302 is typically located outdoors on the opposite side of the indoor unit 301, separated by a wall.

[0149] In some embodiments, as shown in Figures 15A and 15B, the multi-split air conditioning unit 100 further includes a controller 303. The controller 303 is a device that can generate an operation control signal based on an instruction operation code and a sequence signal to instruct the multi-split air conditioning unit 100 to execute the control instruction. For example, the controller 303 can be 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 unit, or a programmable logic device (PLD). The controller 303 can also be other devices with processing functions, such as circuits, devices, or software devices.

[0150] In addition, the outdoor unit 302 and the indoor unit 301 are respectively connected to the controller 303 for communication, and can perform related operations according to instructions of the controller 303.

[0151] In some embodiments, as shown in Figures 15A and 15B, the outdoor unit 302 includes a four-way valve 404. The four-way valve 404 is connected to the refrigerant circuit. The four-way valve 404 includes four ports. The four ports are respectively a first port, a second port, a third port, and a fourth port. The fourth port of the four-way valve 404 is connected to the first end of the first heat exchanger 406 to allow the refrigerant to circulate between the first heat exchanger 406 and the four-way valve 404. The four-way valve 404 is configured to achieve mutual conversion between cooling and heating by changing the flow direction of the refrigerant in the pipeline.

[0152] In some embodiments, as shown in Figures 15A and 15B, the outdoor unit 302 further includes a compressor 401. The compressor 401 is configured to compress the refrigerant so that the low-pressure refrigerant is compressed to form a high-pressure refrigerant. The high-pressure refrigerant can flow from the compressor 401 to the first port of the four-way valve 404.

[0153] As shown in Figures 15A and 15B, the outdoor unit 302 further includes a accumulator 403. The accumulator 403 is configured to store refrigerant. A first end of the accumulator 403 is connected to a second end of the compressor 401 so that the accumulator 403 supplies refrigerant gas to the suction port of the compressor 401. A second end of the accumulator 403 is connected to a third port of a four-way valve 404 so that the refrigerant flowing out of the third port of the four-way valve 404 is output to the accumulator 403.

[0154] In some embodiments, the liquid reservoir 403 is further configured to separate the refrigerant in the liquid reservoir 403 into gaseous refrigerant and liquid refrigerant, so that the gaseous refrigerant flows into the compressor 401 .

[0155] In some embodiments, as shown in Figures 15A and 15B , the outdoor unit 302 further includes an outdoor heat exchanger 402. A first end of the outdoor heat exchanger 402 is connected to the second port of a four-way valve 404. As shown in Figures 15A and 15B , the outdoor unit 302 further includes a throttling device 405. The throttling device 405 is connected to the second end of the second heat exchanger 402. The outdoor heat exchanger 402 has a first inlet and outlet for circulating refrigerant between the outdoor heat exchanger 402 and the suction port of the compressor 401 via a liquid accumulator 403, and a second inlet and outlet for circulating refrigerant between the outdoor heat exchanger 402 and the throttling device 405. The outdoor heat exchanger 402 exchanges heat between the heat exchanger flowing in the heat transfer pipe connected between the first inlet and the second inlet and the outdoor air. In cooling mode, the outdoor heat exchanger 402 operates as a condenser.

[0156] Since the four-way valve 404 is connected to the compressor 401 and the second heat exchanger 402 , the refrigerant output from the accumulator 403 can flow among the compressor 401 , the four-way valve 404 and the second heat exchanger 402 .

[0157] In some embodiments, as shown in Figures 15A and 15B, the indoor unit 301 includes an indoor heat exchanger 406. The indoor heat exchanger 406 has a third inlet and outlet for liquid refrigerant to flow between the four-way valve 404 and a fourth inlet and outlet for gaseous refrigerant to flow between the compressor 401's discharge port. The indoor heat exchanger 406 exchanges heat between the refrigerant flowing in the heat transfer pipe connected between the third and fourth inlets and outlets and the indoor air. The following describes how the cloud server 103 diagnoses faults in the multi-split air conditioning unit 100 based on its operating data.

[0158] In some embodiments, as shown in FIG16 , the cloud server 103 is configured to perform steps 11 to 15 .

[0159] In step 11, a first sample set and a second sample set are obtained.

[0160] The first sample set includes multiple first samples. The multiple first samples are generated based on the normal operating data (first operating data) of the first multi-split air conditioning unit 101. The second sample set includes multiple second samples, which are generated based on the normal operating data (first operating data) and fault operating data (second operating data) of the second multi-split air conditioning unit 102. The second samples have a first label (operating mode label) or a second label (fault type label). The normal operating data of the first multi-split air conditioning unit 101 is the operating data of the first multi-split air conditioning unit 101 in the first operating state. The normal operating data of the second multi-split air conditioning unit 102 is the operating data of the second multi-split air conditioning unit 102 in the first operating state. The fault operating data of the second multi-split air conditioning unit 102 is the operating data of the second multi-split air conditioning unit 102 in the second operating state. The second operating state can refer to a component failure in the indoor unit 301 or the outdoor unit 302, such as a compressor 401 congestion failure or a throttling device 405 failure. This may involve various factors such as electrical failure, mechanical failure, and sensor failure. The first operating state may mean that no fault occurs in any component of the indoor unit 301 or the outdoor unit 302 and each component operates under its rated operating condition.

[0161] The first tag is used to indicate the specific operating modes corresponding to the multiple second samples generated based on the normal operating data of the second multi-split air conditioning unit 102. During normal operation of the second multi-split air conditioning unit 102, different operating modes may exist, such as cooling mode, heating mode, and ventilation mode. The ventilation mode may refer to exhausting indoor air to the outside or supplying outdoor air to the room to achieve ventilation. By setting the first tag, different operating modes can be distinguished, thereby enabling monitoring and identification of the operating mode of the second multi-split air conditioning unit 102.

[0162] The second tag is used to indicate a specific fault type corresponding to a sample generated based on the fault operation data of the second multi-split air conditioning unit 102. When a fault occurs in the second multi-split air conditioning unit 102, it may involve different fault types, such as compressor fault, sensor fault, circuit fault, etc.

[0163] By adding the corresponding first label or second label to any second sample, the sample data can be made more readable, which helps to accurately monitor, analyze and maintain the multi-split air-conditioning unit 100.

[0164] In some embodiments, the cloud server 103 is also configured to construct a first sample set based on the first operating data in the first test data of the first multi-split air-conditioning unit, and may also construct a second sample set based on the first operating data and the second operating data in the second test data of the second multi-split air-conditioning unit.

[0165] For example, as shown in FIG. 17 , step 11 includes steps 111 to 113 .

[0166] In step 111 , first test data of the first multi-split air-conditioning unit 101 and second test data of the second multi-split air-conditioning unit 102 are obtained.

[0167] The first test data includes first operating data of the first multi-split air-conditioning unit 101 , and the second test data includes first operating data and second operating data (fault operating data) of the second multi-split air-conditioning unit.

[0168] It is understood that the first and second test data each include multiple categories of operating data. This diversity of operating data means data diversity, covering a wide range of scenarios. This helps the fault diagnosis model comprehensively learn the patterns and characteristics of the multi-split air conditioning unit 100 in different scenarios, thereby improving the accuracy of the fault diagnosis model's fault prediction and enhancing the generalization ability of the high-speed fault diagnosis model.

[0169] In some embodiments, a tester can test the first multi-split air conditioning unit 101 under different test environments to obtain first test data for the first multi-split air conditioning unit 101 under the different test environments. The first test data includes first operating data of the first multi-split air conditioning unit 101 under the different test environments. The cloud server 103 can obtain the first test data of the first multi-split air conditioning unit 101 in real time and save the first test data of the first multi-split air conditioning unit 101 in a database of the cloud server 103 for later retrieval and analysis. In this way, the controller 303 or the cloud server 103 can obtain the first test data of the first multi-split air conditioning unit 101 from the database.

[0170] Similarly, the cloud server 103 can also obtain the second test data of the second multi-split air-conditioning unit 102 from the database.

[0171] In some embodiments, the first test data may include at least one of the following: the operating conditions (indoor temperature and outdoor temperature), load factor, set temperature, and air volume level of the first multi-split air conditioning unit 101 under different test environments; and the second test data may include at least one of the following: the indoor temperature and outdoor temperature, load factor, set temperature, and air volume level of the second multi-split air conditioning unit 102 under different test environments. The load factor may refer to the ratio of the number of powered indoor units 301 to the total number of indoor units 301 included in the multi-split air conditioning unit 100. The set temperature may refer to the set indoor temperature.

[0172] In step 112 , the first test data and the second test data are preprocessed to obtain processed first test data and second test data.

[0173] In some embodiments, preprocessing includes one or more of the following: data cleaning, data reduction, data coding, and data normalization.

[0174] In some embodiments, data cleaning is configured to indicate the removal of non-steady-state data from the first test data and the second test data. Taking data cleaning of the first test data as an example, the cloud server 103 may calculate the slope of the first test data within a sliding window according to a preset window length based on a sliding window algorithm. If the slope is outside a preset range, the cloud server 103 determines the first test data within the sliding window as non-steady-state operation data and removes the first test data within the sliding window.

[0175] In some embodiments, the data reduction is configured to reduce the size of the first test data and the second test data while retaining key information and features of the first test data and the second test data. Taking data reduction of the first test data as an example, the cloud server 103 can map the first test data from a high-dimensional space to a low-dimensional space based on a dimensionality reduction method (such as a principal component analysis method) to reduce the dimensionality of the first test data, thereby reducing the complexity of the first test data.

[0176] In some embodiments, data encoding can be configured to convert the first and second test data into other formats for use during transmission or storage. Taking the encoding of the first test data as an example, if the first test data includes wind speed data, data encoding can be used to convert the wind speed data into a digital label. For example, "first wind speed" can be encoded as 0, "second wind speed" as 1, "third wind speed" as 2, and so on; the first wind speed is less than the second wind speed, and the second wind speed is less than the third wind speed.

[0177] In some embodiments, data normalization is configured to convert the first test data and the second test data into a uniform scale range to eliminate dimensional differences between different variables and facilitate data analysis and processing.

[0178] In some embodiments, because the capacities of different indoor units 301 in the first multi-split air conditioning unit 101 vary, the cloud server 103 may also perform weighted averaging on the first test data of the first multi-split air conditioning unit 101 based on the capacities of the indoor units 301, thereby considering the performance of the entire first multi-split air conditioning unit rather than the performance of each indoor unit 301 individually. The capacity of an indoor unit 301 may refer to the cooling capacity of the indoor unit 301. Similarly, the cloud server 103 may also perform weighted averaging on the second test data of the second multi-split air conditioning unit 102.

[0179] In step 113 , a first sample set and a second sample set are generated according to the processed first test data and the second test data.

[0180] In some embodiments, the cloud server 103 may divide the processed first test data into a first training sample set and a first validation sample set, wherein the first training sample set is referred to as the first sample set. The first validation sample set is configured to verify the accuracy of the fault diagnosis model. The cloud server 103 may also divide the processed second test data into a second training sample set and a second validation sample set, wherein the second training sample set is referred to as the second sample set.

[0181] In step 12, when the proportion of the target samples in the second sample set is less than or equal to a preset threshold, the second sample set is expanded to adjust the proportion of the target samples in the second sample set.

[0182] The target sample may refer to a second sample having a first label or a second label.

[0183] It can be understood that when the proportion of target samples in the second sample set is less than or equal to the preset threshold, it indicates that the number of target samples is small.

[0184] In some embodiments, the proportion of target samples in the second sample set affects the accuracy of the fault diagnosis model. Therefore, the second sample set can be expanded to adjust the proportion of target samples in the second sample set to improve the accuracy of the fault diagnosis model.

[0185] In some embodiments, as shown in FIG18 , the cloud server 103 further includes a first adjustment unit 120. The first adjustment unit 120 may be established based on a synthetic minority oversampling technique (SMOTE).

[0186] In some embodiments, the cloud server 103 may input the second sample set into the first adjustment unit 120 to obtain an expanded second sample set.

[0187] As shown in FIG18 , the first adjustment unit 120 includes a first sub-adjustment unit 1201 (a rate setting module). The first sub-adjustment unit 1201 is configured to set a sampling rate. The sampling rate refers to the ratio of the number of new samples required to generate any one of the multiple target samples to the target number of samples when generating new samples.

[0188] As shown in FIG18 , the first adjustment unit 120 further includes a second sub-adjustment unit 1202 (K-nearest neighbor calculation module). The second sub-adjustment unit 1202 has a built-in K-nearest neighbor (KNN) algorithm. The second sub-adjustment unit 1202 is configured to obtain at least one neighboring sample of the target sample.

[0189] As shown in Figure 18, the first adjustment unit 120 further includes a third sub-adjustment unit 1203 (new sample construction module). The third sub-adjustment unit 1203 is configured to perform random linear interpolation processing between the target sample and any neighboring sample to generate a new target sample.

[0190] As shown in FIG18 , the first adjustment unit 120 further includes a fourth sub-adjustment unit 1204 (data ratio adjustment module). The fourth sub-adjustment unit 1204 is configured to add the newly added target samples to the second sample set to adjust the ratio of the target samples in the second sample set. The fourth sub-adjustment unit 1204 is further configured to adjust the first sample set based on the feature space distribution of the second sample set.

[0191] In some embodiments, as shown in FIG. 19 , step 12 includes steps 121 to 123 .

[0192] In step 121 , at least one neighboring sample of the target sample is obtained by the second sub-adjustment unit 1202 .

[0193] In some embodiments, before obtaining at least one neighboring sample of the target sample through the second sub-adjustment unit 1202 , the cloud server 103 may set a sampling magnification through the first sub-adjustment unit 1201 .

[0194] In some embodiments, the cloud server 103 may use a K-nearest neighbor algorithm to find at least one neighboring sample of the second sample having the first label. For example, if the at least one target second sample includes multiple target samples, the cloud server 103 may calculate the distance between any one of the multiple target samples and any one of the other target samples. Based on the distance, the cloud server 103 may find the K nearest neighboring samples of the target sample.

[0195] In step 122 , the third sub-adjustment unit 1203 performs random linear interpolation processing between the target sample and any neighboring sample to generate a new sample with the first label or the second label.

[0196] In some embodiments, a point is randomly selected on the line between the target sample and any neighboring sample of the target sample, and the point is a newly added sample with the first label.

[0197] For example, the neighboring sample of the first target sample A is sample B. On the line between the first target sample A and sample B, a point is randomly selected, which is the newly added sample C, and sample C has the first label.

[0198] In some embodiments, the cloud server 103 may also perform random linear interpolation processing according to formula (1) to obtain newly added minority class training samples (ie, newly added target samples). new =x+rand(0,1)*|x-xn| Formula (1)

[0199] x new is a newly added sample with the first label or the second label; x is a newly added sample with the first label or the second label in the second sample set; rand(0, 1) is used to randomly select a value between 0 and 1; xn is any one of the at least one neighboring samples of the newly added sample with the first label.

[0200] In some embodiments, the cloud server 103 may also perform random linear interpolation processing according to formula (2) to obtain newly added minority class training samples (ie, newly added target samples). new =x+rand(0,1)*|x-xn| Formula (2)

[0201] x new is a newly added sample with the first label or the second label; x is any target sample in the second sample set; rand(0, 1) is used to randomly select a value between 0 and 1; xn is any neighboring sample of at least one neighboring sample of the newly added sample with the first label.

[0202] In step 123 , the newly added samples with the first label or the second label are added to the second sample set to adjust the proportion of the target samples in the second sample set.

[0203] In step 13, the first sample set is adjusted according to the feature space distribution of the second sample set.

[0204] In some embodiments, the fourth sub-adjustment unit 1204 adjusts the first sample set according to the feature space distribution of the second sample set.

[0205] In some embodiments, the cloud server 103 may adjust the first sample set according to the feature space distribution of the second sample set so that the feature space distribution of the first sample set is consistent with or close to the feature space distribution of the second sample set.

[0206] It can be understood that the feature distribution of the first training sample is converted to be consistent or close to the feature distribution of the second training sample. In this way, after the fault diagnosis model is trained based on the feature distributions that are consistent or close to the first training sample and the second training sample, the fault diagnosis model can perform fault diagnosis on both the first multi-split air-conditioning unit 101 and the second multi-split air-conditioning unit 102, thereby realizing cross-unit fault diagnosis.

[0207] In some embodiments, the cloud server 103 may input the first sample set and the second sample set into a domain adaptation model respectively to adjust the first sample set.

[0208] For example, the domain adaptation model adopts the DANN (Domain Adaptation Neural Network) structure.

[0209] As shown in Figure 20, the adversarial neural network includes a feature extractor. For example, the feature extractor is a convolutional neural network (CNN) composed of multiple one-dimensional convolutional layers and one-dimensional max-pooling layers. The data input to the feature extractor undergoes convolution operations in the convolutional layers and dimensionality reduction in the max-pooling layers to extract effective features. These effective features can be obtained using feature extraction and feature selection methods.

[0210] In some embodiments, in the feature extractor, nonlinear mapping and feature extraction can be performed on the input data to facilitate capturing and representing the complexity of the input data. This process can be expressed by formula (6). G_f = sigm(Wx+b) Formula (6)

[0211] As shown in Figure 20, the adversarial neural network also includes a label classifier. The label classifier is a fully connected neural network composed of multiple fully connected layers. The data is mapped into an output vector with a length equal to the number of categories through the fully connected neural network. The label classifier is configured to distinguish the label type of the input data, which includes a first label and a second label.

[0212] As shown in Figure 20, the adversarial neural network also includes a domain classifier. The domain classifier is a fully connected neural network composed of a gradient reversal layer and multiple fully connected layers. During feature extraction, the data is mapped into an output value of length 1 through the fully connected neural network. The domain classifier is configured to distinguish the source of input data. For example, the domain classifier can distinguish whether the input data is data from the first multi-split air conditioner unit 101 or data from the second multi-split air conditioner unit 102.

[0213] It should be noted that the adversarial neural network adopts the idea of ​​adversarial learning, so that the label classifier and the domain classifier compete with each other during the training process, achieve a mutual balance between the label classification loss and the domain classification loss, and optimize the feature extractor, so that the feature space distribution of the features extracted by the feature extractor from the first sample in the first sample set is consistent with or close to the feature space distribution of the features extracted from the second sample in the second sample set, thereby enabling the fault diagnosis model to perform fault diagnosis on both the first multi-split air-conditioning unit 101 and the second multi-split air-conditioning unit 102, thereby realizing cross-unit fault diagnosis.

[0214] For example, as shown in FIG. 21 , step 13 includes steps 131 to 135 .

[0215] In step 131, a feature extractor is used to extract a first feature from the first sample, a second feature is extracted from the second sample generated based on the first operating data of the second multi-split air-conditioning unit 102, and a third feature is extracted from the second sample generated based on the second operating data of the second multi-split air-conditioning unit 102.

[0216] It should be noted that step 131 is an extraction operation.

[0217] In step 132 , the first feature and the second feature are respectively input into a domain classifier to obtain a domain loss function, and the third feature is input into a label classifier to obtain a label loss function.

[0218] It should be noted that step 132 is an input operation.

[0219] In some embodiments, the cloud server 103 inputs the first feature into a domain classifier to obtain a first classification probability, and a domain loss function may be determined based on the first classification probability.

[0220] In some embodiments, in the domain classifier, a nonlinear transformation may be performed on the input features so as to distinguish in the feature distribution space whether the input features come from the first multi-split air-conditioning unit 101 or the second multi-split air-conditioning unit 102. This process can be represented by formula (7). d =sigm(u T G f (x)+z) Formula (7)

[0221] Similarly, in some embodiments, the cloud server 103 may input the second feature into a label classifier to obtain a second classification probability, and determine a label loss function based on the second label classification probability.

[0222] In some embodiments, in the label classifier, a nonlinear transformation is performed on the input features so that the label classifier can adapt to various complex data distributions and patterns, thereby distinguishing the corresponding label types of the input features. This process can be expressed by formula (8). y =softmax(VG f (x)+c) Formula (8)

[0223] In step 133, the parameters of the feature extractor, the domain classifier, and the label classifier are updated according to the domain loss function and the label loss function through the back propagation algorithm.

[0224] It should be noted that step 133 is an update operation.

[0225] In some embodiments, during backpropagation, the gradient of the label classifier is first calculated using the label loss function. This gradient is then propagated to the feature extractor according to the chain rule to update the parameters of the feature extractor. Next, the gradient of the domain classifier is calculated using the domain loss function. This gradient is also propagated to the feature extractor to update the parameters of the feature extractor.

[0226] During the above-mentioned domain adaptation process, the feature extractor is continuously optimized so that the feature space distribution of the features extracted by the feature extractor from the first sample in the first sample set is consistent with or close to the feature space distribution of the features extracted from the second sample in the second sample set.

[0227] In step 134, a determination is made as to whether the domain classification loss function value of the domain classifier is greater than or equal to a first threshold, and whether the label loss function value of the label classifier is less than a second threshold; or whether the number of iterations is greater than a preset number. If yes, the process proceeds to step 135; if no, the process returns to step 132. In step 135, convergence conditions are met.

[0228] It should be noted that, in some embodiments, the cloud server 103 may also repeatedly perform the input operation and the update operation until the domain classification loss function value of the domain classifier is maximized and the label loss function value of the label classifier is minimized.

[0229] It can be understood that, on the one hand, the domain classification loss function value of the domain classifier is greater than or equal to the first threshold, indicating that the domain classifier cannot distinguish whether the features of the sample are from the first multi-split air-conditioning unit 101 or from the second multi-split air-conditioning unit 102. This means that the fault diagnosis model established based on the adjusted first sample set and second sample set can diagnose both the first multi-split air-conditioning unit 101 and the second multi-split air-conditioning unit 102.

[0230] On the other hand, when the loss function value of the label classifier is less than or equal to the second threshold, it indicates that the label classifier has high label classification accuracy for the sample, which helps to improve the accuracy of the fault diagnosis model.

[0231] In step 14, the initial model is trained based on the first sample set and the second sample set to obtain a trained fault diagnosis model.

[0232] In some embodiments, the first sample set and the second sample set include a training set and a test set, respectively. An initial model is selected, and training and parameter optimization are performed on the initial model to obtain a trained fault diagnosis model.

[0233] For example, the ratio of the number of samples in the training set to the number of samples in the test set may be 9: 1. The initial model may be a support vector machine (SVM) model.

[0234] In step 15, a fault diagnosis model is used to perform fault diagnosis on the first multi-split air-conditioning unit 101.

[0235] In some embodiments, after obtaining the trained fault diagnosis model, the cloud server 103 can obtain the real-time operation data of the first multi-split air-conditioning unit 101, input the real-time operation data into the fault diagnosis model, and obtain the fault diagnosis result of the first multi-split air-conditioning unit 101.

[0236] In some embodiments, after obtaining the trained fault diagnosis model, the cloud server 103 can also use the real-time operation data of the first multi-split air-conditioning unit 101 as the first test data of the first multi-split air-conditioning unit 101, and re-execute steps 11 to 14, steps 111 to 113, steps 121 to 123, and steps 131 to 136 to update the fault diagnosis model.

[0237] For example, the fault diagnosis model includes a target domain input channel and a source domain input channel. The target domain input channel is configured to input real-time operating data of the first multi-split air conditioning unit 101, and the source domain input channel is configured to input other labeled operating data. The real-time operating data of the first multi-split air conditioning unit can be input into the fault diagnosis model through the target domain input channel to repeatedly execute steps 11 to 14, steps 111 to 113, steps 121 to 123, and steps 131 to 136 to update the fault diagnosis model.

[0238] It is understandable that other labeled operating data include multiple types of operating data, such as various types of fault data. When the real-time operating data of the first multi-split air-conditioning unit 101 input through the target domain input channel is insufficient or incomplete, transfer learning can help the fault diagnosis model learn a wide range of fault modes and characteristics, thereby improving the fault diagnosis model's ability to diagnose target domain faults, as well as improving the performance and generalization ability of the model.

[0239] It should be noted that the preceding description uses cloud server 103 to diagnose faults in the multi-split air conditioning unit 100 based on the operating data of the multi-split air conditioning unit 100. Of course, in some embodiments, controller 303 can also diagnose faults in the multi-split air conditioning unit 100 based on the operating data of the multi-split air conditioning unit 100. In this way, controller 303 can also perform the steps in Figures 16, 17, 19, and 21, and can input the first sample set and the second sample set into the domain adaptive model to adjust the first sample set.

[0240] Some embodiments of the present disclosure further provide a fault diagnosis device 1200. The fault diagnosis device 1200 can diagnose faults of the multi-split air conditioning unit 100 based on the operating data of the multi-split air conditioning unit 100.

[0241] As shown in Fig. 22, the fault diagnosis apparatus 1200 includes an acquisition component 1301. The acquisition component 1301 is configured to acquire a first sample set and a second sample set.

[0242] The fault diagnosis apparatus 1200 further includes an expansion component 1302. The expansion component 1302 is configured to expand the second sample set to adjust the proportion of the target sample in the second sample set when the proportion of the target sample in the second sample set is less than or equal to a preset threshold.

[0243] The fault diagnosis apparatus 1200 further includes a second adjustment unit 1303. The second adjustment unit 1303 is configured to adjust the first sample set according to the feature space distribution of the second sample set.

[0244] The fault diagnosis apparatus 1200 further includes a training component 1304. The training component 1304 is configured to train the initial model based on the first sample set and the second sample set to obtain a trained fault diagnosis model.

[0245] The fault diagnosis device 1200 further includes a diagnosis component 1305. The diagnosis component 1305 is configured to perform fault diagnosis on the first multi-split air-conditioning unit 101 using a fault diagnosis model.

[0246] In some embodiments, the fault diagnosis device 1200 further includes a storage component configured to store program code and data of the fault diagnosis device. The fault diagnosis device 1200 further includes a communication component, which may be a transceiver, a transceiver circuit, or a communication interface.

[0247] Some embodiments of the present disclosure also provide a method for controlling a cloud server. As shown in FIG16 , the method includes:

[0248] In step 11, a first sample set and a second sample set are obtained.

[0249] In step 12, the first test data and the second test data are preprocessed to obtain processed first test data and second test data.

[0250] In step 13, a first sample set and a second sample set are generated according to the processed first test data and the second test data.

[0251] In step 14, the initial model is trained based on the first sample set and the second sample set to obtain a trained fault diagnosis model.

[0252] In step 15, a fault diagnosis model is used to perform fault diagnosis on the first multi-split air-conditioning unit 101.

[0253] For example, as shown in Figure 23, first, the first test data of the first multi-split air-conditioning group 101 (such as the first operating data of the first multi-split air-conditioning group) and the second test data of the second multi-split air-conditioning group 102 (such as the first operating data and the second operating data of the second multi-split air-conditioning group) are preprocessed to obtain preprocessed data, and the processed data are divided into a training sample set (i.e., a first sample set and a second sample set) and a verification sample set.

[0254] Next, the second sample set is input into the first adjustment unit 120, which performs expansion processing on the second sample set to adjust the proportion of the target sample in the second sample set. In this way, the second sample set can be enriched.

[0255] Then, the first sample set and the expanded second sample set are input into the domain adaptation model to adjust the first sample set so that the feature space distribution of the first sample set is consistent with or close to the feature space distribution of the expanded second sample set.

[0256] Finally, after training the initial model based on the first sample set and the second sample set whose feature space distribution is consistent or close to each other, and obtaining the trained fault diagnosis model, the real-time operation data of the first multi-split air-conditioning unit 101 can be used as the first test data of the first multi-split air-conditioning unit 101, and steps 11 to 14, steps 111 to 113, steps 121 to 123, and steps 131 to 136 can be re-executed to update the fault diagnosis model, and fault diagnosis can be performed on the first multi-split air-conditioning unit 101.

[0257] It should be noted that any one of the technical solutions disclosed in the present disclosure can, to a certain extent, solve one or more of the above-mentioned technical problems and achieve certain disclosure purposes; multiple technical disclosures can also be combined into an overall solution to solve one or more of the above-mentioned technical problems and achieve certain disclosure purposes; some of the technical disclosures can also be selected to be combined into an overall solution, while adopting related technologies and inferior solutions, but the inferior trend can be compensated by the means disclosed in this technology, and the above-mentioned one or more technical problems can be solved to a certain extent as a whole and certain disclosure purposes can be achieved; each technical disclosure combined into a complete technical solution constitutes an organic and inseparable overall solution, which solves technical problems as a whole and achieves certain disclosure purposes.

[0258] Any technical disclosure in this disclosure, as well as the recombination of multiple technical disclosures, can form a complete technical solution and can solve one or more of the above-mentioned technical problems and achieve the purpose of disclosure. They all belong to the content of this disclosure and are the content that is directly and unambiguously determined based on the content of this disclosure.

[0259] Those skilled in the art will understand that the scope of the present disclosure 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 disclosure. The scope of the present disclosure is limited by the appended claims.

Claims

1. A cloud server, comprising: A communicator configured to establish a communication connection with an air conditioner; The air conditioner includes at least one indoor unit and at least one outdoor unit; And A processor communicatively connected to the communicator and configured to: Obtain the operation data of the air conditioner; wherein, the operation data of the air conditioner includes the operation data of the at least one indoor unit and the operation data of the at least one outdoor unit; Combine the operation data of the at least one indoor unit with the operation data of the at least one outdoor unit to obtain first combined data in a first dimension and second combined data in a second dimension; Input the first combined data into a first fault detection model to obtain first fault information of the air conditioner in the first dimension, and input the second combined data into a second fault detection model to obtain second fault information of the air conditioner in the second dimension; Determine the fault detection result of the air conditioner according to the first fault information and the second fault information.

2. The cloud server according to claim 1, wherein, The processor is further configured to: Respectively combine the operation data of any one of the at least one indoor units with the operation data of the at least one outdoor unit to obtain the first combined data; the first combined data includes the data corresponding to each of the at least one indoor units; Combine the operation data of the at least one indoor unit with the operation data of the at least one outdoor unit to obtain the second combined data.

3. The cloud server according to claim 1 or 2, wherein, The processor is further configured to: after obtaining the operation data of the air conditioner and before combining the operation data of the at least one indoor unit with the operation data of the at least one outdoor unit, process the operation data of the air conditioner, extract the non-steady-state operation data in the operation data of the air conditioner; eliminate the non-steady-state operation data.

4. The cloud server according to claim 3, wherein, The processor is further configured to: Periodically detect the operation data of the air conditioner with a preset time period before eliminating the non-steady-state operation data; For any one detection period, when the standard deviation of the operation data within the any one detection period is within a preset range, determine the operation data within the detection period as non-steady-state operation data.

5. The cloud server according to any one of claims 1 to 4, wherein The processor is further configured to: after obtaining the operation data of the air conditioner and before combining the operation data of the at least one indoor unit with the operation data of the at least one outdoor unit, process the operation data of the air conditioner into a target data type; The target data type is a detectable data type of at least one of the first fault detection model or the second fault detection model.

6. The cloud server according to claim 5, wherein, The processor is further configured: When the operation data of the air conditioner is first data, perform weighted averaging on the operation data of the air conditioner according to the weight value corresponding to the operation data of the air conditioner, so that the operation data of the air conditioner matches the target data type; wherein, the weight value is used to represent the on or off state of the air conditioner, and the first data includes temperature data and pressure data; When the operating data of the air conditioner is second data, the operating data of the air conditioner is added so that the operating data of the air conditioner matches the target data type; wherein the second data includes other operating data except the temperature data and the pressure data.

7. The cloud server according to any one of claims 1 to 6, wherein, The operation data of the air conditioner includes at least one of temperature data, pressure data, valve opening data, frequency data, air volume level data, and current data.

8. The cloud server according to any one of claims 1 to 7, wherein, The processor is further configured to: If the second fault information is fault information, determining that the air conditioner is faulty in the second dimension; If at least one of the first fault information or the second fault information is fault information, it is determined that the air conditioner is faulty in the first dimension.

9. A method for controlling a cloud server, the method being applied to the cloud server, wherein the cloud service comprises: a communicator configured to establish a communication connection with the air conditioner; The air conditioner includes at least one indoor unit and at least one outdoor unit; as well as a processor, communicatively connected to the communicator; The method comprises: acquiring operating data of the air conditioner; wherein the operating data of the air conditioner comprises operating data of the at least one indoor unit and operating data of the at least one outdoor unit; combining the operating data of the at least one indoor unit with the operating data of the at least one outdoor unit to obtain first combined data of a first dimension and second combined data of a second dimension; Inputting the first combined data into a first fault detection model to obtain first fault information of the air conditioner in the first dimension, and inputting the second combined data into a second fault detection model to obtain second fault information of the air conditioner in the second dimension; A fault detection result of the air conditioner is determined according to the first fault information and the second fault information.

10. The method according to claim 9, wherein, After acquiring the operating data of the air conditioner, and before combining the operating data of the at least one indoor unit with the operating data of the at least one outdoor unit to obtain first combined data of the first dimension and second combined data of the second dimension, the method further includes: Combining the operating data of any one of the at least one indoor unit with the operating data of the at least one outdoor unit to obtain the first combined data; the first combined data includes the data corresponding to each of the at least one indoor unit; The operating data of the at least one indoor unit is combined with the operating data of the at least one outdoor unit to obtain the second combined data.

11. The method according to claim 9 or 10, wherein After acquiring the operating data of the air conditioner and before combining the operating data of the at least one indoor unit with the operating data of the at least one outdoor unit, the method further includes: processing the operating data of the air conditioner to extract non-steady-state operating data from the operating data of the air conditioner; Eliminate the non-steady-state operation data.

12. The method according to claim 11, wherein, The processing of the operating data of the air conditioner to extract the non-steady-state operating data from the operating data of the air conditioner comprises: Performing periodic detection on the operating data of the air conditioner at a preset time; For any detection period, when the standard deviation of the operation data within the any detection period is within a preset range, determine the operation data within the detection period as non-steady-state operation data.

13. The method according to any one of claims 9 to 12, wherein, After obtaining the operation data of the air conditioner and before combining the operation data of the at least one indoor unit with the operation data of the at least one outdoor unit, the method further includes: Processing the operation data of the air conditioner into a target data type; the target data type is a detectable data type of at least one of the first fault detection model or the second fault detection model.

14. The method according to claim 13, wherein, The processing the operation data of the air conditioner into a target data type includes: When the operation data of the air conditioner is first data, perform weighted averaging on the operation data of the air conditioner according to the weight value corresponding to the operation data of the air conditioner, so that the operation data of the air conditioner matches the target data type; wherein, the weight value is used to represent the on or off state of the air conditioner, and the first data includes temperature data and pressure data; When the operation data of the air conditioner is second data, add the operation data of the air conditioner, so that the operation data of the air conditioner matches the target data type; wherein, the second data includes other operation data except the temperature data and the pressure data.

15. The method according to any one of claims 9 to 14, wherein The operation data of the air conditioner includes at least one of temperature data, pressure data, valve opening data, frequency data, air volume level data, and current data.

16. The method according to any one of claims 9 to 15, wherein, The method further includes: For the fault of the air conditioner in the second dimension, when the second fault information is fault information, determine the diagnosis result of the air conditioner as a fault result; For the fault of the air conditioner in the first dimension, when at least one of the first fault information and the second fault information is the fault information, determine the diagnosis result of the air conditioner as the fault result.

17. A cloud server, comprising: A communicator configured to communicate with a first multi-connected air conditioner unit and a second multi-connected air conditioner unit; And A processor connected to the communicator and configured to: Obtain a first sample set and a second sample set; the first sample set includes a plurality of first samples, and the plurality of first samples are generated according to the first operation data of the first multi-connected air conditioner unit; the second sample set includes a plurality of second samples, and the plurality of second samples are generated according to the first operation data and the second operation data of the second multi-connected air conditioner unit; any one of the plurality of second samples has a first label or a second label; the first label is configured to indicate the operation mode corresponding to the second sample; the second label is configured to indicate the fault type corresponding to the second sample; When the proportion of the target sample in the second sample set is less than or equal to a preset threshold, perform an expansion process on the second sample set to adjust the proportion of the target sample in the second sample set; The target sample is the second sample having the first label or the second label; Adjust the first sample set according to the feature space distribution of the second sample set; Train an initial model based on the first sample set and the second sample set to obtain a trained fault diagnosis model; Use the fault diagnosis model to perform fault diagnosis on the first multi-connected air conditioner unit.

18. The cloud server according to claim 17, wherein, The processor is further configured to: obtain at least one neighboring sample of the target sample; Perform random linear interpolation processing between the target sample and any neighboring sample in the at least one neighboring sample respectively to generate a new target sample; Add the new target sample to the second sample set to adjust the proportion of the target sample in the second sample set.

19. The cloud server according to claim 17 or 18, wherein, The processor is further configured to: input the first sample set and the second sample set into a domain adaptation model to adjust the first sample set; the domain adaptation model adopts an adversarial neural network structure.

20. The cloud server according to any one of claims 17 to 19, wherein, The processor is further configured to: Perform an extraction operation; the extraction operation includes the processor extracting a first feature from the first sample through a feature extractor, extracting a second feature from a second sample generated from the first operation data of the second multi-connected air conditioner unit, and extracting a third feature from a second sample generated from the second operation data of the second multi-connected air conditioner unit; Perform an input operation; the input operation includes the processor inputting the first feature and the second feature into a domain classifier to obtain a domain loss function, and inputting the third feature into a label classifier to obtain a label loss function; Perform an update operation; the update operation includes the processor updating the parameters of the feature extractor, the domain classifier, and the label classifier through a backpropagation algorithm according to the domain loss function and The label loss function; Repeat the input operation and the update operation until a convergence condition is reached; the convergence condition includes at least one of the following: the value of the domain classification loss function of the domain classifier is greater than or equal to a first threshold, and the value of the label loss function of the label classifier is less than or equal to a second threshold; or, the number of iterations reaches a preset number.

21. The cloud server according to any one of claims 17 to 20, wherein, The processor is further configured to: Obtain first test data of the first multi-connected air conditioner unit and second test data of the second multi-connected air conditioner unit; the first test data includes the first operation data of the first multi-connected air conditioner unit, and the second test data includes the first operation data and the second operation data of the second multi-connected air conditioner unit; Preprocess the first test data and the second test data to obtain preprocessed first test data and second test data; The preprocessing includes at least one of data cleaning, data reduction, data encoding, and data normalization; Generate the first sample set and the second sample set based on the preprocessed first test data and second test data.

22. The cloud server according to claim 21, wherein, The first test data includes at least one of the following: the operating conditions, load rate, set temperature, and air volume level of the first multi-connected air conditioner unit under different test environments; the second test data includes at least one of the following: the operating conditions, load rate, set temperature, and air volume level of the second multi-connected air conditioner unit under different test environments.

23. A first multi-connected air conditioner unit, comprising: at least one indoor unit; at least one outdoor unit; and a controller configured to: obtain a first sample set and a second sample set; the first sample set includes a plurality of first samples, and the plurality of first samples are generated according to the first operation data of the first multi-connected air conditioner unit; the second sample set includes a plurality of second samples, and the plurality of second samples are generated according to the first operation data and the second operation data of the second multi-connected air conditioner unit; any one of the plurality of second samples has a first label or a second label; the first label is configured to indicate the operation mode corresponding to the second sample; the second label is configured to indicate the fault type corresponding to the second sample; in the case where the proportion of the target sample in the second sample set is less than or equal to a preset threshold, perform an expansion process on the second sample set to adjust the proportion of the target sample in the second sample set; the target sample is the second sample having the first label or the second label; adjust the first sample set according to the feature space distribution of the second sample set; train an initial model according to the first sample set and the second sample set to obtain a trained fault diagnosis model; use the fault diagnosis model to perform fault diagnosis on the first multi-connected air conditioner unit.

24. The first multi-connected air conditioner unit according to claim 23, wherein, The controller is further configured to: obtain at least one nearest neighbor sample of the target sample; perform random linear interpolation processing between the target sample and any one of the at least one nearest neighbor sample to generate a new sample having the first label or the second label; add the new sample having the first label or the second label to the second sample set to adjust the proportion of the target sample in the second sample set.

25. The first multi-connected air conditioner unit according to claim 23 or 24, wherein, The controller is further configured to: input the first sample set and the second sample set into a domain adaptation model to adjust the first sample set; the domain adaptation model adopts an adversarial neural network structure.

26. The first multi-connected air conditioner unit according to any one of claims 23 to 25, wherein, The controller is further configured to: perform an extraction operation; the extraction operation includes the controller extracting a first feature from the first sample through a feature extractor, extracting a second feature from a second sample generated according to the first operation data of the second multi-connected air conditioner unit, and extracting a third feature from a second sample generated according to the second operation data of the second multi-connected air conditioner unit; perform an input operation; the input operation includes the controller inputting the first feature and the second feature into a domain classifier to obtain a domain loss function, and inputting the third feature into a label classifier to obtain a label loss function; Perform an update operation: The update operation includes the controller updating the parameters of the feature extractor, the domain classifier, and the label classifier according to the domain loss function and the label loss function through the backpropagation algorithm; Repeat the input operation and the update operation until the domain classification loss function value of the domain classifier is maximized and the label loss function value of the label classifier is minimized.

Images