Fault early warning method, control device and air conditioner

Abstract

The application provides a fault early warning method, a control device and an air conditioner, and belongs to the field of air conditioners. The method provided by the application comprises the following steps: acquiring multi-dimensional running state parameters of the air conditioner, the multi-dimensional running state parameters comprising at least two of filter screen state parameters, refrigeration cycle parameters and heat exchange environment parameters; constructing a running state feature vector based on the multi-dimensional running state parameters; comparing the current running state feature vector with reference data clusters of normal intervals corresponding to historical normal working conditions of the air conditioner; and determining that the air conditioner has a corresponding fault when at least two parameters in the multi-dimensional running state parameters corresponding to the same fault mode deviate from their respective normal intervals simultaneously. The method provided by the application can be used for energy-saving air conditioners, and can avoid misjudgment interference caused by single sensor instantaneous fluctuation and environmental mutation by using multi-parameter joint discrimination logic, thereby greatly reducing the fault early warning false alarm rate from the root cause and improving the implicit fault recognition accuracy.

Description

Technical Field

[0001] This application relates to the field of air conditioners, and more specifically, to a fault warning method, a control device, and an air conditioner. Background Technology

[0002] With the rapid development of the smart home industry, air conditioners, as core temperature control appliances in homes and offices, have seen a continuous increase in market share and a wide range of applications. Currently, most mainstream smart air conditioners are equipped with IoT communication modules such as Wi-Fi and Bluetooth, enabling basic operations such as remote power on / off, temperature setting, and switching operating modes via a mobile app, effectively optimizing user convenience.

[0003] However, the current research and development of smart air conditioners focuses mainly on remote interaction functions, and there are still significant technical shortcomings in the areas of overall machine health monitoring and early fault prediction. Existing air conditioner fault detection solutions rely solely on a single temperature sensor or a single pressure sensor to collect data and complete fault determination. The monitoring dimensions are limited, making them highly susceptible to interference from external factors such as fluctuations in indoor and outdoor temperature and humidity and changes in load, resulting in a high false alarm rate. Summary of the Invention

[0004] This application provides a fault early warning method, control device, and air conditioner to at least solve the technical problem of high false alarm rate in current air conditioner fault detection schemes. Specifically: This invention provides a method for early warning of air conditioner faults, comprising the following steps: acquiring multi-dimensional operating status parameters of the air conditioner, wherein the multi-dimensional operating status parameters include at least two of filter status parameters, refrigeration cycle parameters, and heat exchange environment parameters; constructing an operating status feature vector based on the multi-dimensional operating status parameters, wherein the operating status feature vector includes at least one of the average value, slope of change, and fluctuation amplitude of each parameter within a preset time period; comparing the current operating status feature vector with a benchmark data cluster of the normal range corresponding to the historical normal operating conditions of the air conditioner; and determining that the air conditioner has a corresponding fault when at least two parameters of the multi-dimensional operating status parameters corresponding to the same fault mode simultaneously deviate from their respective normal ranges.

[0005] According to one embodiment of this application, Long-term user behavior data is collected in advance, including usage time period, usage frequency, set temperature preference, operating mode selection, and usage scenario. A user behavior baseline is constructed based on this long-term user behavior data, including usage frequency characteristics, temperature preference characteristics, and scenario characteristics. According to each characteristic in the user behavior baseline, the benchmark data cluster corresponding to the historical normal operating conditions is divided into at least two behavioral feature subsets. Each behavioral feature subset corresponds to a user behavior mode provided by this invention. The statistical distribution characteristics of the same parameter in each of the behavioral feature subsets are calculated, and the normal interval threshold corresponding to this parameter in each user behavior mode is adjusted, so that the normal interval threshold is obtained hierarchically from the benchmark data cluster based on the user behavior baseline.

[0006] According to one embodiment of this application, the step of calculating the statistical distribution characteristics of the same parameter in each of the behavioral feature subsets and adjusting the normal interval threshold corresponding to the parameter under each user behavior mode includes: dividing the behavioral feature subsets into high-frequency usage scenario subsets and low-frequency usage scenario subsets based on the usage frequency characteristics in the user behavior baseline; calculating the statistical distribution characteristics of each parameter in the high-frequency usage scenario subset and the low-frequency usage scenario subset respectively to obtain a high-frequency scenario baseline threshold and a low-frequency scenario baseline threshold; matching the corresponding scenario baseline threshold according to the current usage scenario; and updating the normal interval threshold corresponding to each parameter according to the current normal interval threshold = corresponding scenario baseline threshold × usage scenario weight; wherein, the tolerance of the normal interval threshold is relaxed in high-frequency usage scenarios, and the standard of the normal interval threshold is tightened in low-frequency usage scenarios.

[0007] According to one embodiment of this application, after determining that the air conditioner has a fault, the method further includes: fitting the time-series change trend of the fault-related indicators corresponding to the fault to establish an indicator prediction model; extrapolating the indicator prediction model to a critical fault threshold to calculate the remaining service life required to reach the critical fault threshold from the current time; and generating preventive maintenance suggestions when the remaining service life is less than a preset safety window.

[0008] According to one embodiment of this application, the preset safety window is 24-168 hours.

[0009] According to one embodiment of this application, after determining that the air conditioner has a fault, the method further includes: determining the fault risk level based on the degree of impact of the fault on equipment safety and user health; if the absolute value of the refrigerant pressure is lower than a preset pressure value or the rate of refrigerant pressure drop is greater than a preset rate of drop, it is determined to be a high-risk fault of refrigerant leakage, and an emergency safety strategy is executed; if the accumulation trend of dust concentration is greater than a preset accumulation trend, it is determined to be a low-risk fault of filter clogging, and a conventional early warning strategy is executed.

[0010] According to one embodiment of this application, the emergency safety strategy includes cutting off the power supply to the air conditioner and activating the ambient air replacement device; the routine early warning strategy includes generating a health report containing the cause of the fault and quantifying energy consumption loss and pushing it to the user terminal.

[0011] According to one embodiment of this application, after executing the conventional early warning strategy, the method further includes: responding to a user-triggered maintenance request instruction and sending a maintenance request carrying fault information to the service platform so that the service platform can match maintenance resources and generate a maintenance work order.

[0012] According to one embodiment of this application, after maintenance is completed and maintenance feedback is received, the equipment health record is updated and the normal operating condition reference is recalibrated.

[0013] According to one embodiment of this application, the air conditioner locally caches the operating status parameters for the most recent preset duration; when the air conditioner is disconnected from the cloud network, fault determination logic is independently completed based on the locally cached data; wherein, the preset duration is 24-168 hours.

[0014] According to one embodiment of this application, the air conditioner malfunction includes filter clogging malfunction and system refrigerant leakage malfunction; when the dust concentration on the filter increases and the heat exchange efficiency of the evaporator is lower than a preset threshold, it is determined to be filter clogging; when the refrigerant pressure drops and the evaporator temperature rises abnormally, it is determined to be system refrigerant leakage.

[0015] According to one embodiment of this application, before constructing the air conditioner operating status feature vector, the collected time-series parameters are denoised and normalized, and an analysis window is constructed. The analysis window is used to cut continuous and irregular time-series data into structured and standardized window datasets.

[0016] According to one embodiment of this application, the filter status parameters are collected by a dust sensor located on the air inlet side of the indoor unit filter; the refrigeration cycle parameters are collected by a pressure sensor connected in series with the high-pressure pipeline of the compressor; and the heat exchange environment parameters are collected by a temperature and humidity sensor located on the evaporator fin side.

[0017] The present invention also provides a control device for an air conditioner, comprising: The parameter acquisition module is used to acquire multi-dimensional operating status parameters of the air conditioner, including at least two of the following: filter status parameters, refrigeration cycle parameters, and heat exchange environment parameters. The feature construction module is electrically connected to the parameter acquisition module and is used to construct an operating state feature vector based on the multi-dimensional operating state parameters. The operating state feature vector includes at least one of the average value, slope of change, and fluctuation amplitude of each parameter within a preset time period. The storage module stores a base data cluster of normal ranges corresponding to the historical normal operating conditions of the air conditioner; The comparison and judgment module is connected to the feature construction module and the storage module respectively, and is used to compare the current running state feature vector with the reference data cluster of the normal interval; The fault determination module, connected to the comparison and judgment module, is configured to determine that the air conditioner has a corresponding fault when at least two parameters corresponding to the same fault mode in the multi-dimensional operating status parameters deviate synchronously from their respective normal ranges.

[0018] The present invention also provides an air conditioner, comprising: a controller and a memory, wherein the controller is coupled to the memory, the memory stores a computer program, and the controller executes the computer program to implement the control method of the above embodiments.

[0019] The present invention also provides a computer-readable storage medium having stored thereon computer program instructions, which, when executed by a processor, implement the method of the above embodiments.

[0020] The present invention also provides an electronic device, comprising: a memory storing computer program instructions thereon; and a processor coupled to the memory, wherein the processor executes the instructions in the memory to implement the method of the above embodiments.

[0021] By adopting the above technical solution, this application has the following beneficial effects compared with the prior art: The fault early warning method provided in this application collects at least two types of multi-dimensional operating status parameters from filter status parameters, refrigeration cycle parameters, and heat exchange environment parameters. Based on these parameters, an operating status feature vector is constructed. The operating status feature vector includes at least one of the average value, slope of change, and fluctuation amplitude of each parameter within a preset time period. By comparing and verifying the feature vector with a benchmark data cluster formed by the air conditioner's historical normal operating conditions, a corresponding fault is determined in the air conditioner only when at least two indicators matching the same fault mode within the multi-dimensional parameters simultaneously deviate from their respective normal ranges. This multi-parameter joint discrimination logic can avoid misjudgment interference caused by instantaneous fluctuations of a single sensor and sudden environmental changes, significantly reducing the false alarm rate of fault early warning from the root cause and improving the accuracy of hidden fault identification. Attached Figure Description

[0022] Figure 1 This is a flowchart of the fault warning method in the embodiments of this application; Figure 2 This is a schematic diagram of the air conditioner in the embodiments of this application; Figure 3 This is a schematic diagram of the internal structure of the air conditioner in an embodiment of this application.

[0023] Explanation of reference numerals in the attached figures: 1. Dust sensor; 2. Pressure sensor; 3. Temperature and humidity sensor. Detailed Implementation

[0024] The embodiments of this application are described in detail below. Examples of these embodiments are shown in the accompanying drawings, wherein the same or similar reference numerals denote the same or similar elements or elements having the same or similar functions throughout. The embodiments described below with reference to the accompanying drawings are exemplary and intended to explain this application, and should not be construed as limiting this application.

[0025] In the description of this application, the phrase "in one embodiment" does not necessarily refer to the same embodiment, although it may refer to the same embodiment. Similarly, the phrase "in some embodiments," as used herein, does not necessarily refer to the same embodiment when used multiple times, although it may refer to the same embodiment. As used herein, the term "or" is an inclusive "or" operator and is equivalent to the term "and / or," unless the context clearly specifies otherwise. The term "based on" is not exclusive and allows for reliance on additional factors not described, unless the context clearly specifies otherwise. The word "exemplary" herein means "serving as an example, instance, or illustration." Any embodiment described herein as "exemplary" is not necessarily to be construed as superior to or better than other embodiments. The scope of this application is limited only by the scope of the appended claims, and any examples set forth in this specification are not intended to be limiting, but merely to illustrate some of the many possible embodiments of the claimed application. The various embodiments provided in this application should not be construed as limiting the scope of protection of this application.

[0026] Furthermore, the terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of technical features indicated. Thus, a feature defined as "first" or "second" may explicitly or implicitly include one or more of that feature. In the description of this application, "multiple" means two or more, unless otherwise explicitly specified.

[0027] In this application, unless otherwise expressly specified and limited, the terms "installation," "connection," "linking," and "fixing," etc., should be interpreted broadly. For example, they can refer to a fixed connection, a detachable connection, or an integral part; they can refer to a mechanical connection or an electrical connection; they can refer to a direct connection or an indirect connection through an intermediate medium; they can refer to the internal communication of two components or the interaction between two components. Those skilled in the art can understand the specific meaning of the above terms in this application according to the specific circumstances.

[0028] With the rapid development of the smart home industry, air conditioners, as core temperature control appliances in homes and offices, have seen a continuous increase in market share and a wide range of applications. Currently, most mainstream smart air conditioners are equipped with IoT communication modules such as Wi-Fi and Bluetooth, enabling basic operations such as remote power on / off, temperature setting, and switching operating modes via a mobile app, effectively optimizing user convenience.

[0029] However, the current research and development of smart air conditioners focuses mainly on remote interaction functions, and there are still significant technical shortcomings in the areas of overall machine health monitoring and early fault prediction. Existing air conditioner fault detection solutions rely solely on a single temperature sensor or a single pressure sensor to collect data and complete fault determination. The monitoring dimensions are limited, making them highly susceptible to interference from external factors such as fluctuations in indoor and outdoor temperature and humidity and changes in load, resulting in a high false alarm rate.

[0030] In addition, current air conditioner fault detection solutions have the following shortcomings: 1. Delayed fault identification: Traditional solutions do not have the ability to actively monitor hidden faults such as filter dust accumulation and minor refrigerant leakage. They can only be perceived by users after obvious fault symptoms such as reduced air conditioning performance and abnormal noise appear. This not only significantly increases the cost of later maintenance, but also accelerates the wear and tear of core components and shortens the service life of the whole machine.

[0031] 2. Inadequate energy efficiency management: The existing system cannot establish a correlation analysis logic between abnormal equipment energy consumption and equipment health issues such as filter blockage and refrigerant degradation. It cannot output customized energy-saving optimization solutions based on equipment operating status, making it difficult to achieve refined and precise energy efficiency management goals.

[0032] 3. Fixed early warning logic: Conventional fault alarms generally adopt global fixed threshold judgment rules, which are easily affected by external factors such as fluctuations in ambient temperature and humidity and airflow disturbances, resulting in false alarms; at the same time, they cannot adapt to different user habits and dynamically adjust the judgment criteria, and do not have the technical capability of early fault prediction and preventive maintenance.

[0033] 4. Disruption of the operation and maintenance service link: The existing fault reminder only realizes the function of information push notification, lacks an integrated handling plan after the fault occurs, cannot automatically match the nearest maintenance network, generate standardized maintenance work orders, and does not integrate fault classification linkage protection logic (such as automatic compressor shutdown and linkage of fresh air replacement when a fault occurs).

[0034] In summary, how to provide an intelligent air conditioner fault early warning and maintenance method that can collect multi-dimensional operating data in real time, predict potential faults in advance using intelligent algorithms, and provide adaptive optimization suggestions and automated service integration has become an urgent technical problem to be solved in the current home appliance technology field.

[0035] Based on this, such as Figure 1As shown, this application provides a method for early warning of air conditioner malfunctions, including the following steps: Step S1: Obtain multi-dimensional operating status parameters of the air conditioner. The multi-dimensional operating status parameters include at least two of the following: filter status parameters, refrigeration cycle parameters, and heat exchange environment parameters.

[0036] In some examples, the multi-dimensional operating status parameters are continuously collected time-series parameters, enabling continuous tracking of the equipment's operating status. These parameters include: filter status parameters such as dust concentration on the filter inlet side, filter dust accumulation rate, and return airflow attenuation; refrigeration cycle parameters such as refrigerant high-pressure pipeline pressure, pressure fluctuation amplitude, pressure time-series attenuation slope, and compressor operating current; and heat exchange environment parameters such as evaporator surface temperature and humidity, indoor ambient temperature and humidity, inlet and outlet air temperature difference, and ambient airflow disturbance parameters. These parameters are sampled in real-time and continuously by sensors mounted on the air conditioner unit, forming a complete time-series data sequence, providing continuous and comprehensive data support for subsequent feature construction and fault diagnosis. It is understandable that in other examples, the multi-dimensional operating status parameters may also be instantaneous static values.

[0037] Step S2: Construct an operating state feature vector based on multi-dimensional operating state parameters. The operating state feature vector includes at least one of the following: the average value, the slope of change, and the fluctuation amplitude of each parameter within a preset time period.

[0038] In some examples, the operating status feature vector is a structured multidimensional time-series feature set, which not only integrates the original instantaneous sampled values ​​of various parameters collected in step S1, but also calculates derived statistical and trend features based on continuous time-series data. Specifically, it includes features such as short-term average values, long-term operating average values, time-series change slopes, numerical fluctuation ranges, abnormal fluctuation amplitudes, and fluctuation frequencies. By fusing the original basic parameters with time-series derived features, the current instantaneous operating status, long-term operating trend, and operational stability of the air conditioner can be simultaneously characterized, solving the problem that single original parameter features have limited dimensions and cannot identify latent, gradual faults.

[0039] Step S3: Compare the current operating status feature vector with the benchmark data cluster of the normal interval corresponding to the historical normal operating conditions of the air conditioner.

[0040] In some examples, the baseline data cluster is a dynamic dataset generated through device adaptive learning, rather than fixed factory threshold data. The baseline data cluster is trained based on full-volume time-series operational data from the air conditioner's normal operating cycle (default one week) after installation, accurately adapting to the hardware characteristics and initial operating conditions of a single device. Simultaneously, the baseline data cluster supports dynamic iterative updates. After the user completes maintenance operations such as filter cleaning, refrigerant charging, and component inspection, the system will re-collect normal operating data and update the normal range of the baseline data cluster, ensuring that the comparison benchmark always closely matches the device's real-time health status and operating conditions.

[0041] Step S4: When at least two parameters in the multi-dimensional operating status parameters that correspond to the same fault mode deviate synchronously from their respective normal ranges, it is determined that the air conditioner has a corresponding fault.

[0042] In some examples, a multi-parameter correlation and linkage fault determination logic is employed, with preset parameter correlation matching models for different fault modes. Each fault mode corresponds to a set of strongly correlated operating parameters; for example, filter clogging faults correspond to dust concentration parameters, return air volume parameters, and inlet / outlet air temperature difference parameters; refrigerant leakage faults correspond to pipeline pressure parameters, pressure decay slope parameters, and heat exchange temperature difference parameters. Equipment faults are only confirmed when at least two types of correlated timing parameters under the same fault mode synchronously and continuously deviate from the normal range corresponding to the baseline data cluster, rather than experiencing instantaneous single-point value fluctuations. This effectively shields against false anomalies caused by sudden environmental changes and instantaneous sensor interference, significantly improving the accuracy of fault determination.

[0043] The fault early warning method provided in this application collects at least two types of multi-dimensional operating status parameters from filter status parameters, refrigeration cycle parameters, and heat exchange environment parameters. Based on these parameters, an operating status feature vector is constructed. The operating status feature vector includes at least one of the average value, slope of change, and fluctuation amplitude of each parameter within a preset time period. By comparing and verifying the feature vector with a benchmark data cluster formed by the air conditioner's historical normal operating conditions, a corresponding fault is determined in the air conditioner only when at least two indicators matching the same fault mode within the multi-dimensional parameters simultaneously deviate from their respective normal ranges. This multi-parameter joint discrimination logic can avoid misjudgment interference caused by instantaneous fluctuations of a single sensor and sudden environmental changes, significantly reducing the false alarm rate of fault early warning from the root cause and improving the accuracy of hidden fault identification.

[0044] In some implementations, the air conditioner fault early warning method further includes the following steps: Pre-collect long-term user behavior data, including usage time, usage frequency, set temperature preference, operating mode selection, and usage scenario; construct a user behavior baseline based on the long-term user behavior data, including usage frequency characteristics, temperature preference characteristics, and scenario characteristics. Based on the features in the user behavior baseline, the benchmark data cluster corresponding to the historical normal working conditions is divided into at least two behavioral feature subsets, each of which corresponds to a user behavior pattern. The statistical distribution features of the same parameter in each behavioral feature subset are calculated, and the normal interval threshold corresponding to the parameter in each user behavior pattern is adjusted so that the normal interval threshold is obtained by hierarchical mapping from the benchmark data cluster based on the user behavior baseline.

[0045] Specifically, based on the features in the user behavior baseline, the benchmark data cluster corresponding to the historical normal working conditions is divided into at least two behavioral feature subsets, each of which corresponds to a user behavior pattern. The statistical distribution features of the same parameter in each behavioral feature subset are calculated. The statistical distribution features include the mean, standard deviation, and preset quantile values ​​of the parameter in the corresponding subset. The interval determined by the mean plus or minus a preset multiple of the standard deviation is used as the initial normal interval of the parameter in the user behavior pattern. Then, the normal interval threshold corresponding to the parameter in each user behavior pattern is adjusted so that the normal interval threshold is obtained by hierarchical mapping from the benchmark data cluster based on the user behavior baseline.

[0046] In some examples, the statistical distribution characteristics of the same parameter in each subset of behavioral features are calculated separately, and the normal interval threshold corresponding to the parameter under each user behavior mode is adjusted. This includes: dividing the behavioral feature subset into high-frequency usage scenario subset and low-frequency usage scenario subset based on the usage frequency characteristics in the user behavior baseline; calculating the statistical distribution characteristics of each parameter in the high-frequency usage scenario subset and the low-frequency usage scenario subset respectively to obtain the high-frequency scenario baseline threshold and the low-frequency scenario baseline threshold; matching the corresponding scenario baseline threshold according to the current usage scenario; and updating the normal interval threshold corresponding to each parameter according to the formula: current normal interval threshold = corresponding scenario baseline threshold × usage scenario weight. Among these, the tolerance of the normal interval threshold is relaxed for high-frequency usage scenarios, and the standard of the normal interval threshold is tightened for low-frequency usage scenarios.

[0047] Specifically, based on the usage frequency characteristics in the user behavior baseline, the behavioral feature subset is divided into a high-frequency usage scenario subset and a low-frequency usage scenario subset. The high-frequency usage scenario subset corresponds to the condition where the percentage of runtime per unit time exceeds a preset proportion, and the low-frequency usage scenario subset corresponds to the condition where the percentage of runtime per unit time is lower than a preset proportion. The statistical distribution characteristics of each parameter in the high-frequency and low-frequency usage scenario subsets are calculated to obtain the high-frequency scenario baseline threshold and the low-frequency scenario baseline threshold. The corresponding scenario baseline threshold is matched according to the current usage scenario. The normal interval threshold corresponding to each parameter is updated according to the formula: current normal interval threshold = corresponding scenario baseline threshold × usage scenario weight. The usage scenario weight is determined by comprehensively considering the occurrence proportion of the scenario in historical usage behavior and the historical false alarm feedback rate. The higher the occurrence proportion and the lower the historical false alarm feedback rate, the closer the weight of the scenario is to 1. Among them, the tolerance of the normal interval threshold is relaxed for high-frequency usage scenarios, that is, the multiple of the preset standard deviation is increased to increase the width of the normal interval, while the standard of the normal interval threshold is tightened for low-frequency usage scenarios, that is, the multiple of the preset standard deviation is decreased to narrow the width of the normal interval. The reason is that in high-frequency usage scenarios, the air conditioner compressor starts and stops frequently and the load fluctuates greatly. The normal fluctuation range of parameters is relatively wide, and a loose tolerance can avoid misjudging normal load fluctuations as faults. In low-frequency usage scenarios, the air conditioner's operating conditions are relatively simple and the parameters are stable. Moreover, the sample size of operating data is small, and tightening the standards can improve the sensitivity of detecting potential anomalies.

[0048] In this design approach, by adding a user behavior adaptive optimization mechanism, the system continuously records and collects user behavior data on the daily operation of the air conditioner. This includes, but is not limited to, the frequency of daily on / off cycles, the cumulative daily / single-hour usage time constituting the device's usage frequency, the user's long-term fixed cooling / heating temperature settings, the set temperature preferences corresponding to temperature adjustment habits, and differentiated usage scenarios such as daytime regular use, nighttime sleep use, and seasonal high / low load use. The system continuously updates user behavior data in periodic statistical windows (default one week), intelligently distinguishing between high-frequency and low-frequency usage scenarios and assigning differentiated quantitative scenario weights to the two types of scenarios. High-frequency usage scenarios correspond to smaller scenario weight values, while low-frequency usage scenarios correspond to larger scenario weight values. The system uses a fixed calculation formula: Current warning threshold = Baseline threshold × Usage scenario weight, to iteratively update the normal range thresholds corresponding to filter status parameters, cooling cycle parameters, and heat exchange environment parameters in real time.

[0049] The specific adaptation logic is as follows: For high-frequency, high-load usage scenarios, the normal fluctuation range of parameters is amplified by weighting coefficients, and the threshold tolerance of the normal range is relaxed to adapt to the normal operating condition fluctuations caused by long-term continuous operation of equipment and avoid frequent false alarms; For low-frequency, low-load or standby usage scenarios, the allowable fluctuation range of parameters is narrowed by weighting coefficients, and the judgment standard of the normal range threshold is tightened to improve the ability to detect gradual hidden faults such as trace refrigerant decay and slow dust accumulation on the filter.

[0050] This example uses a quantification formula to achieve adaptive and refined dynamic adjustment of the threshold, so that the sensitivity of the fault warning is highly matched with the user's actual operating conditions and the equipment's operating status, effectively balancing the accuracy of the warning and the user experience, and further reducing the probability of false alarms and missed alarms.

[0051] In some implementations, after determining that the air conditioner has a fault, the process further includes: fitting the time-series change trend of the fault-related indicators corresponding to the fault to establish an indicator prediction model; extrapolating the indicator prediction model to the critical fault threshold to calculate the remaining service life required to reach the critical fault threshold from the current moment; and generating preventive maintenance suggestions when the remaining service life is less than a preset safety window.

[0052] In this design approach, based on the initial determination of air conditioner faults, a mechanism for assessing remaining lifespan and triggering preventive maintenance based on time-series trend prediction is added, enabling proactive fault prediction and maintenance.

[0053] Specifically, the system accurately identifies the core related indicators corresponding to the currently identified faults. For example, a filter blockage fault corresponds to the dust accumulation rate and airflow reduction indicators, while a refrigerant leakage fault corresponds to the refrigerant pressure reduction slope and heat exchange efficiency reduction indicators. For the above fault-related indicators, the system retrieves the continuous time-series operating data stored in the equipment over a long period of time, and uses time-series analysis algorithms such as linear regression and moving average to perform curve fitting on the dynamic trend of the indicators changing over time, thereby constructing a dedicated dynamic prediction model for the indicators that is adapted to the current operating conditions of the equipment.

[0054] This model can accurately replicate the gradual decay or cumulative deterioration of fault indicators, and can truly reflect the progressive development characteristics of equipment failures. Subsequently, the system extrapolates the completed indicator prediction model backward along the time axis, using the industry- and equipment-calibrated critical fault thresholds as the judgment boundary, to accurately calculate the remaining service life (RUL) from the current equipment operating time until the indicators reach the critical fault threshold, the equipment experiences functional failure, or fails to meet energy efficiency standards.

[0055] The system has a preset safety window, ranging from 24 to 168 hours, specifically 72 hours. When the calculated remaining service life is less than this preset safety window, it indicates that the equipment failure is about to rapidly worsen. The system automatically generates structured, actionable preventative maintenance recommendations, providing forward-looking energy-saving guidance. It can also quantify the additional energy consumption and equipment damage risks resulting from delayed maintenance, guiding users to complete maintenance operations such as filter cleaning, refrigerant replenishment, and component inspection before the equipment completely fails and its performance significantly degrades. This effectively reduces maintenance costs, minimizes energy waste, and extends the overall lifespan of the air conditioner.

[0056] In some implementations, after determining that the air conditioner has a fault, the following steps are also taken: determining the fault risk level based on the degree of impact of the fault on equipment safety and user health; if the absolute value of the refrigerant pressure is lower than the preset pressure value or the rate of refrigerant pressure drop is greater than the preset rate of drop, it is determined to be a high-risk fault of refrigerant leakage, and an emergency safety strategy is executed; if the cumulative trend of dust concentration is greater than the preset cumulative trend, it is determined to be a low-risk fault of filter clogging, and a regular early warning strategy is executed.

[0057] In this approach, after fault identification and judgment are completed, a fault risk classification and differentiated handling mechanism is added. The risk level is divided according to the fault hazard attributes, the degree of impact on equipment operation safety and the physical health of indoor users, so as to achieve precise and hierarchical management of fault handling and avoid the confusion and safety hazards caused by uniform pop-up warnings for all faults.

[0058] High-risk faults focus on fault types that pose safety hazards and are prone to causing equipment damage or indoor environmental risks. A typical scenario is refrigerant leakage: the system monitors the absolute value of refrigerant pressure and the rate of pressure decrease over time in the refrigeration cycle pipeline. When the absolute value of the static / dynamic refrigerant pressure is lower than the system's preset safety pressure threshold, or the rate of pressure decrease per unit time is greater than the preset decay rate threshold, it indicates a rapid and significant refrigerant leak in the air conditioning system. This type of fault can easily cause compressor burnout due to insufficient liquid, refrigerant accumulation polluting indoor air, and other safety hazards. Therefore, the system identifies it as a high-risk refrigerant leakage fault and immediately activates emergency safety strategies. This automatically cuts off the compressor's power supply and simultaneously activates the fresh air system for indoor air replacement, mitigating equipment damage and occupant safety risks from the source.

[0059] Low-risk faults only affect the air conditioner's heat exchange efficiency, energy consumption, and user experience, without posing immediate safety hazards. A typical scenario for diagnosing this is filter clogging: the system continuously monitors the cumulative trend of dust concentration on the filter. When the rate of increase in dust concentration and the total long-term accumulation exceed a preset cumulative trend threshold, and no abnormal equipment safety parameters are triggered, it is determined to be a low-risk filter clogging fault. For this type of non-safety, gradual performance loss fault, the system only executes standard early warning strategies. A health report containing the cause of the fault, quantified energy loss, and filter cleaning and maintenance recommendations is generated and pushed to the user's terminal via the user's app. Aggressive operations such as forced equipment shutdown are avoided, ensuring long-term healthy operation of the equipment while maximizing the user's normal operating experience.

[0060] This tiered response mechanism enables differentiated management and control of emergency protection against safety risk-related faults and flexible alerts for performance loss-related faults, thus improving the closed-loop system for the entire process of air conditioning fault early warning and handling.

[0061] In some examples, after executing the regular warning strategy, the process also includes: responding to a user-triggered maintenance request instruction and sending a maintenance request carrying fault information to the service platform so that the service platform can match maintenance resources and generate a maintenance work order.

[0062] This approach addresses the shortcomings of traditional air conditioner fault warning systems, which only inform users of faults but lack direct access to repair services and involve cumbersome user repair procedures, by establishing a seamless information exchange link between the equipment, user, and after-sales service departments. After sending routine warnings such as equipment health reports and maintenance suggestions to users, the system can monitor and respond in real-time to one-click repair requests triggered by users on their mobile apps. The repair request data uploaded to the service platform is structured and complete fault information, including the currently determined fault type, fault level, abnormal equipment operating parameters, equipment model information, user equipment location information, and estimated energy consumption. Users do not need to manually describe the fault symptoms. Upon receiving the repair request, the cloud service platform automatically calls the service provider resource matching interface, intelligently matching optimal repair resources based on user location, service radius of repair outlets, technician ratings, and order status. This quickly generates standardized electronic repair work orders, which are then simultaneously pushed to the corresponding technician and user terminals. This achieves a fully automated closed loop from fault warning, user confirmation, resource matching to work order generation, significantly simplifying the user repair process and improving after-sales repair response efficiency and service standardization.

[0063] In some specific examples, after maintenance is completed and maintenance feedback is received, the equipment health record is updated and the normal operating condition baseline is recalibrated.

[0064] This approach establishes a post-repair iterative update mechanism, forming a complete closed-loop system for equipment monitoring, early warning, maintenance, and self-calibration. After maintenance personnel complete equipment inspection, fault repair, component replacement, or cleaning and maintenance, the service provider or user submits maintenance completion feedback information in the background, and the system receives and confirms the maintenance closed-loop signal in the cloud.

[0065] First, the system automatically updates the cloud-based device health record, fully recording the fault type, abnormal parameter changes, fault occurrence time, repair items, maintenance results, and energy consumption improvement. This forms a traceable, full-lifecycle device operation and maintenance ledger, providing data support for subsequent equipment aging analysis, fault pattern statistics, and personalized maintenance strategy iteration. Simultaneously, the system initiates a normal operating condition baseline recalibration process, clearing abnormal data interference from the fault-prone phase, re-collecting normal operating time-series data for a period after the air conditioner's repair, and adaptively learning to generate a new normal operating condition baseline data cluster, updating the normal range thresholds for each operating parameter. This post-repair baseline reset mechanism effectively avoids operating condition deviations caused by historical fault residual data, component aging and replacement, and maintenance interventions. It ensures that subsequent fault warnings, threshold judgments, and trend analyses are based on the latest equipment health conditions, continuously guaranteeing the system's fault identification accuracy and dynamic threshold adaptability, ensuring that warning judgments remain accurate and reliable throughout long-term equipment operation.

[0066] In some implementations, the air conditioner locally caches the operating status parameters for the most recent preset duration; when the air conditioner disconnects from the cloud network, the fault determination logic is independently completed based on the locally cached data; wherein, the preset duration is 24-168 hours.

[0067] This approach adds a local data cache and an offline independent analysis mechanism, effectively solving the shortcomings of traditional cloud-dependent fault monitoring solutions that cannot work in scenarios such as network disconnection, weak signal, and abnormal network connection, and greatly improving the stability and scenario adaptability of the system.

[0068] During normal operation, the air conditioner continuously stores multi-dimensional time-series operating status data such as filter status parameters, refrigeration cycle parameters, and heat exchange environment parameters in the local main control cache area. The effective data duration of the local cache is limited to 24-168 hours, preferably 24 hours. This not only retains a sufficient amount of continuous time-series data for trend analysis and fault comparison, but also avoids the long-term storage of massive amounts of data occupying local storage resources, thus achieving a balance between storage resources and analysis accuracy.

[0069] When the device detects network outages such as interruptions in cloud communication links (WiFi, MQTT, etc.) or network handshake failures, the system automatically switches to offline mode. It operates without relying on cloud databases, cloud model computing power, or cloud benchmark data. Instead, it independently utilizes locally cached recent complete time-series datasets to independently complete a full set of fault warning logic, including feature vector construction, benchmark data comparison, multi-parameter correlation fault determination, and fault trend prediction. This ensures uninterrupted and unaffected device health monitoring, latent fault identification, and preventative maintenance assessment functions during network outages. Once the device's network is restored, all monitoring data, warning records, and device health logs from the offline period can be uniformly synchronized to the cloud, completing data retransmission and file updates, achieving uninterrupted monitoring across all online and offline scenarios.

[0070] In some implementations, air conditioner malfunctions include filter clogging and system refrigerant leakage; filter clogging is determined when the dust concentration on the filter increases and the evaporator heat exchange efficiency is lower than a preset threshold; and system refrigerant leakage is determined when the refrigerant pressure drops and the evaporator temperature rises abnormally.

[0071] This application addresses common air conditioner settings with multi-parameter linkage judgment logic, relying on a multi-sensor data fusion mechanism to achieve accurate fault identification. Among these, filter clogging faults fall under the category of progressive performance degradation faults, and the fault determination employs a dual-parameter linkage verification mechanism. The system monitors the real-time changes in dust concentration on the filter inlet side. If the dust concentration continues to accumulate and increase, and the actual heat exchange efficiency of the evaporator, calculated based on parameters such as the evaporator inlet and outlet air temperature difference and heat exchange airflow, remains consistently lower than the preset efficiency threshold corresponding to normal operating conditions, and both parameters show abnormal deviations simultaneously, then a filter blockage fault is accurately determined. This determination method can effectively distinguish between instantaneous fluctuations in ambient dust and actual filter blockage, accurately identifying hidden faults such as impaired heat exchange, reduced airflow, and decreased energy efficiency caused by filter dirt.

[0072] System refrigerant leakage is a core abnormality in the refrigeration system, and it is also judged using a combined logic of related parameters: Under stable operating conditions of the air conditioner, if the refrigerant pressure in the high-pressure refrigerant line is continuously decreasing, the pressure decay slope exceeds the normal range, and this is accompanied by an abnormal increase in the evaporator surface temperature and a decrease in refrigeration heat exchange capacity, and both related parameters are abnormal simultaneously, then a system refrigerant leakage fault can be determined. Because refrigerant leakage directly leads to insufficient refrigeration circulating medium and inadequate heat exchange in the evaporator, it results in a chain reaction of decreasing pressure and abnormally rising evaporator temperature.

[0073] This linkage judgment logic can accurately capture hidden faults such as minute and slow refrigerant leaks, avoid misjudgment due to single pressure fluctuations, and achieve accurate differentiation and reliable identification of the two types of core faults, providing accurate fault type basis for subsequent risk classification, safety protection and preventive maintenance.

[0074] In some implementations, before constructing the feature vector of the air conditioner's operating status, the collected time-series parameters are denoised and normalized, and an analysis window is constructed, with the analysis window ranging from 24 to 168 hours.

[0075] In this approach, a complete data preprocessing process is added before feature vector construction. This addresses issues such as noise interference, inconsistent dimensions, and strong data dispersion in the raw time-series parameters collected by sensors in real time. Standardized preprocessing operations improve data quality and ensure the accuracy of subsequent feature construction and fault assessment.

[0076] The noise reduction process is mainly used to filter out invalid and abnormal noise data caused by sudden environmental interference, instantaneous sensor jitter, and circuit signal fluctuations, eliminate invalid values ​​caused by single-point sudden changes, retain effective time-series data that can truly reflect the changes in the operating conditions of the air conditioner, and avoid abnormal noise data interfering with the feature extraction results.

[0077] Normalization is used to unify the dimensions and numerical ranges of multi-dimensional heterogeneous data such as filter state parameters, refrigeration cycle parameters, and heat exchange environment parameters, eliminating the feature weight imbalance caused by differences in units and numerical magnitudes of different physical parameters, and allowing various parameters to participate in unified feature fusion calculation and comparative analysis.

[0078] Meanwhile, the system constructs a fixed-duration sliding analysis window for the preprocessed time series data, cutting the continuous and irregular time series data into a structured and standardized window dataset. This facilitates the subsequent extraction of statistical and trend features such as time series mean, slope of change, and fluctuation amplitude, achieving the structured regularization of time series data. This provides high-quality and highly regularized basic data support for accurately constructing multi-dimensional operating status feature vectors and conducting multi-parameter fusion fault judgment, thereby reducing the probability of fault misjudgment and missed judgment from the data source.

[0079] In some implementations, combined Figure 2 and Figure 3 As shown, filter status parameters are collected by a dust sensor 1 located on the air inlet side of the indoor unit filter; refrigeration cycle parameters are collected by a pressure sensor 2 connected in series with the compressor high-pressure pipeline; and heat exchange environment parameters are collected by a temperature and humidity sensor 3 located on the evaporator fin side.

[0080] Specifically, the dust sensor 1 is embedded on the edge of the air inlet side of the indoor unit's filter, with the sampling port facing the return airflow path. It can capture changes in the concentration of air dust entering the filter in real time, accurately and in real time collect the time-series parameters of the filter's dust accumulation status, and can truly reflect the degree of filter blockage and the rate of dust accumulation, avoiding data lag and sampling distortion caused by post-collection.

[0081] Pressure sensor 2 is connected in series via a dedicated tee connector on the high-pressure pipeline between the outdoor unit compressor discharge port and the condenser. This location is the core location where the refrigeration cycle system is most sensitive to operating conditions and where pressure fluctuations most accurately reflect the refrigerant state. It can accurately collect core parameters such as the instantaneous pressure value, pressure fluctuation amplitude, and pressure decay slope of the high-pressure side of the refrigeration cycle, and sensitively detect pressure changes caused by hidden anomalies such as minor refrigerant leaks and pipeline blockages.

[0082] Temperature and humidity sensor 3 is closely attached to the surface of the evaporator fins, which can directly sense the real temperature and humidity conditions on the heat exchange side of the evaporator, accurately collect heat exchange environment parameters, effectively feed back the heat exchange efficiency of the evaporator, the state of heat exchange and cold and heat exchange and the disturbance of environmental conditions, and avoid the problems of lag and large deviation in the collection of indoor remote environmental temperature and humidity.

[0083] This layered and zoned sensor layout ensures the targeted and accurate acquisition of each type of parameter, providing a reliable hardware data acquisition foundation for subsequent multi-sensor data fusion, fault trend analysis, and accurate early warning.

[0084] This application also provides a control device for an air conditioner, including a parameter acquisition module, a feature construction module, a storage module, a comparison and judgment module, and a fault judgment module.

[0085] This control device is the hardware logic architecture for realizing the aforementioned air conditioner fault early warning method. Each module has a clear division of labor and a closed-loop logical linkage, which can independently complete equipment status monitoring, feature analysis, data comparison and intelligent fault determination. The system comprises several modules: a parameter acquisition module, a feature construction module, and a fault determination module. The parameter acquisition module is responsible for acquiring multi-dimensional operating status parameters of the air conditioner in real time and continuously, including filter status parameters, refrigeration cycle parameters, and heat exchange environment parameters, providing raw data input for device calculations. The feature construction module establishes an electrical connection with the parameter acquisition module, receives multi-dimensional raw operating data, integrates features, and constructs a structured operating status feature vector for the air conditioner. The operating status feature vector includes at least one of the average value, slope of change, and fluctuation amplitude of each parameter within a preset time period. The storage module is used to persistently store historical normal operating condition data of the air conditioner, forming a normal range benchmark data cluster that can be dynamically iteratively updated, providing a standard basis for fault comparison. The comparison and judgment module establishes data connections with the feature construction module and the storage module respectively, retrieves the real-time operating status feature vector and the benchmark data cluster to complete differential comparison analysis. The fault determination module is coupled to the comparison and judgment module, receives the comparison results, and executes joint fault determination logic. Only when at least two parameters corresponding to the same fault mode in the multi-dimensional operating status parameters deviate synchronously from their respective normal ranges is the corresponding fault of the air conditioner finally confirmed.

[0086] This modular architecture has clear logic and strong coupling, enabling intelligent analysis of the entire process at the local edge of the air conditioner without long-term reliance on cloud computing power, effectively improving the real-time performance and stability of fault diagnosis.

[0087] This application also provides an air conditioner, which is a physical device capable of implementing the aforementioned fault warning and preventive maintenance control methods. The core of the device adopts a hardware-software co-engineering architecture that couples a controller and a memory. The memory serves as a program and data storage carrier, and contains a built-in computer program that implements a complete set of control logic, including multi-parameter acquisition and preprocessing, feature vector construction, dynamic benchmark comparison, adaptive threshold adjustment, fault classification and determination, preventive life prediction, service closed-loop linkage, and offline data caching. The controller, as the core computing and execution unit of the device, can call and run the computer program stored in the memory, thereby accurately executing the air conditioner fault warning and preventive maintenance control methods of any of the aforementioned embodiments.

[0088] Through this hardware and software adaptation architecture, the air conditioner can autonomously complete a series of intelligent functions such as real-time monitoring of equipment health status, accurate identification of hidden faults, personalized operating condition adaptive correction, fault risk classification and handling, preventive maintenance push and after-sales service linkage, effectively improving the air conditioner's operational reliability, fault early warning accuracy and intelligent operation and maintenance level, and reducing equipment energy consumption and after-sales maintenance costs.

[0089] This application also provides a computer-readable storage medium, which is a non-volatile storage carrier for the program logic of an air conditioner fault early warning method and can be applied to the control hardware system of various intelligent air conditioners. The storage medium contains embedded computer program instructions that can be recognized, loaded, and executed by the device processor. These instructions completely correspond to all the steps of the aforementioned method, including multi-parameter acquisition and preprocessing, feature vector construction, dynamic benchmark comparison, adaptive threshold adjustment, joint fault determination, remaining service life prediction, fault classification and handling, closed-loop linkage of maintenance services, and offline data caching. When the program instructions stored in the storage medium are read and executed by the processor of the air conditioning device, the air conditioner fault early warning and preventive maintenance control process disclosed in any embodiment of this application can be stably and completely implemented. This provides software program support for the air conditioning device to achieve intelligent fault monitoring, proactive preventive maintenance, adaptive operating condition correction, and full-link service closed-loop functions. It features strong versatility, high portability, and stable operation, and can be independently adapted to various air conditioning devices with intelligent control modules.

[0090] This application also provides an electronic device, which is an intelligent control hardware device capable of independently executing an air conditioner fault early warning method, applicable to various intelligent air conditioner main control systems and supporting edge control terminals. The electronic device adopts a hardware architecture coupling memory and processor. The memory is used to persistently store computer program instructions that implement all control logic of this application, including complete algorithm programs such as data preprocessing, feature vector construction, dynamic benchmark comparison, user adaptive threshold adjustment, joint fault judgment, remaining life prediction, risk classification and handling, after-sales service closed-loop linkage, offline data caching, and benchmark calibration. The processor, as the core computing and execution unit, can call and read the computer program instructions stored in the memory, and execute the air conditioner fault early warning and preventive maintenance control method corresponding to any of the aforementioned embodiments through computation and parsing. Based on a hardware and software collaborative architecture, this electronic device can independently complete full-condition monitoring of air conditioning equipment, accurate prediction of latent faults, intelligent adaptive correction, security protection control, and operation and maintenance service linkage on the local edge side, without relying on real-time computing power in the cloud. It has strong independence and stability, high versatility, and can be widely adapted to different models and series of intelligent air conditioning products.

[0091] In the above embodiments of this application, the descriptions of each embodiment have their own emphasis. Parts not described in detail in a certain embodiment can be referred to in the relevant descriptions of other embodiments. The steps illustrated in the related flowcharts can be executed in a computer system such as a set of computer-executable instructions. Furthermore, although a logical order is shown in the flowcharts, in some cases, the steps shown or described may be performed in a different order than that shown here. In other words, the order of steps described in the foregoing embodiments is merely an example. Reasonable adjustments to the order of steps based on the content of the embodiments of this application are also within the protection scope of the embodiments of this application.

[0092] Although embodiments of this application have been shown and described above, it is understood that the above embodiments are exemplary and should not be construed as limiting this application. Those skilled in the art can make changes, modifications, substitutions, and variations to the above embodiments within the scope of this application. Furthermore, those skilled in the art can combine and integrate the different embodiments or examples described in this specification.

Claims

1. A method for early warning of air conditioner malfunctions, characterized in that, Includes the following steps: The multi-dimensional operating status parameters of the air conditioner are obtained, including at least two of the following: filter status parameters, refrigeration cycle parameters, and heat exchange environment parameters. An operation state feature vector is constructed based on the multi-dimensional operation state parameters. The operation state feature vector includes at least one of the average value, slope of change, and fluctuation amplitude of each parameter within a preset time period. The current operating status feature vector is compared with the benchmark data cluster of the normal interval corresponding to the historical normal operating conditions of the air conditioner. When at least two parameters in the multi-dimensional operating status parameters that correspond to the same fault mode deviate synchronously from their respective normal ranges, it is determined that the air conditioner has a corresponding fault.

2. The air conditioner fault early warning method according to claim 1, characterized in that, The air conditioner malfunction early warning method also includes the following steps: Pre-collect long-term user behavior data, including usage time period, usage frequency, set temperature preference, operating mode selection and usage scenario; A user behavior baseline is constructed based on the long-term usage behavior data. The user behavior baseline includes usage frequency characteristics, temperature preference characteristics, and scenario characteristics. Based on the features in the user behavior baseline, the benchmark data cluster corresponding to the historical normal working conditions is divided into at least two behavioral feature subsets, and each behavioral feature subset corresponds to a user behavior pattern. Calculate the statistical distribution characteristics of the same parameter in each of the aforementioned behavioral feature subsets, adjust the normal interval threshold corresponding to the parameter under each user behavior pattern, and obtain the normal interval threshold from the benchmark data cluster in a hierarchical mapping based on the user behavior baseline.

3. The air conditioner fault early warning method according to claim 2, characterized in that, The step of calculating the statistical distribution characteristics of the same parameter in each of the aforementioned behavioral feature subsets and adjusting the normal interval threshold corresponding to the parameter under each user behavior pattern includes: Based on the usage frequency characteristics in the user behavior baseline, the subset of behavioral characteristics is divided into a high-frequency usage scenario subset and a low-frequency usage scenario subset; Calculate the statistical distribution characteristics of each parameter in the high-frequency usage scenario subset and the low-frequency usage scenario subset respectively to obtain the high-frequency scenario benchmark threshold and the low-frequency scenario benchmark threshold; Match the corresponding scenario baseline threshold based on the current usage scenario; Update the normal range thresholds for each parameter according to the formula: current normal range threshold = corresponding scenario baseline threshold × usage scenario weight; In high-frequency usage scenarios, the tolerance of the normal range threshold is relaxed, while in low-frequency usage scenarios, the standard of the normal range threshold is tightened.

4. The air conditioner fault early warning method according to claim 1, characterized in that, After determining that the air conditioner is faulty, the method further includes: The time-series change trends of the fault-related indicators corresponding to the fault are fitted to establish an indicator prediction model; Extrapolate the index prediction model to the critical failure threshold and calculate the remaining service life required to reach the critical failure threshold from the current moment. When the remaining service life is less than a preset safety window, a preventative maintenance recommendation is generated.

5. The air conditioner fault early warning method according to claim 4, characterized in that, The preset safety window is 24-168 hours.

6. The air conditioner fault early warning method according to claim 1, characterized in that, After determining that the air conditioner is faulty, the method further includes: The fault risk level is determined based on the degree of impact of the fault on equipment safety and user health; If the absolute value of the refrigerant pressure is lower than the preset pressure value or the rate of decrease of the refrigerant pressure is greater than the preset rate of decrease, it is determined to be a high-risk fault of refrigerant leakage, and an emergency safety strategy is executed. If the cumulative trend of dust concentration is greater than the preset cumulative trend, it is determined to be a low-risk fault of filter clogging, and the normal early warning strategy is executed.

7. The air conditioner fault early warning method according to claim 6, characterized in that, The emergency safety strategy includes cutting off the power supply to the air conditioner and activating the ambient air replacement device. The conventional early warning strategy includes generating a health report containing the cause of the fault and quantifying energy loss, and pushing it to the user's terminal.

8. The air conditioner fault early warning method according to claim 6, characterized in that, After executing the aforementioned conventional early warning strategy, it also includes: In response to a user-triggered maintenance request, a maintenance request carrying fault information is sent to the service platform so that the service platform can match maintenance resources and generate a maintenance work order.

9. The air conditioner fault early warning method according to claim 8, characterized in that, After maintenance is completed and maintenance feedback is received, update the equipment health record and recalibrate the normal operating condition benchmark.

10. The air conditioner fault early warning method according to claim 1, characterized in that, The air conditioner locally caches the operating status parameters for the most recent preset time. When the air conditioner disconnects from the cloud network, the fault determination logic is independently completed based on the local cached data. The preset duration is 24-168 hours.

11. The air conditioner fault early warning method according to claim 1, characterized in that, The air conditioner malfunctions include filter blockage and system refrigerant leakage. When the dust concentration on the filter increases and the heat exchange efficiency of the evaporator is lower than the preset threshold, it is determined that the filter is clogged. When the refrigerant pressure drops and the evaporator temperature rises abnormally, it is determined that the system is leaking refrigerant.

12. The air conditioner fault early warning method according to claim 1, characterized in that, Before constructing the feature vector of the air conditioner's operating status, the collected time-series parameters are denoised and normalized, and an analysis window is constructed. The analysis window is used to cut the continuous and irregular time-series data into a structured and standardized window dataset.

13. The air conditioner fault early warning method according to claim 1, characterized in that, The filter status parameters are collected by a dust sensor (1) located on the air inlet side of the indoor unit filter; the refrigeration cycle parameters are collected by a pressure sensor (2) connected in series with the high-pressure pipeline of the compressor; and the heat exchange environment parameters are collected by a temperature and humidity sensor (3) located on the evaporator fin side.

14. A control device for an air conditioner, characterized in that, include: The parameter acquisition module is used to acquire multi-dimensional operating status parameters of the air conditioner, including at least two of the following: filter status parameters, refrigeration cycle parameters, and heat exchange environment parameters. The feature construction module is electrically connected to the parameter acquisition module and is used to construct an operating state feature vector based on the multi-dimensional operating state parameters. The operating state feature vector includes at least one of the average value, slope of change, and fluctuation amplitude of each parameter within a preset time period. The storage module stores a base data cluster of normal ranges corresponding to the historical normal operating conditions of the air conditioner; The comparison and judgment module is connected to the feature construction module and the storage module respectively, and is used to compare the current running state feature vector with the reference data cluster of the normal interval; The fault determination module, connected to the comparison and judgment module, is configured to determine that the air conditioner has a corresponding fault when at least two parameters corresponding to the same fault mode in the multi-dimensional operating status parameters deviate synchronously from their respective normal ranges.

15. An air conditioner, characterized in that, include: A controller and a memory, the controller being coupled to the memory, the memory storing a computer program, the controller executing the computer program to implement the control method according to any one of claims 1-13.

16. A computer-readable storage medium having computer program instructions stored thereon, characterized in that, When the computer program instructions are executed by the processor, they implement the method of any one of claims 1 to 13.

17. An electronic device, characterized in that, include: A memory that stores computer program instructions; A processor, coupled to the memory, implements the method of any one of claims 1 to 13 when executing instructions in the memory.

Images