Air conditioner fault detection method and device, air conditioner, equipment and product

Abstract

The application provides an air conditioner fault detection method, device, air conditioner, equipment and product, which belong to the air conditioner technology, and the method comprises the following steps: acquiring a target operation parameter of an air conditioner collected historically; inputting the target operation parameter into a prediction model to obtain an energy consumption prediction value output by the prediction model; and performing fault detection on the air conditioner based on the energy consumption prediction value, an energy consumption relative threshold value and an actual energy consumption value collected currently. According to the scheme, the operation parameter that is strongly related to energy consumption is selected, the prediction model is constructed, the fault detection is converted into monitoring of energy consumption abnormalities, the dependence on specific fault simulation experiments and data labeling can be reduced, the energy consumption relative threshold value is adaptively determined according to the relative relationship between the prediction value and the actual value in the test stage, the adaptability of the prediction model under different operation conditions is improved, and therefore, the cost of fault detection is reduced.

Description

Technical Field

[0001] This application relates to air conditioning technology. More specifically, it relates to an air conditioning fault detection method, apparatus, air conditioner, equipment, and product. Background Technology

[0002] Multi-split air conditioning systems are widely used in commercial buildings such as shopping malls, office buildings, industrial plants, and data centers. These systems have complex structures and are susceptible to performance degradation or soft failures due to factors such as ambient temperature and humidity, component aging, refrigerant leaks, and abnormal valve openings. This can result in a situation where the system can still operate but with reduced efficiency.

[0003] Current fault detection methods for multi-split air conditioning systems primarily involve artificially creating faults in a laboratory setting, collecting operational data under fault conditions, and then inputting the normal and fault data into a classification model to train a fault discrimination model. However, this approach requires simulating multiple fault types, each necessitating the design of independent experiments, which is time-consuming and costly. Furthermore, the laboratory environment cannot fully replicate real-world operating scenarios. Therefore, the current challenge is to reduce the cost of fault detection. Summary of the Invention

[0004] This application provides an air conditioning fault detection method, device, air conditioner, equipment, and product, which can reduce the cost of fault detection.

[0005] In a first aspect, embodiments of this application provide an air conditioning fault detection method, including:

[0006] The system acquires historically collected target operating parameters of the air conditioner, which are the parameters with the highest correlation to the air conditioner's energy consumption under various collection conditions, including operating conditions. These target operating parameters are then input into a prediction model to obtain the predicted energy consumption value. The prediction model is pre-trained and tested using a training set, which includes the target operating parameters and corresponding energy consumption parameters under different collection conditions. Based on the predicted energy consumption value, a relative energy consumption threshold, and the currently collected actual energy consumption value, fault detection is performed on the air conditioner. The relative energy consumption threshold is determined based on the relative magnitude of the predicted energy consumption value and the actual energy consumption value output by the prediction model during the testing phase.

[0007] In some embodiments of this application, when the air conditioner is a multi-split air conditioner, the data collection conditions also include the number of indoor and outdoor units being turned on; the method further includes: acquiring multiple operating parameters and corresponding energy consumption parameters of the air conditioner under different data collection conditions; wherein, the operating parameters include environmental data and air conditioner status data; the environmental data includes at least one of the following: indoor and outdoor temperature, indoor and outdoor relative humidity, and indoor and outdoor pressure; the status data includes at least one of the following: air conditioner temperature, pressure, frequency, current, expansion valve opening, air volume level, compressor power, subcooling degree, and exhaust superheat degree; classifying the energy consumption parameters according to the data collection conditions, and classifying the operating parameters under each data collection condition according to the maximum and minimum values ​​of the operating parameters; performing correlation analysis on the data pairs formed by the classified operating parameters and the corresponding energy consumption parameters, and taking the operating parameter with the highest correlation to the energy consumption parameter as the target operating parameter.

[0008] In some embodiments of this application, the air conditioner's operating parameters and corresponding energy consumption parameters are obtained under different acquisition conditions, including: when the air conditioner is operating without faults, the air conditioner's operating parameters and corresponding energy consumption parameters are obtained when the acquisition conditions change; and / or, when the air conditioner is operating without faults, the acquisition conditions are changed and the air conditioner's operating parameters and corresponding energy consumption parameters are obtained.

[0009] In some embodiments of this application, the operating parameters under each acquisition condition are classified according to their maximum and minimum values. This includes: obtaining a data set of the first operating parameter under all historical acquisition conditions; dividing the data points of the first operating parameter into multiple categories based on the distribution characteristics of the data points in the data set using a clustering analysis algorithm, wherein each category corresponds to a data interval, and the boundaries of the data intervals of different categories are determined by the clustering process; sequentially determining the clustering category to which the data points of the first operating parameter belong under each acquisition condition, and using the belonged category as the category of the first operating parameter under that acquisition condition; and sequentially determining the categories of other operating parameters under each acquisition condition based on the classification logic of the first operating parameter.

[0010] In some embodiments of this application, the method further includes: during the testing phase of the prediction model, calculating the difference between the predicted energy consumption value and the actual energy consumption value output by the prediction model, and the ratio of this difference to the actual energy consumption value; and using the sum of the maximum ratio obtained during the testing phase and multiples of the differences between the maximum and minimum ratios as the relative energy consumption threshold; where N is a positive integer.

[0011] In some embodiments of this application, the air conditioner is fault detected based on the predicted energy consumption value, the relative energy consumption threshold, and the currently collected actual energy consumption value. This includes: calculating the product of the predicted energy consumption value and the relative energy consumption threshold, and using the difference between the predicted energy consumption value and the product as the left endpoint of the fault interval, and using the sum of the predicted energy consumption value and the product as the right endpoint of the fault interval; if the actual energy consumption value is outside the fault interval, the air conditioner is determined to be faulty.

[0012] Secondly, embodiments of this application provide an air conditioning fault detection device, comprising:

[0013] The acquisition module is used to acquire historically collected target operating parameters of the air conditioner. These target operating parameters are those with the highest correlation to the air conditioner's energy consumption parameters under various acquisition conditions, including operating conditions. The prediction module is used to input the target operating parameters into the prediction model to obtain the predicted energy consumption value output by the prediction model. The prediction model is pre-trained and tested based on a training set, which includes the target operating parameters and corresponding energy consumption parameters under different acquisition conditions. The detection module is used to perform fault detection on the air conditioner based on the predicted energy consumption value, the relative energy consumption threshold, and the currently collected actual energy consumption value. The relative energy consumption threshold is determined based on the relative magnitude of the predicted energy consumption value and the actual energy consumption value output by the prediction model during the testing phase.

[0014] Thirdly, embodiments of this application provide an air conditioner, including the air conditioner fault detection device described above.

[0015] Fourthly, embodiments of this application provide an electronic device, including:

[0016] The processor, and the memory that is in communication with the processor;

[0017] The memory stores the instructions that the computer executes;

[0018] The processor executes computer-executable instructions stored in memory to achieve the method described above.

[0019] Fifthly, embodiments of this application provide a computer-readable storage medium storing computer-executable instructions, which, when executed by a processor, implement the aforementioned method.

[0020] The air conditioner fault detection method, apparatus, air conditioner, equipment, and product provided in this application include the following method: acquiring historically collected target operating parameters of the air conditioner; inputting the target operating parameters into a prediction model to obtain the energy consumption prediction value output by the prediction model; and performing fault detection on the air conditioner based on the energy consumption prediction value, the relative energy consumption threshold, and the currently collected actual energy consumption value. The solution in this application, by selecting operating parameters strongly correlated with energy consumption and constructing a prediction model, transforms fault detection into monitoring of energy consumption anomalies. This reduces the reliance on specific fault simulation experiments and data labeling. Simultaneously, by adaptively determining the relative energy consumption threshold based on the relative relationship between the predicted and actual values ​​during the testing phase, it improves the adaptability of the prediction model under different operating conditions, thereby reducing the cost of fault detection. Attached Figure Description

[0021] To more clearly illustrate the implementation methods in the embodiments of this application or related technologies, the accompanying drawings used in the description of the embodiments or related technologies will be briefly introduced below. Obviously, the accompanying drawings described below are some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings.

[0022] Figure 1 This is a schematic diagram of a multi-split air conditioner operation scenario provided in an embodiment of this application.

[0023] Figure 2 An exemplary block diagram of the configuration of the control device according to an exemplary embodiment is shown.

[0024] Figure 3 This is a schematic diagram of the structure of a multi-split air conditioner provided in an embodiment of this application.

[0025] Figure 4 This is a flowchart illustrating an air conditioner fault detection method provided in this application.

[0026] Figure 5 This is a flowchart illustrating an air conditioner fault detection method provided in this application.

[0027] Figure 6 This is a flowchart illustrating an air conditioner fault detection method provided in this application.

[0028] Figure 7 This is a flowchart illustrating an air conditioner fault detection method provided in this application.

[0029] Figure 8 This is a flowchart illustrating an air conditioner fault detection method provided in this application.

[0030] Figure 9 This is a flowchart illustrating an air conditioner fault detection method provided in this application.

[0031] Figure 10 This is a schematic diagram of the structure of an air conditioning fault detection device provided in this application.

[0032] Figure 11 This is a schematic diagram of the structure of an electronic device provided in an embodiment of this application. Detailed Implementation

[0033] To make the objectives, implementation methods and advantages of this application clearer, the exemplary implementation methods of this application will be clearly and completely described below with reference to the accompanying drawings of the exemplary embodiments of this application. Obviously, the described exemplary embodiments are only some embodiments of this application, and not all embodiments.

[0034] It should be noted that the brief descriptions of terms in this application are only for the convenience of understanding the embodiments described below, and are not intended to limit the embodiments of this application. Unless otherwise stated, these terms should be understood in their ordinary and common meaning.

[0035] Furthermore, the terms “comprising” and “having”, and any variations thereof, are intended to cover but not exclusively include, for example, a product or device that includes a series of components is not necessarily limited to those that are explicitly listed, but may include other components that are not explicitly listed or that are inherent to such product or device.

[0036] Multi-split air conditioning systems are central air conditioning systems that connect multiple indoor units to a single outdoor unit. They are widely used in commercial buildings such as office buildings, shopping malls, industrial plants, data centers, hospitals, and cold chain logistics, where high temperature control stability is required. Their core advantages lie in energy efficiency and spatial adaptability, enabling independent temperature control in multiple zones through flexible refrigerant piping connections. However, the complexity of multi-split systems, including components such as compressors, condensers, evaporators, and four-way valves, makes them susceptible to changes in ambient temperature and humidity, component aging, refrigerant leaks, and abnormal valve openings during operation. This can lead to performance degradation or soft failures (in which case the system can still operate but with reduced efficiency). For example, in commercial buildings, condenser fouling or refrigerant leaks can cause a sharp increase in energy consumption; in industrial production environments, decreased compressor efficiency can lead to abnormal equipment precision or reduced product yield; and in cold chain logistics, system failures can cause food spoilage, resulting in economic losses.

[0037] Related fault detection technologies typically involve artificially creating faults in a laboratory setting, such as compressor wear, refrigerant leakage, or valve jamming, to collect operational data under these fault conditions. The normal operating data and fault data are then input into a classification model, such as SVM, random forest, or neural network, to train a fault discrimination model. Finally, real-time monitoring of parameters (such as current, pressure, and temperature) determines whether a fault has occurred.

[0038] However, this approach is costly to implement: it requires simulating multiple fault types, each requiring independent testing, resulting in long experimental cycles and high expenses. Furthermore, the laboratory environment differs from actual operating scenarios, exhibiting variations in temperature and humidity, and sudden load changes, making it difficult for the model to adapt to complex operating conditions and limiting its generalization ability. Additionally, this approach can only detect the fault types simulated in the experiments, failing to cover all faults affecting energy consumption, and fault data requires manual annotation; moreover, some soft faults, such as slow degradation of component performance, are difficult to capture through short-term experiments.

[0039] This application provides an air conditioner fault detection method, apparatus, air conditioner, equipment, and product. The method includes: acquiring historically collected target operating parameters of the air conditioner; inputting the target operating parameters into a prediction model to obtain the energy consumption prediction value output by the prediction model; and performing fault detection on the air conditioner based on the energy consumption prediction value, a relative energy consumption threshold, and the currently collected actual energy consumption value. The solution in this application, by selecting operating parameters strongly correlated with energy consumption and constructing a prediction model, transforms fault detection into monitoring of energy consumption anomalies. This reduces the reliance on specific fault simulation experiments and data labeling. Simultaneously, it adaptively determines the relative energy consumption threshold based on the relative relationship between the predicted and actual values ​​during the testing phase, improving the adaptability of the prediction model under different operating conditions, thereby reducing the cost of fault detection.

[0040] The technical solutions of this application will be described in detail below with reference to specific embodiments. The following specific embodiments can be combined with each other, and the same or similar concepts or processes may not be described again in some embodiments.

[0041] Below, in conjunction with Figures 1 to 3 The multi-split air conditioning system provided in the embodiments of this application will be described.

[0042] Figure 1 This is a schematic diagram illustrating a multi-split air conditioner operation scenario provided in an embodiment of this application. Figure 1 As shown, users can operate the multi-split air conditioner 200 through the smart device 300 or the control device 100.

[0043] In some embodiments, the control device 100 may be a remote control. Communication between the remote control and the display device includes infrared protocol communication, Bluetooth protocol communication, and other short-range communication methods, controlling the multi-split air conditioner 200 wirelessly or via wired means. Users can control the multi-split air conditioner 200 by inputting user commands through buttons on the remote control, voice input, control panel input, etc.

[0044] In some embodiments, a smart device 300 (such as a mobile terminal, tablet computer, computer, laptop computer, etc.) can also be used to control the multi-split air conditioner 200. For example, an application running on the smart device can be used to control the multi-split air conditioner 200.

[0045] In some embodiments, multi-split air conditioners may receive commands without using the aforementioned smart devices or control devices, but instead receive user control via touch or gestures.

[0046] In some embodiments, the multi-split air conditioner 200 can also be controlled in ways other than the control device 100 and the smart device 300. For example, it can be controlled by directly receiving the user's voice commands through a module configured inside the multi-split air conditioner 200, or it can be controlled by receiving the user's voice commands through a voice control device set outside the multi-split air conditioner 200.

[0047] In some embodiments, the multi-split air conditioner 200 also communicates with the server 400. The multi-split air conditioner 200 may communicate via a local area network (LAN), wireless local area network (WLAN), and other networks. The server 400 may provide various content and interactive features to the multi-split air conditioner 200. The server 400 may be a cluster or multiple clusters, and may include one or more types of servers.

[0048] Figure 2 An exemplary block diagram of the configuration of the control device according to an exemplary embodiment is shown. Figure 2 As shown, the control device 100 includes a controller 110, a communication interface 130, a user input / output interface 140, a memory, and a power supply. The control device 100 can receive user input operation commands and convert the operation commands into commands that the multi-split air conditioner 200 can recognize and respond to, thus acting as an intermediary for interaction between the user and the multi-split air conditioner 200.

[0049] Figure 3 This is a structural schematic diagram of a multi-split air conditioner provided as an embodiment of this application. Figure 3 As shown, the multi-split air conditioner 200 may include an indoor unit 220 and an outdoor unit 210.

[0050] The indoor unit 220 is installed indoors and is used for heat exchange with the indoor environment.

[0051] The outdoor unit 210 is installed outdoors to carry indoor heat to the outside.

[0052] like Figure 3 As shown, the indoor unit 220 includes multiple indoor units 221 connected in parallel, and the multiple indoor units 221 are connected to the outdoor unit 210. The multiple indoor units 221 are respectively installed in different rooms.

[0053] Indoor unit 221 can include, but is not limited to, wall-mounted, ducted, and curtain units.

[0054] Furthermore, the multi-split air conditioning system may include a controller. The controller is electrically connected to the indoor and outdoor units to control the operation of their internal components, so that the various components of the multi-split air conditioning system can perform their predetermined functions, including receiving user commands, operating in cooling mode, heating mode, fan mode, etc., and uploading the operating status of the multi-split air conditioning system to the cloud.

[0055] The controller includes an indoor controller and an outdoor controller. The indoor controller is installed in the indoor unit. The indoor controller is used to control the operating status of various components inside the indoor unit. The outdoor controller is used to control the operating status of various components inside the outdoor unit.

[0056] The indoor controller and the outdoor controller communicate via wired or wireless communication.

[0057] In some embodiments of this application, no specific distinction is made between indoor controllers and outdoor controllers; both indoor controllers and outdoor controllers are collectively referred to as controllers.

[0058] The following explanation uses the processor of a multi-split air conditioner as an example to illustrate the automatic control strategy of a multi-split air conditioner.

[0059] The technical solutions of this application will be described in detail below with reference to specific embodiments. The following specific embodiments can be combined with each other, and the same or similar concepts or processes may not be described again in some embodiments.

[0060] Figure 4 This is a flowchart illustrating an air conditioner fault detection method provided in this application. Figure 4 As shown, the method includes the following steps:

[0061] S101. Obtain the target operating parameters of the air conditioner collected in history. The target operating parameters are the parameters that have the highest correlation with the energy consumption parameters of the air conditioner under various different collection conditions. The collection conditions include the operating conditions.

[0062] S102. Input the target operating parameters into the prediction model to obtain the energy consumption prediction value output by the prediction model; wherein, the prediction model is pre-trained and tested based on the training set, which includes the target operating parameters and corresponding energy consumption parameters under different collection conditions;

[0063] S103. Based on the predicted energy consumption value, the relative energy consumption threshold, and the currently collected actual energy consumption value, perform fault detection on the air conditioner; wherein, the relative energy consumption threshold is determined according to the relative magnitude of the predicted energy consumption value and the actual energy consumption value output by the prediction model during the testing phase.

[0064] It should be noted that the embodiments of this application can be applied to multi-split air conditioning systems, as well as other air conditioning systems such as central air conditioning systems, unit air conditioning systems, and household split air conditioning systems. This application does not limit these applications.

[0065] For example, the target operating parameters collected in history may come from long-term monitoring data records from the air conditioner's built-in sensors, historical operating data of the unit stored in the cloud, or maintenance and operation datasets accumulated by the maintenance platform.

[0066] In some embodiments, correlation analysis (such as Pearson correlation coefficient calculation) can be used to screen out the set of parameters with the strongest statistical correlation to energy consumption parameters from a large number of operating parameters. Specifically, the screening process can be carried out in a data pool that combines all collection conditions to obtain globally highly correlated parameters; or it can be analyzed separately for specific collection conditions to obtain highly correlated parameters specific to that condition. For example, target operating parameters that may be generally highly correlated with energy consumption under various conditions include compressor operating frequency, indoor-outdoor temperature difference, and condenser tube temperature; while under specific high temperature and high humidity conditions, the running time of dehumidification mode or fan speed may become target operating parameters with prominent correlation under that condition.

[0067] For example, operating conditions mainly include the air conditioner's operating mode (such as cooling, heating, dehumidifying, and air supply), set temperature, fan speed, and other system-specific operating status settings. Optionally, in addition to operating conditions, the data collection conditions may also include environmental conditions (such as indoor and outdoor temperature, humidity, and sunlight intensity) and equipment status conditions (such as the air conditioner's service life, filter cleanliness, and refrigerant charge history).

[0068] In some embodiments, the predictive model can be a variety of machine learning or statistical models. For example, one possibility is a multiple linear regression model, which uses multiple target operating parameters as input features and learns the weight coefficients of each parameter through training to output a predicted energy consumption value. Another possibility is a neural network model, such as a Rongguo multilayer perceptron, which captures a more complex nonlinear relationship between target operating parameters and energy consumption. During the training phase, the predictive model uses a large amount of historical data in the training set (containing target operating parameters under different acquisition conditions and corresponding actual energy consumption values) to adjust its internal parameters to minimize prediction error. During the testing phase, another set of data not used in training is used to evaluate the model's generalization performance and verify its predictive accuracy.

[0069] In this example, the training set includes target operating parameters and corresponding energy consumption parameters under different data collection conditions. This means the training data needs to cover as many operating scenarios as possible that the air conditioner might encounter. For example, the training set should include data on high summer temperatures and low winter temperatures, data on high-load cooling and low-load nighttime operation, and data on filter cleaning and minor clogging conditions. Such a training set helps the model learn the mapping relationship between target operating parameters and energy consumption under various data collection conditions, thus enabling it to make relatively robust energy consumption predictions when faced with new and unseen combinations of operating conditions.

[0070] In some embodiments, when the model passes the testing phase validation, a relative energy consumption threshold is determined by calculating the distribution of the deviation between the model's predicted energy consumption and the actual energy consumption. This threshold is not a fixed theoretical value, but rather derived statistically based on the performance of the test data. For example, the relative error of "(actual value - predicted value) / predicted value" for all samples in the test set can be calculated, and then a high quantile (such as the 95th percentile) of this relative error distribution can be taken as the threshold. This indicates that during the testing phase, the model predicts that in 95% of cases, the actual value does not exceed the "allowable range calculated by adding the predicted value to the relative threshold". The threshold determined by the above method can quantify the predictive uncertainty of the model under normal conditions.

[0071] In some embodiments, when fault detection is performed based on predicted energy consumption, a relative energy consumption threshold, and the current actual energy consumption, the relative deviation between the current actual energy consumption and the predicted energy consumption output by the model can be calculated in real time. If this deviation continues to exceed a preset relative energy consumption threshold, a fault warning is triggered. For example, if the threshold is 15%, and at a certain moment the predicted value is 1 kWh and the actual value is 1.2 kWh, the relative deviation is 20% (>15%), which may be determined as an abnormal energy consumption.

[0072] In some optional embodiments, a normal energy consumption range (e.g., [predicted value × (1 - threshold), predicted value × (1 + threshold)]) can be calculated based on the predicted energy consumption value and the relative energy consumption threshold, and the actual energy consumption value can be continuously monitored to see if it falls within this range. If the actual value exceeds the upper limit of the range for several consecutive cycles (indicating low energy efficiency), it may indicate a fault in the system such as refrigerant leakage or heat exchanger blockage; conversely, if it remains below the lower limit of the range, it may also suggest abnormal sensor readings or other control problems. Both of the above methods achieve preliminary fault identification through quantitative comparison.

[0073] The air conditioner fault detection method provided in this application includes: acquiring historically collected target operating parameters of the air conditioner; inputting the target operating parameters into a prediction model to obtain the energy consumption prediction value output by the prediction model; and performing fault detection on the air conditioner based on the energy consumption prediction value, the relative energy consumption threshold, and the currently collected actual energy consumption value. This solution, by selecting operating parameters strongly correlated with energy consumption and constructing a prediction model, transforms fault detection into monitoring of energy consumption anomalies. This reduces the reliance on specific fault simulation experiments and data labeling. Furthermore, by adaptively determining the relative energy consumption threshold based on the relative relationship between the predicted and actual values ​​during the testing phase, it improves the adaptability of the prediction model under different operating conditions, thereby reducing the cost of fault detection.

[0074] Figure 5 This is a flowchart illustrating an air conditioner fault detection method provided in this application. When the air conditioner is a multi-split air conditioner, the data collection conditions also include the number of indoor and outdoor units currently in operation; for example... Figure 5 As shown, the method also includes the following steps:

[0075] S201. Acquire multiple operating parameters and corresponding energy consumption parameters of the air conditioner under different acquisition conditions; wherein, the operating parameters include environmental data and air conditioner status data; the environmental data includes at least one of the following: indoor and outdoor temperature, indoor and outdoor relative humidity, and indoor and outdoor pressure; the status data includes at least one of the following: air conditioner temperature, pressure, frequency, current, expansion valve opening, air volume level, compressor power, subcooling degree, and exhaust superheat degree.

[0076] S202. Classify energy consumption parameters according to the acquisition conditions, and classify operating parameters under each acquisition condition according to the maximum and minimum values ​​of the operating parameters.

[0077] S203. Based on the data pairs consisting of the categorized operating parameters and the corresponding energy consumption parameters, perform correlation analysis and select the operating parameters with the highest correlation to the energy consumption parameters as the target operating parameters.

[0078] For example, for a multi-split air conditioner, it is first necessary to obtain multiple operating parameters and corresponding energy consumption parameters under different data collection conditions. For instance, when the data collection conditions are cooling mode, indoor set temperature 26 degrees Celsius, high fan speed, and 4 units operating (indoor and outdoor), the operating parameters might include outdoor temperature 38 degrees Celsius, indoor temperature 28 degrees Celsius, and indoor relative humidity 55% from environmental data, and compressor frequency 50 Hz, condensing pressure 1.8 MPa, evaporator temperature 10 degrees Celsius, total current 12 amps, electronic expansion valve opening 70%, and compressor power 3 kW from status data; the corresponding energy consumption parameter might be a total energy consumption of 4.2 kWh for that period. Another data collection condition might be heating mode, indoor set temperature 22 degrees Celsius, low fan speed, and 2 units operating (indoor and outdoor). In this case, the operating parameters might include outdoor temperature 5 degrees Celsius, indoor temperature 20 degrees Celsius, compressor frequency 40 Hz, and exhaust superheat 15 degrees Celsius, etc., and the corresponding energy consumption parameter might be 3.0 kWh. This data typically comes from the historical operating database of the air conditioning system or a real-time monitoring platform, covering a variety of typical and edge operating scenarios.

[0079] In some embodiments, energy consumption parameters are categorized according to the acquisition conditions. This means combining the operating conditions (e.g., cooling, heating) and the number of indoor and outdoor units in operation (e.g., 1, 2, etc.) in the acquisition conditions into discrete category labels, such as "4 units cooling" or "2 units heating". Then, the energy consumption parameter data corresponding to each acquisition condition is assigned to the corresponding category, forming a subset of energy consumption data indexed by the acquisition conditions. Secondly, the operating parameters under each acquisition condition are categorized based on the maximum and minimum values ​​of the historical operating parameters under that condition. For example, under the acquisition condition of "4 units cooling", for the operating parameter of compressor frequency, if its historical maximum and minimum values ​​are 60 Hz and 30 Hz respectively, its numerical range can be divided into three categories such as "low frequency" (30-40 Hz), "medium frequency" (40-50 Hz), and "high frequency" (50-60 Hz). Other operating parameters such as condensing pressure and current are also divided into several categories according to their respective maximum and minimum values ​​in the same way, thereby converting continuous parameter values ​​into discrete category identifiers for easier subsequent analysis.

[0080] Furthermore, there are various ways to perform correlation analysis on the data pairs consisting of the categorized operating parameters and the corresponding energy consumption parameters.

[0081] For example, one implementation employs a statistical correlation calculation method, such as using Spearman's rank correlation coefficient, to analyze the correlation strength between each operating parameter category and the energy consumption parameter category. Under each acquisition condition, the correlation coefficient between all operating parameter category sequences (such as compressor frequency category sequences) and energy consumption parameter category sequences is calculated, and the operating parameter with the highest absolute value of the correlation coefficient is determined as the target operating parameter under that condition. For example, under the condition of "4 units of cooling", it may be found that the compressor frequency category and the energy consumption category have the strongest correlation.

[0082] Alternatively, another implementation method is to evaluate the predictive ability of each operating parameter category to the energy consumption parameter category based on the information gain or mutual information index in information theory; by calculating the reduction in information entropy of the energy consumption parameter category after each operating parameter category is divided, the operating parameter with the largest information gain is selected as the target operating parameter. For example, under the condition of "heating 2 units", it may be found that the exhaust superheat category provides the highest information gain, so it is selected as the target operating parameter.

[0083] The solution in this example categorizes operating parameters and energy consumption parameters according to the collection conditions, and then uses correlation analysis to filter out the core operating parameters that are most relevant to energy consumption under different conditions. This approach can improve the identification accuracy of key operating parameters by taking into account the variable load characteristics of multi-split air conditioners through conditional category analysis and correlation filtering, thereby improving the accuracy of the prediction model in fault detection under complex operating conditions.

[0084] Figure 6 This is a flowchart illustrating an air conditioner fault detection method provided in this application. Figure 6 As shown, S201 acquires the air conditioner's operating parameters and corresponding energy consumption parameters under different acquisition conditions, including:

[0085] S301. When the air conditioner is running without faults, acquire the air conditioner's operating parameters and corresponding energy consumption parameters when the data acquisition conditions change.

[0086] And / or,

[0087] S302. When the air conditioner is running without faults, change the data collection conditions and obtain the air conditioner's operating parameters and corresponding energy consumption parameters.

[0088] In some embodiments, the operating parameters and corresponding energy consumption parameters are acquired when the acquisition conditions change. This is mainly achieved by long-term monitoring and recording of data generated by spontaneous changes in acquisition conditions caused by external environment or user behavior during the natural operation of the air conditioner.

[0089] For example, during a trouble-free summer operation, the data collection conditions for a multi-split air conditioner on a certain day might be "cooling mode, outdoor temperature 35 degrees Celsius, relative humidity 70%, 3 indoor units on." Operating parameters at this moment, such as compressor frequency, total current, and condensing pressure, can be collected, and the corresponding instantaneous power or time-period energy consumption parameters can be recorded. Subsequently, when the weather turns to thunderstorms, the data collection conditions naturally change to "cooling mode, outdoor temperature 28 degrees Celsius, relative humidity 95%, still 3 indoor units on," and a new set of operating parameters and energy consumption parameters under the new conditions can be collected again. This implementation relies on a continuous data monitoring system that passively captures and accumulates data pairs that naturally form over time, with seasonal changes, and user operations (such as adjusting temperature settings and turning indoor units on and off), covering multiple data collection conditions, thereby reflecting the correlations under real operating conditions.

[0090] In some embodiments, changing the acquisition conditions and obtaining the operating parameters and corresponding energy consumption parameters involves active testing or controlled experiments.

[0091] For example, after the air conditioner is installed and commissioned, or during routine maintenance, technicians can, provided the equipment is fault-free, proactively and systematically change one or more data collection conditions in a specific environmental simulation chamber or a selected actual environment. For instance, they can first set the air conditioner to "heating mode, set temperature 24 degrees Celsius, two indoor units on," and record a set of operating parameters and energy consumption parameters after a period of stable operation. Then, they can proactively change the data collection conditions to "heating mode, set temperature 20 degrees Celsius, four indoor units on," and record a second set of data after the system stabilizes. They can further proactively change environmental conditions, such as adjusting the humidity in the environmental simulation chamber, and record a third set of data. This approach is a proactive data production method designed to systematically and efficiently construct a dataset that covers a more comprehensive and balanced combination of data collection conditions. It is particularly effective in supplementing data from edge or extreme operating conditions that are less common in natural operation, ensuring that the data foundation for subsequent analysis is sufficiently representative.

[0092] This example solution ensures that all basic data used for correlation analysis comes from the fault-free operation of the air conditioner, and combines two data acquisition strategies: passively recording natural changes and actively intervening in conditions. This lays a data foundation for accurately screening target operating parameters and establishing a reliable energy consumption prediction model.

[0093] Figure 7 This is a flowchart illustrating an air conditioner fault detection method provided in this application. Figure 7 As shown in S202, the operating parameters under each acquisition condition are categorized based on their maximum and minimum values, including:

[0094] S401. Obtain the data set of the first operating parameter under all historical acquisition conditions. Based on the distribution characteristics of the data points in the data set, divide the data points of the first operating parameter into multiple categories through a clustering analysis algorithm. Each category corresponds to a data interval, and the boundary of the data intervals of different categories is determined by the clustering process.

[0095] S402. Sequentially determine the cluster category to which the data points of the first operating parameter belong under each acquisition condition, and use the category to which they belong as the category of the first operating parameter under that acquisition condition.

[0096] S403. Based on the category classification logic of the first operating parameter, determine the category of other operating parameters under each acquisition condition in turn.

[0097] In some embodiments, the clustering analysis algorithm can be K-means, Gaussian mixture model clustering, or density-based clustering. The boundaries of the intervals are not simply divided equally based on the maximum and minimum values ​​or subjectively set, but are automatically formed by the clustering algorithm after analyzing the distribution characteristics of all historical data points.

[0098] For example, when using the K-means algorithm, the algorithm iteratively finds multiple cluster centers based on the Euclidean distance between data points. In the end, each data point will be assigned to the category of the cluster center closest to it, and the midpoint of adjacent cluster centers or the boundary point determined based on the data distribution density will naturally form the data interval boundary corresponding to different categories.

[0099] To illustrate with a specific example of the first operating parameter, let's assume it's the compressor frequency. First, we acquire a dataset containing all data points for the compressor frequency under all historical data collection conditions (e.g., different modes, different numbers of indoor and outdoor units operating). Next, we analyze this dataset using a clustering algorithm (e.g., K-means). The algorithm might automatically categorize these frequency values ​​into three classes: "low frequency," "medium frequency," and "high frequency." Each class corresponds to a specific data range (e.g., "low frequency" corresponds to 25-35 Hz, "medium frequency" to 35-50 Hz, and "high frequency" to 50-60 Hz). These range boundaries are calculated by the clustering algorithm based on the actual clustering of the data points. Furthermore, for each specific data collection condition (e.g., "cooling mode, 4 indoor units operating"), we determine which clustering range ("medium frequency") the compressor frequency data point collected under that condition (e.g., 42 Hz) falls into, thus defining its category.

[0100] Similarly, for the second operating parameter, such as condensing pressure, the same logic as that for compressor frequency is followed: first, based on all its historical data, the global category interval is determined by clustering (which may be divided into categories such as "low pressure", "medium pressure", "high pressure" etc.), and then the category to which the specific data point belongs is determined for each collection condition.

[0101] The solution presented in this example uses a data-driven clustering analysis method to classify operating parameters. This allows for a more objective and reasonable definition of category boundaries based on the actual distribution of parameter values, thereby improving the accuracy and interpretability of the physical meaning of the mapping between parameter categories and energy consumption parameter categories in subsequent correlation analysis.

[0102] Figure 8 This is a flowchart illustrating an air conditioner fault detection method provided in this application. Figure 8 As shown, the method also includes the following steps:

[0103] S501. During the testing phase of the prediction model, calculate the difference between the predicted energy consumption value and the actual energy consumption value output by the prediction model, and the ratio of the difference to the actual energy consumption value.

[0104] S502. The sum of the maximum ratio calculated during the testing phase and multiples of the difference between the maximum and minimum ratios is taken as the relative energy consumption threshold; where N is a positive integer.

[0105] For example, during the testing phase of the prediction model, for each sample in the test set, the relative error between its predicted energy consumption value and the corresponding actual energy consumption value is calculated. The specific formula is the ratio of the absolute value of (actual energy consumption - predicted energy consumption) to the actual energy consumption value. Alternatively, sometimes the ratio of (actual energy consumption - predicted energy consumption) directly to the actual energy consumption value is used to preserve the direction of the deviation. Furthermore, by iterating through all test samples, this ratio is calculated for each sample, thus obtaining a sequence of ratios.

[0106] Furthermore, the maximum value (denoted as Max_R) and minimum value (denoted as Min_R) are identified from this ratio sequence. The formula for calculating the relative energy consumption threshold (denoted as T) is: T = Max_R + N × (Max_R - Min_R). Here, N is a positive integer adjustment factor that needs to be preset. By adjusting the size of N, the leniency or strictness of the final threshold can be controlled. The threshold calculated in this way is not a simple maximum value, but rather an additional buffer proportional to the range of deviation fluctuations in the test data (i.e., the difference between the maximum and minimum values) added on top of the maximum observation deviation.

[0107] In practical applications, the determination of parameter N usually relies on engineering experience or optimization through historical data verification. For example, different N values ​​such as 1, 2, and 3 can be tried to observe the effect of distinguishing known normal data and known fault data under different thresholds. A value of N can be selected that can effectively capture real abnormal phenomena while ensuring a low false alarm rate (that is, misjudging normal fluctuations as faults).

[0108] The energy consumption relative threshold in this example not only depends on the maximum error observed during the testing phase, but also takes into account the dispersion of the model prediction error. This is equivalent to providing a tolerance band for the prediction uncertainty of the model under normal conditions, so that the fault judgment criteria can adapt to the performance fluctuations of the model itself and are more robust.

[0109] The scheme in this example determines the relative energy consumption threshold by introducing adaptive computational logic based on the statistical distribution (especially the range) of prediction errors during the testing phase. This enables the fault detection criteria to better match the actual prediction performance and uncertainty of the model, thereby helping to achieve more stable and adaptive anomaly identification in changing environments.

[0110] Figure 9 This is a flowchart illustrating an air conditioner fault detection method provided in this application. Figure 9 As shown, in S103, fault detection of the air conditioner is performed based on the predicted energy consumption value, the relative energy consumption threshold, and the currently collected actual energy consumption value, including:

[0111] S601. Calculate the product of the predicted energy consumption value and the relative energy consumption threshold, and take the difference between the predicted energy consumption value and the product as the left endpoint of the fault interval, and take the sum of the predicted energy consumption value and the product as the right endpoint of the fault interval.

[0112] S602. If the actual energy consumption value is outside the fault range, then the air conditioner is considered faulty.

[0113] In practical applications, it is necessary to obtain the predicted energy consumption value output by the prediction model for the current target operating parameters, obtain the determined relative energy consumption threshold from the analysis results of the historical test phase, and read the actual energy consumption value collected at present from the sensor in real time or determine it through calculation.

[0114] For example, the product of the predicted energy consumption value and the relative threshold of energy consumption is calculated, which represents the allowable absolute deviation based on the predicted value; then, the value of the predicted energy consumption value minus the product is used as the left endpoint of the fault interval, and the value of the predicted energy consumption value plus the product is used as the right endpoint of the fault interval, thereby forming a numerical interval symmetrical about the predicted energy consumption value, which defines the range in which the actual energy consumption value should be under normal conditions.

[0115] Furthermore, the currently collected actual energy consumption value is compared with the fault range: if the actual energy consumption value is less than the left endpoint or greater than the right endpoint, meaning the actual value falls outside this range, the system determines that the air conditioner is operating abnormally and has a fault; if the actual energy consumption value is greater than or equal to the left endpoint and less than or equal to the right endpoint, the system determines that the air conditioner is operating normally. This process achieves automatic fault identification based on quantized thresholds.

[0116] In some optional embodiments, another method for fault diagnosis based on a relative energy consumption threshold is the one-sided threshold comparison method. Specifically, the system calculates an allowable upper limit for energy consumption, i.e., upper limit = predicted energy consumption × (1 + relative energy consumption threshold). Then, it directly compares the currently collected actual energy consumption value with this upper limit: if the actual energy consumption value consistently exceeds the upper limit, it is determined that the air conditioner has a fault such as low energy efficiency or abnormal load; if the actual energy consumption value is lower than or equal to the upper limit, it is considered normal. This method eliminates the need to calculate a lower limit and a complete range, simplifying the judgment logic, and is particularly suitable for application scenarios that primarily focus on whether energy consumption is abnormally high.

[0117] The solution in this example uses energy consumption predictions and relative energy consumption thresholds derived from testing phase statistics to construct a clear fault determination range or set a comparison benchmark. This provides the fault detection process with an objective and quantitative basis for decision-making, thereby helping to improve the automation level of fault identification and the consistency of judgment, and reducing excessive reliance on human experience.

[0118] The air conditioner fault detection method provided in this application includes: acquiring historically collected target operating parameters of the air conditioner; inputting the target operating parameters into a prediction model to obtain the energy consumption prediction value output by the prediction model; and performing fault detection on the air conditioner based on the energy consumption prediction value, the relative energy consumption threshold, and the currently collected actual energy consumption value. This solution, by selecting operating parameters strongly correlated with energy consumption and constructing a prediction model, transforms fault detection into monitoring of energy consumption anomalies. This reduces the reliance on specific fault simulation experiments and data labeling. Furthermore, by adaptively determining the relative energy consumption threshold based on the relative relationship between the predicted and actual values ​​during the testing phase, it improves the adaptability of the prediction model under different operating conditions, thereby reducing the cost of fault detection.

[0119] Figure 10 This is a schematic diagram of the structure of an air conditioner fault detection device provided in this application. Figure 10 As shown, the device includes:

[0120] The acquisition module 91 is used to acquire the target operating parameters of the air conditioner collected in history. The target operating parameters are the parameters that have the highest correlation with the energy consumption parameters of the air conditioner under various different acquisition conditions. The acquisition conditions include the operating conditions.

[0121] The prediction module 92 is used to input the target operating parameters into the prediction model and obtain the energy consumption prediction value output by the prediction model; wherein, the prediction model is pre-trained and tested based on the training set, which includes the target operating parameters and corresponding energy consumption parameters under different collection conditions;

[0122] The detection module 93 is used to perform fault detection on the air conditioner based on the predicted energy consumption value, the relative energy consumption threshold, and the currently collected actual energy consumption value; wherein, the relative energy consumption threshold is determined according to the relative magnitude of the predicted energy consumption value and the actual energy consumption value output by the prediction model during the testing phase.

[0123] In one example, when the air conditioner is a multi-split air conditioner, the data collection conditions also include the number of indoor and outdoor units that are turned on; the calculation module 92 is also used for:

[0124] Acquire multiple operating parameters and corresponding energy consumption parameters of the air conditioner under different data acquisition conditions; among which, the operating parameters include environmental data and air conditioner status data; the environmental data includes at least one of the following: indoor and outdoor temperature, indoor and outdoor relative humidity, and indoor and outdoor pressure; the status data includes at least one of the following: air conditioner temperature, pressure, frequency, current, expansion valve opening, air volume level, compressor power, subcooling degree, and exhaust superheat degree.

[0125] Energy consumption parameters are categorized based on the acquisition conditions, and operating parameters are categorized based on the maximum and minimum values ​​of the operating parameters under each acquisition condition.

[0126] Based on the data pairs consisting of the categorized operating parameters and the corresponding energy consumption parameters, a correlation analysis is performed, and the operating parameters with the highest correlation to the energy consumption parameters are taken as the target operating parameters.

[0127] In one example, prediction module 92 is specifically used for:

[0128] When the air conditioner is running without faults, the operating parameters and corresponding energy consumption parameters of the air conditioner are acquired when the data collection conditions change.

[0129] And / or, when the air conditioner is running without faults, change the data collection conditions and obtain the air conditioner's operating parameters and corresponding energy consumption parameters.

[0130] In one example, prediction module 92 is specifically used for:

[0131] Obtain the data set of the first operating parameter under all historical collection conditions. Based on the distribution characteristics of the data points in the data set, divide the data points of the first operating parameter into multiple categories through a clustering analysis algorithm. Each category corresponds to a data interval, and the boundaries of the data intervals of different categories are determined by the clustering process.

[0132] The cluster category to which the data points of the first operating parameter belong under each acquisition condition are determined sequentially, and the category to which they belong is taken as the category of the first operating parameter under that acquisition condition.

[0133] Based on the classification logic of the first operating parameter, the categories of other operating parameters under each acquisition condition are determined sequentially.

[0134] In one example, detection module 93 is also used for:

[0135] During the testing phase of the prediction model, the difference between the predicted energy consumption value and the actual energy consumption value output by the prediction model is calculated, and the ratio of the difference to the actual energy consumption value is calculated.

[0136] The maximum ratio calculated during the testing phase, plus N times the difference between the maximum and minimum ratios, will be used as the relative energy consumption threshold; where N is a positive integer.

[0137] In one example, detection module 93 is specifically used for:

[0138] Calculate the product of the predicted energy consumption value and the relative energy consumption threshold, and use the difference between the predicted energy consumption value and the product as the left endpoint of the fault interval, and the sum of the predicted energy consumption value and the product as the right endpoint of the fault interval.

[0139] If the actual energy consumption value is outside the fault range, then the air conditioner is considered faulty.

[0140] It should be noted that the above Figure 10 The division of modules shown is merely illustrative. This application does not limit the division of modules or the naming of modules.

[0141] This application also provides an air conditioner that includes the air conditioner fault detection device as described in any of the above embodiments.

[0142] This application also provides an electronic device. This electronic device can be the controller for the aforementioned multi-split air conditioner. Figure 11 This is a schematic diagram of the structure of an electronic device provided in an embodiment of this application. Figure 11 As shown, the electronic device includes:

[0143] The electronic device includes a processor 291 and a memory 292; it may also include a communication interface 293 and a bus 294. The processor 291, memory 292, and communication interface 293 can communicate with each other via the bus 294. The communication interface 293 can be used for information transmission. The processor 291 can invoke logical instructions stored in the memory 292 to execute the methods of the above embodiments.

[0144] Furthermore, the logic instructions in the aforementioned memory 292 can be implemented as software functional units and, when sold or used as independent products, can be stored in a computer-readable storage medium.

[0145] The memory 292, as a computer-readable storage medium, can be used to store software programs and computer-executable programs, such as program instructions / modules corresponding to the methods in the embodiments of this application. The processor 291 executes functional applications and data processing by running the software programs, instructions, and modules stored in the memory 292, thereby implementing the methods in the above-described method embodiments.

[0146] The memory 292 may include a program storage area and a data storage area. The program storage area may store the operating system and application programs required for at least one function; the data storage area may store data created based on the use of the terminal device. Furthermore, the memory 292 may include high-speed random access memory and may also include non-volatile memory.

[0147] This application also provides a computer-readable storage medium, which may include various media capable of storing program code, such as a USB flash drive, a portable hard drive, a read-only memory (ROM), a random access memory (RAM), a magnetic disk, or an optical disk. Specifically, the computer-readable storage medium stores program instructions, which are used in the methods described in the above embodiments.

[0148] This application also provides a program product including executable instructions stored in a readable storage medium. At least one control module of a display device can read the executable instructions from the readable storage medium, and the at least one control module executes the executable instructions to cause the display device to implement the handwriting erasure methods provided in the various embodiments described above.

[0149] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of this application, and are not intended to limit them. Although this application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some or all of the technical features therein. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of this application.

[0150] For ease of explanation, the above description has been provided in conjunction with specific embodiments. However, the above exemplary discussion is not intended to be exhaustive or to limit the embodiments to the specific forms disclosed above. Various modifications and variations can be obtained based on the above teachings. The selection and description of the above embodiments are for the purpose of better explaining the principles and practical applications, thereby enabling those skilled in the art to better utilize the embodiments and various different variations of embodiments suitable for specific application considerations.

Claims

1. A method for detecting air conditioner faults, characterized in that, include: The target operating parameters of the air conditioner are obtained from historical data, wherein the target operating parameters are the parameters that have the highest correlation with the energy consumption parameters of the air conditioner under various different data collection conditions, and the data collection conditions include operating conditions. The target operating parameters are input into the prediction model to obtain the energy consumption prediction value output by the prediction model; wherein, the prediction model is pre-trained and tested based on a training set, which includes the target operating parameters and corresponding energy consumption parameters under different collection conditions; Based on the predicted energy consumption value, the relative energy consumption threshold, and the currently collected actual energy consumption value, the air conditioner is subjected to fault detection; wherein, the relative energy consumption threshold is determined according to the relative magnitude of the predicted energy consumption value and the actual energy consumption value output by the prediction model during the testing phase.

2. The method according to claim 1, characterized in that, When the air conditioner is a multi-split air conditioner, the data collection conditions also include the number of indoor and outdoor units being turned on; the method further includes: The system acquires multiple operating parameters and corresponding energy consumption parameters of the air conditioner under different data acquisition conditions. The operating parameters include environmental data and air conditioner status data. The environmental data includes at least one of the following: indoor and outdoor temperature, indoor and outdoor relative humidity, and indoor and outdoor pressure. The status data includes at least one of the following: air conditioner temperature, pressure, frequency, current, expansion valve opening, air volume level, compressor power, subcooling degree, and exhaust superheat degree. The energy consumption parameters are categorized according to the acquisition conditions, and the operating parameters are categorized according to the maximum and minimum values ​​of the operating parameters under each acquisition condition. Based on the data pairs consisting of the categorized operating parameters and the corresponding energy consumption parameters, a correlation analysis is performed, and the operating parameter with the highest correlation to the energy consumption parameter is taken as the target operating parameter.

3. The method according to claim 2, characterized in that, The acquisition of air conditioner operating parameters and corresponding energy consumption parameters under different data collection conditions includes: When the air conditioner is running without faults, the operating parameters and corresponding energy consumption parameters of the air conditioner are acquired when the acquisition conditions change. And / or, when the air conditioner is operating without faults, change the data acquisition conditions and obtain the operating parameters and corresponding energy consumption parameters of the air conditioner.

4. The method according to claim 2, characterized in that, Based on the maximum and minimum values ​​of the operating parameters, the operating parameters under each acquisition condition are categorized, including: Obtain the data set of the first operating parameter under all historical collection conditions. Based on the distribution characteristics of the data points in the data set, divide the data points of the first operating parameter into multiple categories through a clustering analysis algorithm. Each category corresponds to a data interval, and the boundaries of the data intervals of different categories are determined by the clustering process. The cluster category to which the data points of the first operating parameter belong under each acquisition condition are determined sequentially, and the category to which they belong is taken as the category of the first operating parameter under that acquisition condition. Based on the classification logic of the first operating parameter, the categories of other operating parameters under each acquisition condition are determined sequentially.

5. The method according to any one of claims 1-4, characterized in that, The method further includes: During the testing phase of the prediction model, the difference between the predicted energy consumption value and the actual energy consumption value output by the prediction model is calculated, and the ratio of this difference to the actual energy consumption value is calculated. The energy consumption relative threshold is the sum of N multiples of the difference between the maximum ratio calculated during the testing phase and the minimum ratio; where N is a positive integer.

6. The method according to claim 5, characterized in that, Based on the predicted energy consumption value, the relative energy consumption threshold, and the currently collected actual energy consumption value, fault detection is performed on the air conditioner, including: Calculate the product of the predicted energy consumption value and the relative energy consumption threshold, and use the difference between the predicted energy consumption value and the product as the left endpoint of the fault interval, and use the sum of the predicted energy consumption value and the product as the right endpoint of the fault interval. If the actual energy consumption value is outside the fault range, then the air conditioner is determined to be faulty.

7. An air conditioner fault detection device, characterized in that, include: The acquisition module is used to acquire the target operating parameters of the air conditioner collected in history, wherein the target operating parameters are the parameters that have the highest correlation with the energy consumption parameters of the air conditioner under various different acquisition conditions, and the acquisition conditions include operating conditions; The prediction module is used to input the target operating parameters into the prediction model to obtain the energy consumption prediction value output by the prediction model; wherein, the prediction model is pre-trained and tested based on a training set, the training set including the target operating parameters and corresponding energy consumption parameters under different collection conditions; The detection module is used to perform fault detection on the air conditioner based on the energy consumption prediction value, the energy consumption relative threshold, and the currently collected actual energy consumption value; wherein, the energy consumption relative threshold is determined according to the relative magnitude of the energy consumption prediction value and the actual energy consumption value output by the prediction model during the testing phase.

8. An air conditioner, characterized in that, Includes the air conditioning fault detection device as described in claim 7.

9. An electronic device, characterized in that, include: A processor, and a memory communicatively connected to the processor; The memory stores computer-executed instructions; The processor executes computer execution instructions stored in the memory to implement the method as described in any one of claims 1-6.

10. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores computer-executable instructions that, when executed by a processor, implement the method as described in any one of claims 1-6.

Images