Refrigerator door state detection method, controller, refrigerator, medium, and program product

Abstract

The application provides a refrigerator door state detection method, a controller, a refrigerator, a medium and a program product, and comprises the following steps: collecting historical operation parameters of the refrigerator in a first time period before a current time, preprocessing the historical operation parameters to obtain a first parameter time sequence; inputting the first parameter time sequence into a pre-trained prediction model, and outputting a second parameter time sequence in a second time period after the current time through the prediction model; collecting an actual parameter time sequence of the refrigerator in the second time period, determining a residual error sequence according to the actual parameter time sequence and the second parameter time sequence, and calculating a residual error characteristic value of the residual error sequence; obtaining a door magnetic switch signal of the refrigerator door through a door magnetic sensor, and determining a door opening and closing state of the refrigerator door according to the door magnetic switch signal and the residual error characteristic value. The application can solve the problem that the existing refrigerator cannot identify door body virtual masking and chronic leakage, and accurately distinguish the door body state.

Description

Technical Field

[0001] This application relates to the field of refrigeration equipment technology, and in particular to a method for detecting the status of a refrigerator door, a controller, a refrigerator, a medium, and a program product. Background Technology

[0002] In related technologies, current mainstream refrigerators mostly use door magnetic trigger timer alarms to handle door status. This means that when the door magnetic sensor detects the refrigerator door has been open for more than a preset time, it will emit a buzzer as a reminder. However, existing solutions often have the following problems when it comes to food identification and management: the judgment logic is simplistic, relying solely on the door magnetic status and time. It cannot recognize situations where the refrigerator door is physically closed but not completely closed, i.e., there are tiny gaps. In such cases, the door magnetic sensor indicates the door is closed, but cold air continues to leak, causing the compressor to start frequently, resulting in significant energy waste. Furthermore, traditional solutions cannot detect and warn of chronic leaks caused by aging sealing strips until the user notices severe frost buildup or a lack of cooling. Summary of the Invention

[0003] This application aims to address at least one of the technical problems existing in the prior art. To this end, this application proposes a refrigerator door status detection method, controller, refrigerator, medium, and program product, aiming to solve the problem that existing refrigerators cannot identify door ajar or chronic leakage, and to achieve accurate differentiation of door status.

[0004] In a first aspect, embodiments of this application provide a method for detecting the state of a refrigerator door, the method comprising: Collect the historical operating parameters of the refrigerator for the first time period before the current moment, preprocess the historical operating parameters, and obtain the first parameter time series; The first parameter time series is input into a pre-trained prediction model, and the prediction model outputs the second parameter time series for the second time period after the current time. The prediction model is pre-trained based on the normal operating parameters of a standard refrigerator under various working conditions. Collect the actual parameter time series of the refrigerator during the second time period, determine the residual sequence based on the actual parameter time series and the second parameter time series, and calculate the residual characteristic value of the residual sequence; The door magnetic switch signal of the refrigerator door is obtained by a door magnetic sensor, and the door opening and closing state of the refrigerator door is determined based on the door magnetic switch signal and the residual characteristic value.

[0005] According to some embodiments of this application, the preprocessing of the historical operating parameters to obtain the first parameter time series includes: The historical operating parameters are subjected to median filtering based on a preset window length to remove abnormal peak parameters. The filtered historical operating parameters are normalized to a preset numerical range to obtain a standardized first parameter time series.

[0006] According to some embodiments of this application, the step of inputting the first parameter time series into a pre-trained prediction model and outputting the second parameter time series for a second time period after the current time through the prediction model includes at least one of the following: When the first parameter time series is the first temperature time series of the compartment, the first temperature time series is input into the pre-trained compartment temperature prediction model, and the compartment temperature prediction model outputs the second temperature time series of the second time period after the current time. Here, the first temperature time series and the second temperature time series are multiple compartment temperature parameters arranged in chronological order. When the first parameter time series is the first frequency time series of the compressor, the first frequency time series is input into the pre-trained compressor frequency prediction model, and the compressor frequency prediction model outputs the second frequency time series of the second time period after the current moment. The first frequency time series and the second frequency time series are multiple compressor frequency parameters arranged in chronological order. When the first parameter time series is the third frequency time series of the wind turbine, the third frequency time series is input into the pre-trained wind turbine frequency prediction model, and the wind turbine frequency prediction model outputs the fourth frequency time series of the second time period after the current time. The third frequency time series and the fourth frequency time series are multiple wind turbine frequency parameters arranged in chronological order.

[0007] According to some embodiments of this application, determining a residual sequence based on the actual parameter time series and the second parameter time series, and calculating the residual characteristic values ​​of the residual sequence, includes: For each time point, the residual value is calculated based on the actual parameter time series and the second parameter time series; The residual values ​​are arranged in chronological order to obtain a residual sequence; The moving average of all the residual values ​​in the residual sequence is obtained by performing a moving average process.

[0008] According to some embodiments of this application, determining the door opening / closing state of the refrigerator door based on the door magnetic switch signal and the residual characteristic value includes: The target dynamic threshold is obtained at each first preset time interval; The residual feature value is compared with the target dynamic threshold to obtain a numerical comparison result; Determine the level type of the door magnetic switch signal, and determine the door opening / closing state of the refrigerator door based on the comparison result of the level type and the value.

[0009] According to some embodiments of this application, the target dynamic threshold is obtained through the following steps: Determine the target time period during which the refrigerator is operating normally without any abnormal alarms; Obtain the mean and standard deviation of the residuals of the refrigerator's operating parameters over the target time period; The target dynamic threshold is determined based on the residual mean, the standard deviation, and the preset sensitivity coefficient, wherein the target dynamic threshold is the sum of the product of the standard deviation and the preset sensitivity coefficient and the residual mean.

[0010] According to some embodiments of this application, determining the door opening / closing state of the refrigerator door based on the level type and the numerical comparison result includes at least one of the following: When the level type is the first level type and the numerical comparison result indicates that the residual characteristic value is less than or equal to the target dynamic threshold, it is determined that the refrigerator door is in a normally closed state; When the level type is the first level type, and the numerical comparison result indicates that the residual characteristic value is greater than the target dynamic threshold and remains for a second preset time, it is determined that the refrigerator door is in a leaky state where the door is not closed tightly. When the level type is the second level type, and the numerical comparison result indicates that the residual characteristic value is less than or equal to the target dynamic threshold, and it is determined that no new food has been placed in the compartment, the refrigerator door is determined to be in a normal open state. When the level type is the second level type, and the numerical comparison result indicates that the residual characteristic value is greater than the target dynamic threshold, and it is determined that no new food has been placed in the compartment, it is determined that the refrigerator door is in a long-term open state.

[0011] According to some embodiments of this application, after acquiring the door magnetic switch signal of the refrigerator door through the door magnetic sensor, the method further includes one of the following: When the level type is the second level type and it is determined that new food is placed in the compartment, the incremental operating parameters of the refrigerator are collected during the third time period after the food is placed in. The end time of the third time period is taken as the second moment. Based on the incremental operating parameters, all or part of the historical operating parameters, and the new actual parameter time series collected by the refrigerator during the fourth time period after the second moment, the door opening and closing status of the refrigerator is determined. When the level type is the second level type and it is determined that new food is placed in the compartment, the determination of the door opening and closing status of the refrigerator door based on the door magnetic switch signal and the residual characteristic value is stopped.

[0012] According to some embodiments of this application, when the prediction model includes a compartment temperature prediction model, a compressor frequency prediction model, and a fan frequency prediction model, the method further includes: The first door opening / closing state determined by the compartment temperature prediction model, the second door opening / closing state determined by the compressor frequency prediction model, and the third door opening / closing state determined by the fan frequency prediction model are all determined. Determine the first weighting coefficient corresponding to the compartment temperature prediction model, the second weighting coefficient corresponding to the compressor frequency prediction model, and the third weighting coefficient corresponding to the fan frequency prediction model; A weighted average value is calculated based on the first weighting coefficient, the first door opening / closing state, the second weighting coefficient, the second door opening / closing state, the third weighting coefficient, and the third door opening / closing state. The weighted average value is compared with the preset value to obtain the final door opening and closing state determined after the fusion of multiple prediction models.

[0013] According to some embodiments of this application, the method further includes one of the following: If it is determined that the refrigerator door is in a leaking state due to not being closed properly or has been left open for an extended period of time, a prompt message will be generated via a speaker and / or a display screen. If it is determined that the refrigerator door is in a leaky state due to not being closed properly or has been left open for an extended period of time, an alarm message will be sent to the mobile terminal. The alarm message includes at least one of the following: refrigerator compartment name, time of occurrence, predicted temperature deviation, and prompt message.

[0014] According to some embodiments of this application, the prediction model is trained through the following steps: Auxiliary variables of a standard refrigerator under various operating conditions are obtained, wherein the auxiliary variables include at least one of the following: sensor data, operating status of the refrigeration equipment, and behavioral events; Obtain the dominant variables of the standard refrigerator, wherein the dominant variables include at least one of the following: compartment volume temperature and food surface temperature; The auxiliary variable and the dominant variable are input into the prediction model for training to obtain the trained prediction model.

[0015] According to some embodiments of this application, it also includes one of the following: The operating conditions include at least one of the following: no-load operating condition, half-load operating condition, and full-load operating condition; The sensor data includes at least one of the following: room interior temperature, ambient temperature, and evaporator temperature; The operating status of the refrigeration equipment includes at least one of the following: compressor status, fan speed, valve opening, defrosting status, and heater power; The behavioral events include at least one of the following: door open / closed status, cumulative door open duration, and door open frequency.

[0016] Secondly, embodiments of this application provide a controller, including: a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the refrigerator door status detection method of the first aspect described above when running the computer program.

[0017] Thirdly, embodiments of this application provide a refrigerator that includes the controller described in the second aspect above.

[0018] Fourthly, embodiments of this application provide a computer-readable storage medium storing computer-executable instructions for performing the refrigerator door status detection method as described in the first aspect above.

[0019] Fifthly, embodiments of this application provide a computer program product, including a computer program or computer instructions, wherein the computer program or computer instructions are stored in a computer-readable storage medium, a processor of a computer device reads the computer program or computer instructions from the computer-readable storage medium, and the processor executes the computer program or computer instructions, causing the computer device to perform the refrigerator door status detection method as described in the first aspect above.

[0020] According to the technical solution of this application embodiment, it has at least the following beneficial effects: First, this application embodiment collects the historical operating parameters of the refrigerator for a first time period before the current moment and performs preprocessing to remove interference information in the original data, obtaining a first parameter time series; then, this application embodiment inputs the first parameter time series into a pre-trained prediction model to predict the second parameter time series for a second time period after the current moment; next, this application embodiment collects the actual parameter time series of the refrigerator in the second time period, and calculates the residual sequence and residual characteristics by comparing the actual parameter time series with the predicted second parameter time series. The system uses a eigenvalue sensor to detect deviations in refrigerator operating parameters from normal conditions and promptly detects parameter fluctuations caused by refrigerator door malfunctions. Finally, this embodiment combines the door magnetic switch signal obtained by the door magnetic sensor with the calculated residual characteristic value to comprehensively judge the opening and closing status of the refrigerator door. This changes the traditional approach that relies solely on the door magnetic status and time for judgment, effectively identifying hidden faults such as the door magnetic sensor indicating it is closed but not actually closed tightly, or the presence of tiny gaps. It can also promptly detect chronic leaks caused by aging sealing strips, providing early warning of chronic leaks and avoiding energy waste caused by frequent compressor starts due to continuous cold air leakage.

[0021] Additional aspects and advantages of this application will be set forth in part in the description which follows, and in part will be obvious from the description, or may be learned by practice of this application. Attached Figure Description

[0022] The accompanying drawings are used to provide a further understanding of the technical solutions of this application and constitute a part of the specification. They are used together with the embodiments of this application to explain the technical solutions of this application and do not constitute a limitation on the technical solutions of this application.

[0023] Figure 1 This is a flowchart of a refrigerator door status detection method provided in one embodiment of this application; Figure 2 yes Figure 1 A flowchart of a sub-step of one embodiment of step S110; Figure 3 yes Figure 1 A flowchart of a sub-step of one embodiment of step S130; Figure 4 yes Figure 1 A flowchart of a sub-step of one embodiment of step S140; Figure 5 This is a flowchart illustrating the process of obtaining a target dynamic threshold according to an embodiment of this application; Figure 6 This is a flowchart of a method for determining the opening and closing status of a door using multi-model fusion, provided in one embodiment of this application; Figure 7 This is a flowchart of the training process of a prediction model provided in one embodiment of this application; Figure 8 This is a schematic diagram of an intelligent refrigerator door not fully closed detection system provided in one embodiment of this application; Figure 9 This is an overall flowchart of a refrigerator door status detection method provided in one embodiment of this application; Figure 10 This is a schematic diagram of a controller for performing a refrigerator door status detection method according to an embodiment of this application. Detailed Implementation

[0024] The embodiments of this application are described in detail below. Examples of the 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 are only used to explain this application, and should not be construed as limiting this application.

[0025] In the description of this application, it should be understood that the orientation descriptions, such as up, down, front, back, left, right, etc., indicate the orientation or positional relationship based on the orientation or positional relationship shown in the accompanying drawings. They are only for the convenience of describing this application and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation. Therefore, they should not be construed as limitations on this application.

[0026] In the description of this application, "several" means one or more, "more than" means two or more, "greater than," "less than," and "exceeding" are understood to exclude the stated number, while "above," "below," and "within" are understood to include the stated number. The use of "first" and "second" in the description is merely for distinguishing technical features and should not be construed as indicating or implying relative importance, or implicitly indicating the number of indicated technical features, or implicitly indicating the order of the indicated technical features.

[0027] In the description of this application, unless otherwise expressly defined, terms such as "setup," "installation," and "connection" should be interpreted broadly, and those skilled in the art can reasonably determine the specific meaning of the above terms in this application in conjunction with the specific content of the technical solution.

[0028] In some situations, current mainstream refrigerators primarily use door magnetic sensors to trigger timed alarms when the door is open for a preset time. This triggers an alarm with a beep. However, existing solutions often have the following problems when it comes to food identification and management: The logic is simplistic, relying solely on the door magnetic sensor status and time. This fails to recognize situations where the door is physically closed but not completely shut, meaning there are tiny gaps. In such cases, the sensor indicates the door is closed, but cold air continues to leak, causing the compressor to start frequently and resulting in significant energy waste. Furthermore, traditional solutions cannot detect and warn of chronic leaks caused by aging door seals until the user notices severe frost buildup or a lack of cooling.

[0029] Based on the above, this application proposes a refrigerator door status detection method, controller, refrigerator, medium, and program product, aiming to solve the problem that existing refrigerators cannot identify door ajar and chronic leakage, and to achieve accurate differentiation of door status.

[0030] The various embodiments of the refrigerator door status detection method of this application will be further described below with reference to the accompanying drawings.

[0031] like Figure 1 As shown, Figure 1 This is a flowchart of a refrigerator door state detection method according to an embodiment of this application. The refrigerator door state detection method includes, but is not limited to, steps S110, S120, S130 and S140.

[0032] Step S110: Collect the historical operating parameters of the refrigerator for the first time period before the current time, preprocess the historical operating parameters, and obtain the first parameter time series; Step S120: Input the first parameter time series into the pre-trained prediction model, and output the second parameter time series of the second time period after the current time through the prediction model. The prediction model is pre-trained based on the normal operating parameters of a standard refrigerator under various working conditions. Step S130: Collect the actual parameter time series of the refrigerator in the second time period, determine the residual sequence based on the actual parameter time series and the second parameter time series, and calculate the residual characteristic value of the residual sequence; Step S140: Obtain the door magnetic switch signal of the refrigerator door through the door magnetic sensor, and determine the door opening and closing status of the refrigerator door based on the door magnetic switch signal and the residual characteristic value.

[0033] In one embodiment, firstly, this application embodiment continuously acquires historical operating parameters of the refrigerator during a first time period prior to the current moment, such as internal temperature, compressor frequency, and other parameters related to the refrigerator's operating status. Since the collected historical operating parameters may contain noise data caused by environmental interference, sensor jitter, etc., it is necessary to preprocess the collected historical operating parameters to remove invalid interference data and arrange the historical operating parameters in chronological order to obtain a first parameter time series. Then, this application embodiment inputs the first parameter time series into a pre-trained prediction model, wherein the prediction model has been pre-trained offline using normal operating parameters of a standard refrigerator under various actual working conditions, fully learning and fitting the refrigerator's operating parameters. The model analyzes the parameter changes during normal refrigerator operation, enabling it to predict subsequent operating parameters. After receiving the first parameter time series, the prediction model deduces and outputs the second parameter time series that the refrigerator should have under normal operating conditions during the second time period following the current moment. Next, this embodiment collects the actual operating parameters of the refrigerator during the second time period, organizes them chronologically to form an actual parameter time series, compares this actual parameter time series with the second parameter time series, calculates the difference between the actual and predicted parameters at each corresponding moment to form a residual sequence, and further analyzes the residual sequence to calculate residual characteristic values. These residual characteristic values ​​reflect the degree of deviation between the actual operating state and the normal state. Finally, this embodiment uses a door magnetic sensor to collect the door magnetic switch signal in real time to determine the door's surface opening and closing state (e.g., closed to open), and then combines this with the residual characteristic values ​​to determine the door's opening and closing state.

[0034] It should be noted that the embodiments of this application change the traditional approach of relying solely on the status of the door magnet and time for judgment. It can effectively identify hidden faults such as the door magnet indicating that it is closed but is not actually closed tightly and there are tiny gaps. It can also detect chronic leakage problems caused by aging of the sealing strip in a timely manner, realize early warning of chronic leakage, and avoid energy waste caused by frequent compressor starts due to continuous leakage of cold air.

[0035] It is understood that the aforementioned historical operating parameters may be the internal air temperature, food surface temperature, compressor frequency, fan frequency, or other parameters related to the refrigerator's operating status. This application embodiment does not specifically limit these parameters.

[0036] In addition, it is understood that the residual characteristic value mentioned above can be the residual standard deviation, the residual root mean square or mean square error, the residual median, or other types of numerical values. This application does not specifically limit this.

[0037] In addition, such as Figure 2 As shown, Figure 2 yes Figure 1 A flowchart of the sub-steps of one embodiment of step S110; the preprocessing process in step S110 may include, but is not limited to, steps S210 and S220.

[0038] Step S210: Perform median filtering on historical operating parameters based on a preset window length to remove abnormal peak parameters; Step S220: Normalize the filtered historical operating parameters to a preset numerical range to obtain a standardized first parameter time series.

[0039] In one embodiment, firstly, according to a pre-set fixed window length, the historical operating parameters of the refrigerator collected are subjected to sliding median filtering to screen and remove abnormal peak parameters caused by sensor jitter or environmental equipment interference during the sampling process, filter out irregular abrupt changes in data, and retain the true trend of refrigerator operating parameters; then, the filtered historical operating parameters are normalized and converted to uniformly map and converge to a fixed preset value range, thereby finally generating the first parameter time series.

[0040] It should be noted that the embodiments of this application use median filtering to effectively filter out random noise and abnormal mutations in the original collected data, improve the authenticity and stability of historical operating parameters, and reduce the risk of false detection. Secondly, through normalization, the data can be standardized and regulated in a unified manner, which can effectively improve the accuracy of subsequent parameter prediction.

[0041] It should be noted that the prediction output process of the prediction model in step S120 above may include, but is not limited to, the following implementation scenarios, which are detailed below: The first implementation scenario: When the first parameter time series is the first temperature time series of the compartment, the first temperature time series is input into the pre-trained compartment temperature prediction model, and the compartment temperature prediction model outputs the second temperature time series of the second time period after the current time. Here, the first temperature time series and the second temperature time series are multiple compartment temperature parameters arranged in chronological order.

[0042] Specifically, when the first parameter time series is the first temperature time series of the compartment, this embodiment of the application employs a compartment temperature prediction model. Based on the temperature change patterns learned in the early stages of the model, such as normal refrigerator storage temperature, door opening and closing, and heat exchange, a second temperature time series can be predicted based on the first temperature time series. This embodiment of the application can intuitively reflect the temperature shift caused by abnormalities such as cold leakage through door gaps and aging seals through compartment temperature changes. Furthermore, because temperature data is easy to collect and closely matches the actual leakage faults in the refrigerator, abnormalities in the refrigerator's insulation can be quickly identified.

[0043] The second implementation scenario: When the first parameter time series is the first frequency time series of the compressor, the first frequency time series is input into the pre-trained compressor frequency prediction model, and the compressor frequency prediction model outputs the second frequency time series of the second time period after the current moment. Here, the first frequency time series and the second frequency time series are multiple compressor frequency parameters arranged in chronological order.

[0044] Specifically, when the first parameter time series is the first frequency time series of the compressor, this embodiment employs a compressor frequency prediction model. Based on the compressor start-up, shutdown, and frequency adjustment operation patterns learned in the early stages of the model under normal refrigerator load and normal sealing conditions, a second frequency time series can be predicted from the first frequency time series. This embodiment can directly reflect the refrigerator's cooling load changes through the compressor frequency, and is more sensitive to problems such as frequent compressor start-up and shutdown and high-frequency continuous operation caused by door leakage. It can identify hidden leakage anomalies from the perspective of the refrigeration unit's operating load.

[0045] The third implementation scenario: When the first parameter time series is the third frequency time series of the wind turbine, the third frequency time series is input into the pre-trained wind turbine frequency prediction model, and the wind turbine frequency prediction model outputs the fourth frequency time series of the second time period after the current moment. Here, the third frequency time series and the fourth frequency time series are multiple wind turbine frequency parameters arranged in chronological order.

[0046] Specifically, when the first parameter time series is the third frequency time series of the fan, this embodiment of the application will use a fan frequency prediction model. Based on the operating characteristics of the fan's air delivery and circulating heat dissipation under normal sealed environment learned in the early stage of the model, a fourth frequency time series can be predicted based on the third frequency time series. This embodiment of the application can reflect the adjustment state of the cold air circulation inside the box through the fan frequency. Since cold leakage through the door gaps can cause internal airflow turbulence and abnormal fan speed regulation, the prediction and comparison of fan operating parameters can help identify sealing abnormalities from the perspective of cold air circulation.

[0047] In addition, such as Figure 3 As shown, Figure 3 yes Figure 1 A flowchart of the sub-steps of one embodiment of step S130; the calculation process of the residual characteristic value in step S130 may include, but is not limited to, steps S310, S320 and S330.

[0048] Step S310: For each time point, calculate the residual value based on the actual parameter time series and the second parameter time series; Step S320: Arrange multiple residual values ​​in chronological order to obtain a residual sequence; Step S330: Perform a moving average on all residual values ​​in the residual sequence to obtain the moving average of the residual sequence.

[0049] In one embodiment, firstly, for each corresponding moment within the second time period, the real-time collected operating parameters are compared one by one with the predicted operating parameters to calculate the residual value corresponding to each moment node; then, all residual values ​​are arranged in chronological order to obtain a residual sequence; finally, a moving average calculation is performed on the entire residual sequence to smooth and correct multiple consecutive sets of residual values, ultimately outputting the moving average value of the residuals. This embodiment of the application, by calculating the residual value moment by moment, can understand the degree of deviation between the actual operating parameters and the predicted parameters, clearly showing the temporal variation characteristics of the parameter deviation; furthermore, the moving average smoothing process can effectively filter out interference and improve the accuracy of the residual characteristic values.

[0050] In addition, such as Figure 4 As shown, Figure 4 yes Figure 1 A flowchart of the sub-steps of one embodiment of step S140; the process of determining the opening and closing state of the door in step S140 may include, but is not limited to, steps S410, S420 and S430.

[0051] Step S410: Obtain the target dynamic threshold at each first preset time interval; Step S420: Compare the residual feature value with the target dynamic threshold to obtain the numerical comparison result; Step S430: Determine the level type of the door magnetic switch signal, and determine the opening and closing status of the refrigerator door based on the comparison result of the level type and the value.

[0052] In one embodiment, the present application will periodically acquire the updated target dynamic threshold according to a set first preset time, and then compare the calculated residual feature value with the target dynamic threshold to obtain a numerical comparison result; at the same time, identify and analyze the level type corresponding to the door magnetic switch signal to obtain the surface opening and closing signal of the refrigerator door, and finally combine the door magnetic level type and the numerical comparison result to make a comprehensive judgment to obtain the actual door opening and closing state of the refrigerator door.

[0053] It should be noted that, by periodically updating the target dynamic threshold, this embodiment can adapt to fluctuations in operating parameters caused by seasonal environment, equipment aging, and load changes, thus offering high flexibility. Secondly, by comparing the residual characteristic value with the target dynamic threshold, this embodiment can distinguish between normal parameter fluctuations and abnormal parameter deviations. Finally, by combining the door magnetic signal level type for fusion judgment, unlike the traditional single detection method that relies solely on the door magnetic signal, this embodiment can effectively identify hidden anomalies where the door magnetic indicator shows it is closed but it is not actually closed tightly, thus improving the accuracy of refrigerator door status detection.

[0054] It should be noted that the step S430 above, which involves determining the opening and closing state of the refrigerator door based on the comparison result of the voltage level type and value, can be implemented in several ways, including but not limited to the following: The first implementation scenario: When the level type is the first level type and the numerical comparison result indicates that the residual characteristic value is less than or equal to the target dynamic threshold, it is determined that the refrigerator door is in a normally closed state.

[0055] Specifically, if the door magnetic sensor outputs the first level type representing the refrigerator door being closed, and the residual characteristic value does not exceed the target dynamic threshold, it indicates that the refrigerator door has been reliably closed, and the actual changes of various operating parameters inside the refrigerator are basically consistent with the normal change patterns predicted by the model. The parameter deviation is within a reasonable fluctuation range. Therefore, the embodiments of this application can determine that the refrigerator door is in a normal closed state.

[0056] It should be noted that when a user first purchases and powers on the refrigerator, it is in brand-new condition and typically does not have a door that is not properly closed, thus preventing leakage. Therefore, the target dynamic threshold calculated upon initial power-on is a reference value calculated under the condition that the refrigerator door is properly closed. Furthermore, in subsequent prediction and judgment processes, if the predicted residual characteristic value under the closed-door condition is found to be greater than the target dynamic threshold and remains so for a second preset time, then the refrigerator door is considered to be in a leaky state, and the refrigerator will be returned to the factory for repair. Therefore, the target dynamic threshold will not be calculated based on a reference value under the condition of a leaky door, thus ensuring the accuracy and reliability of the target dynamic threshold.

[0057] The second implementation scenario: When the level type is the first level type, and the numerical comparison result indicates that the residual characteristic value is greater than the target dynamic threshold and is maintained for the second preset time, it is determined that the refrigerator door is in a leaky state where the door is not closed tightly.

[0058] Specifically, under the premise that the door magnetic signal is continuously at the first level type corresponding to the door being closed, when the residual characteristic value is continuously greater than the target dynamic threshold and the abnormal state is stably maintained until the second preset time, after eliminating the interference of instantaneous data disturbance and short-term operating condition fluctuation, it indicates that the parameters inside the refrigerator have a continuous abnormal deviation, which is not caused by normal operation. Therefore, it is determined that the refrigerator door appears to be closed but actually has gaps, and is in an abnormal operating condition of poor sealing and cold air leakage.

[0059] The third implementation scenario: When the level type is the second level type, and the numerical comparison result indicates that the residual characteristic value is less than or equal to the target dynamic threshold, and it is determined that no new food has been placed in the compartment, the refrigerator door is determined to be in the normal open state.

[0060] Specifically, when the door magnetic sensor outputs the second level type representing the refrigerator door being open, and the residual characteristic value is less than or equal to the target dynamic threshold, it indicates that although the user opened the refrigerator door to take out or put in food, the opening time was short, the temperature inside the refrigerator and the equipment operating parameters fluctuated within a limited range, and no significant loss of cold air occurred. Therefore, this embodiment of the application can be determined as normal door opening behavior.

[0061] The fourth implementation scenario: When the level type is the second level type, and the numerical comparison result indicates that the residual characteristic value is greater than the target dynamic threshold, and it is determined that no new food has been placed in the compartment, it is determined that the refrigerator door is in a long-term open state.

[0062] Specifically, based on the second level type corresponding to the door opening signal, if the residual characteristic value exceeds the target dynamic threshold, it means that the refrigerator door has been open for too long, hot air from the outside continues to enter the cabinet, the internal operating parameters deviate significantly from the normal range, and the cold energy loss increases significantly. Therefore, the embodiments of this application can determine that the refrigerator door is in an abnormal state of being open for a long time.

[0063] In one embodiment, the refrigerator door status detection method may further include the following steps: when the level type is the second level type and it is determined that new food is placed in the compartment, the incremental operating parameters of the refrigerator are collected during the third time period after the food is placed in the compartment. The end time of the third time period is taken as the second moment. Based on the incremental operating parameters, all or part of the historical operating parameters, and the new actual parameter time series collected by the refrigerator during the fourth time period after the second moment, the opening and closing status of the refrigerator door is determined.

[0064] Specifically, when the door magnetic signal is of the second level type representing the door being open, and new food is detected being placed inside the refrigerator compartment, the temperature inside the compartment will fluctuate due to the temperature of the new food itself, resulting in a normal deviation between the actual parameters and the predicted parameters. At this time, the system automatically collects the incremental operating parameters in the third time period after the food is placed inside, and takes the end point of the third time period as the second moment. The system integrates the newly added incremental operating parameters, all or part of the original historical operating parameters, and the new actual parameter time series collected in the fourth time period after the second moment, and re-participates in the data comparison and feature calculation to complete the subsequent determination of the refrigerator door's opening and closing status.

[0065] In one embodiment, when the level type is the second level type and it is determined that new food is placed in the compartment, the incremental operating parameters of the refrigerator are collected in the third time period after the food is placed. The incremental operating parameters and all or part of the historical operating parameters of the first time period are preprocessed to obtain the third parameter time series. The third parameter time series is input into a pre-trained prediction model, and the prediction model outputs the fourth parameter time series in the fourth time period after the second time. The new actual parameter time series of the refrigerator in the fourth time period is collected. A new residual sequence is determined based on the new actual parameter time series and the fourth parameter time series. The residual feature value of the new residual sequence is calculated. The door opening and closing state of the refrigerator is determined based on the residual feature value of the new residual sequence.

[0066] In one embodiment, the refrigerator door status detection method may further include the following steps: when the level type is the second level type and it is determined that new food has been placed in the compartment, stop determining the door opening and closing status of the refrigerator door based on the door magnetic switch signal and the residual characteristic value.

[0067] Specifically, when the door magnetic switch signal is the second level type representing the door being open, and it is detected that new food has been put into the refrigerator compartment, the temperature inside the compartment will fluctuate due to the temperature of the new food itself, which will cause a normal deviation between the actual parameters and the predicted parameters. Therefore, in order to prevent misjudgment, parameter prediction and refrigerator door status judgment can be stopped.

[0068] It is understood that, in this application embodiment, images of the compartment environment can be captured by a camera, and the presence or absence of new food can be determined by identifying and comparing these images; alternatively, the user can input food information through a terminal application and send it to the refrigerator, thereby enabling the refrigerator to detect whether new food has been placed in the compartment; other detection methods may also be used in this application embodiment, and this application embodiment does not specifically limit these methods.

[0069] In one embodiment, the above control logic is triggered only when the volume of the food ingredient is greater than a preset volume.

[0070] In addition, such as Figure 5 As shown, Figure 5 This is a flowchart of the process for obtaining a target dynamic threshold according to an embodiment of this application; the process for obtaining the target dynamic threshold may include, but is not limited to, steps S510, S520 and S530.

[0071] Step S510: Determine the target time period during which the refrigerator is operating normally and without any abnormal alarms; Step S520: Obtain the mean and standard deviation of the residuals of the refrigerator's operating parameters over the target time period; Step S530: Determine the target dynamic threshold based on the residual mean, standard deviation and preset sensitivity coefficient, wherein the target dynamic threshold is the sum of the product of the standard deviation and the preset sensitivity coefficient and the residual mean.

[0072] In one embodiment, firstly, this embodiment selects a target time period in which the refrigerator operates stably without human intervention and without generating any abnormal alarms. Then, within the selected target time period, residual data corresponding to the refrigerator's operating parameters can be continuously collected and statistically analyzed in the manner described above, and the corresponding residual mean and standard deviation can be calculated. These two parameters can reflect the overall distribution and fluctuation range of the residuals under normal operating conditions of the refrigerator. Next, this embodiment combines a preset sensitivity coefficient and calculates the latest target dynamic threshold in real time by multiplying the residual mean, standard deviation, and sensitivity coefficient.

[0073] It should be noted that, by selecting a stable period with low interference from the refrigerator for data statistics, the embodiments of this application can reduce data interference caused by human operation and sudden working conditions, thus ensuring the authenticity of the data. Secondly, the embodiments of this application use the residual mean and standard deviation, and flexibly adjust the severity of the judgment based on the sensitivity coefficient, so that the dynamic threshold can be automatically adapted and adjusted according to seasonal environment, equipment aging, and load changes. Finally, compared with a fixed threshold, the embodiments of this application can effectively improve the accuracy and flexibility of refrigerator door status detection.

[0074] In addition, such as Figure 6 As shown, Figure 6 This is a flowchart of a method for determining the opening and closing status of a door using a multi-model fusion method, provided in one embodiment of this application. When the prediction model includes a room temperature prediction model, a compressor frequency prediction model, and a fan frequency prediction model, the method may also include, but is not limited to, steps S610, S620, S630, and S640.

[0075] Step S610: Determine the opening and closing state of the first door as determined by the compartment temperature prediction model, determine the opening and closing state of the second door as determined by the compressor frequency prediction model, and determine the opening and closing state of the third door as determined by the fan frequency prediction model. Step S620: Determine the first weighting coefficient corresponding to the compartment temperature prediction model, determine the second weighting coefficient corresponding to the compressor frequency prediction model, and determine the third weighting coefficient corresponding to the fan frequency prediction model; Step S630: Calculate the weighted average value based on the first weighting coefficient, the first door opening / closing state, the second weighting coefficient, the second door opening / closing state, the third weighting coefficient, and the third door opening / closing state. Step S640: Compare the weighted average value with the preset value to obtain the final door opening / closing state determined after the fusion of multiple prediction models.

[0076] In one embodiment, firstly, this embodiment employs a compartment temperature prediction model, a compressor frequency prediction model, and a fan frequency prediction model to calculate the corresponding first door opening / closing state, second door opening / closing state, and third door opening / closing state, respectively. Next, this embodiment matches a first weighting coefficient, a second weighting coefficient, and a third weighting coefficient to the three prediction models, wherein the weighting coefficient values ​​are reasonably allocated based on the correlation and sensitivity of parameters such as compartment temperature, compressor frequency, and fan frequency to refrigerator door sealing abnormalities. Finally, this embodiment combines the door judgment results corresponding to each prediction model with the weighting coefficients to calculate a comprehensive weighted average value. Finally, this weighted average value is compared with a preset value to output the final door opening / closing state after verification by multiple prediction models.

[0077] It should be noted that the embodiments of this application use multiple models, including compartment temperature, compressor frequency, and fan frequency, to detect and determine the refrigerator's operation simultaneously. This is more accurate and reliable than single-parameter detection, and can detect refrigerator malfunctions from multiple dimensions, such as internal temperature changes, cooling load, and cold air circulation. Secondly, by using different weighting coefficients for fusion calculation, the determination role of important parameters, such as compartment temperature, can be highlighted, while the determination role of secondary parameters, such as compressor frequency and fan frequency, can be weakened. Finally, the multi-model weighted fusion determination method of the embodiments of this application can effectively improve the fault tolerance of door status detection, avoid the problems of misjudgment and omission by a single model, and can more accurately detect abnormal scenarios such as the refrigerator door not being closed tightly or the door being left open for a long time.

[0078] In one embodiment, the refrigerator door status detection method of this application further includes: generating a prompt message through a speaker and / or a display screen when it is determined that the refrigerator door is in a leaky state or has been open for a long time.

[0079] Specifically, when this application embodiment identifies an abnormal condition where the refrigerator door is not closed properly and is leaking or has been open for a long time, it will trigger a local reminder mechanism, which will respond by driving the speaker and / or display screen to issue an audio or visual prompt, and promptly provide feedback to the user on the current abnormal door status.

[0080] In one embodiment, the refrigerator door status detection method of this application further includes: when it is determined that the refrigerator door is in a state of leakage due to not being closed tightly or in a state of being open for a long time, sending alarm information to a mobile terminal, wherein the alarm information includes at least one of the following: refrigerator compartment name, occurrence time, predicted temperature deviation, and prompt message.

[0081] Specifically, when this application embodiment identifies that the refrigerator door is in a state of leakage due to not being closed tightly or an abnormal condition of being open for a long time, it will remotely push alarm information to the user's bound mobile terminal, informing the user of at least one of the following: refrigerator compartment name, time of abnormality, predicted temperature deviation, and text prompt.

[0082] In addition, such as Figure 7 As shown, Figure 7 This is a flowchart of the training process of a prediction model provided in one embodiment of this application; the training process of the prediction model may include, but is not limited to, steps S710, S720 and S730.

[0083] Step S710: Obtain auxiliary variables of a standard refrigerator under various operating conditions, wherein the auxiliary variables include at least one of the following: sensor data, operating status of the refrigeration equipment, and behavioral events; Step S720: Obtain the dominant variables of the standard refrigerator, wherein the dominant variables include at least one of the following: compartment volume temperature and food surface temperature; Step S730: Input the auxiliary variables and the dominant variables into the prediction model for training to obtain the trained prediction model.

[0084] In one embodiment, firstly, this application embodiment collects various data from a standard refrigerator under various operating conditions, i.e., auxiliary variables, including temperature data detected by various sensors, the operating status of components such as the compressor and fan, and usage behavior events such as opening and closing the door, thus collecting all kinds of real operating data under different operating conditions. Secondly, this application embodiment collects important temperature data from the standard refrigerator as dominant variables, including the overall average temperature of the refrigerator and the equivalent temperature of the food surface. Next, this application embodiment inputs the previously collected auxiliary data and important temperature data into a predictive model for continuous learning and training, allowing the model to identify and obtain the changing patterns between various operating data and the internal temperature and food temperature, ultimately training a mature and usable predictive model.

[0085] In one embodiment, the operating condition includes at least one of the following: no-load operating condition, half-load operating condition, and full-load operating condition.

[0086] Specifically, when training the prediction model, this embodiment of the application collects data under different storage conditions of the refrigerator: empty, half-loaded, and fully loaded. Empty means the refrigerator contains no food, half-loaded means it contains only a portion of food, and fully loaded means it is packed with food. Since different loading levels affect the amount of heat stored inside the refrigerator, the speed of cold air circulation, and the rate of temperature change, this embodiment of the application includes all these conditions in the training scope when training the prediction model. This allows the model to learn to adapt to different storage conditions, accurately predicting temperature changes regardless of the amount of food in the refrigerator, and ensuring that temperature control is more closely aligned with actual usage.

[0087] In one embodiment, the sensor data includes at least one of the following: room interior temperature, ambient temperature, and evaporator temperature.

[0088] Specifically, when training the prediction model, this embodiment of the application collects data from one or more of the following sensors as references: internal temperature of the refrigerator compartment, ambient temperature, and evaporator temperature. Since these temperatures affect the refrigerator's cooling effect and internal temperature changes, accurately reflecting the working status of the refrigeration components, external environmental interference, and the internal temperature conditions, this embodiment of the application incorporates this sensor data into the model during training. This allows the model to understand the relationships between different temperatures, resulting in more accurate temperature predictions and more realistic temperature control.

[0089] In one embodiment, the operating state of the refrigeration equipment includes at least one of the following: compressor status, fan speed, valve opening, defrosting status, and heater power.

[0090] Specifically, during the training of the prediction model, this embodiment of the application collects various operating status data of the refrigerator, including compressor operation status, fan speed, valve opening and closing, whether it is in defrost mode, and the power of the heating device. Since these operating parameters determine the cooling strength, cold air circulation speed, and defrosting rhythm, and significantly affect the internal temperature changes of the refrigerator, this embodiment of the application incorporates this operating data into the model training. This allows the model to understand the correlation between the device's operating status and temperature changes, thereby more accurately predicting subsequent temperature trends and making temperature control more reasonable.

[0091] In one embodiment, the behavioral event includes at least one of the following: door open / closed state, cumulative door opening duration, and door opening frequency.

[0092] Specifically, during the training of the prediction model, this embodiment of the application collects usage behavior data such as the refrigerator door opening and closing status, cumulative door opening time, and number of door openings. Since frequent door opening and prolonged door opening time allow hot air from outside to enter the refrigerator, causing the internal temperature to rise and the temperature fluctuations to increase, this embodiment of the application needs to incorporate these behavioral events into the model learning process when training the prediction model. This allows the model to learn about the temperature disturbances caused by door opening, adapt to daily usage habits in advance, more accurately predict temperature changes, and achieve more stable temperature control.

[0093] Based on the refrigerator door state detection methods of the above embodiments, the following presents overall embodiments of the system assembly and the refrigerator door state detection methods of this application.

[0094] This application provides a method and system for detecting a refrigerator door that is not closed properly based on edge computing and temperature prediction, in order to solve the problem that existing refrigerators cannot identify a slightly ajar door and chronic leaks, and to achieve early warning and accurate differentiation.

[0095] This application embodiment uses a lightweight timing prediction model deployed on the refrigerator's main control board (as an edge computing unit) to learn the dynamic temperature change pattern under normal refrigerator operation. By using residual analysis between the predicted temperature and the actual temperature, and combining it with the door magnet status, it accurately identifies the abnormal state where the door is closed but not fully closed.

[0096] Among them, such as Figure 8 As shown, Figure 8 This is a schematic diagram of an intelligent refrigerator door not fully closed detection system according to an embodiment of this application. The system includes, but is not limited to, a sensing layer 100, an edge computing layer 200, and an application layer 300.

[0097] The sensing layer 100 includes temperature sensors 110, such as NTC thermistors, deployed in the refrigerator compartment and freezer compartment; and a door magnetic sensor 120, such as a reed switch or Hall effect sensor, installed on the door. The temperature sensor 110 collects the internal temperature at a fixed frequency, such as every 60 seconds, and the door magnetic sensor 120 outputs the door's open / closed status in real time, where 0 indicates closed and 1 indicates open.

[0098] It should be noted that if the prediction scheme is based on temperature, the temperature sensor 110 mentioned above needs to be configured; if the prediction scheme is based on the frequency of the compressor or fan, the temperature sensor 110 mentioned above can be replaced with a Hall sensor, or a sensor scheme can be used for frequency detection, such as using a current or voltage sampling scheme.

[0099] Additionally, for the edge computing layer 200, i.e., the main control board of the refrigerator, a low-power MCU or an AIoT chip with an integrated NPU can be used. The algorithm of this embodiment runs on the main control board, including a data acquisition module 210, a preprocessing module 220, a temperature prediction model 230, a residual calculation module 240, a fusion judgment module 250, and an alarm generation module 260. The functions of each module are as follows: Data acquisition module 210: via It reads sensor data through interfaces such as GPIO and adds timestamps.

[0100] Preprocessing module 220: Performs median filtering on the raw temperature data to remove occasional noise, and then performs normalization processing to generate a standardized time series for model input.

[0101] Temperature prediction model 230: This is a lightweight LSTM model converted using TensorFlow Lite Micro, with fewer than 50KB of parameters, deployed in the MCU's Flash memory. This model learns the temperature variation patterns under normal refrigerator operating conditions (including compressor start-stop cycles, defrosting cycles, and temperature recovery curves after normal door opening and closing) and can predict future temperature trends based on historical temperatures. It should be noted that if a temperature-based prediction scheme is used, the above-mentioned temperature prediction model 230 needs to be configured; if a compressor or fan frequency-based prediction scheme is used, the above-mentioned temperature prediction model 230 can be replaced with a compressor frequency prediction model or a fan frequency prediction model.

[0102] Residual calculation module 240: compares the predicted temperature output by the model with the actual collected temperature point by point, calculates the residual sequence, and calculates the moving average to smooth short-term fluctuations.

[0103] Fusion Judgment Module 250: Combines the door magnetic state and residual, makes judgments according to state machine logic, and outputs the detection results.

[0104] Alarm generation module 260: Based on the judgment result, it controls the local buzzer 310 to emit prompts at different frequencies, or uploads the alarm information to the cloud server 330 via Wi-Fi, Bluetooth or 4G module, and then pushes it to the user's mobile APP 340.

[0105] Additionally, the application layer 300 includes a local buzzer 310, a display screen 320, a cloud server 330, and a user mobile app 340. The local buzzer 310 is used to instantly alert the user; the cloud server 330 is responsible for collecting abnormal event data and performing big data analysis; the user app can view alarm details and historical records, and set detection sensitivity.

[0106] Based on the above system, such as Figure 9 As shown, Figure 9 This is an overall flowchart of a refrigerator door status detection method provided in one embodiment of this application, which includes the following steps: Step 1: Data Acquisition and Preprocessing The refrigerator's main control board collects the refrigerator compartment temperature T(t) at a frequency of 1 minute, and simultaneously acquires the door magnetic signal D(t) in real time. Median filtering is applied to the temperature data with a window length of 3 to remove abnormal peaks. Then, the temperature values ​​are normalized to the [0,1] interval to obtain a standardized temperature time series.

[0107] Step 2: Temperature Time Series Prediction: Based on the current time t, extract the temperature data of the past 60 minutes, i.e., the window length N=60, to form the input sequence X = [T(t-59), ..., T(t)].

[0108] Input X into a temperature prediction model deployed in an MCU, such as a lightweight LSTM. The model outputs a predicted temperature sequence P = [P(t+1), P(t+2), ..., P(t+15)] for the next 15 minutes, i.e., K=15.

[0109] Before the refrigerator leaves the factory, the model has been pre-trained offline using operating data from a standard refrigerator under various conditions (such as no load, half load, full load, and different ambient temperatures) to learn the dynamic temperature characteristics under normal conditions. After deployment, the model parameters are fixed, but can be fine-tuned through incremental learning.

[0110] Step 3: Residual Calculation and Dynamic Threshold: When the actual time reaches t+1 to t+15, the corresponding actual temperature sequence A = [A(t+1), A(t+2),..., A(t+15)] is collected. The residual r(i) = |A(i) - P(i)| is calculated for each time step, i=1,...,15. To reduce the influence of occasional fluctuations, the moving average of the residual sequence R_avg = mean(r(1),...,r(15)) is calculated.

[0111] Setting the dynamic threshold Th: During periods when the refrigerator is operating normally and without abnormal alarms, such as 2:00 AM to 4:00 AM, the system continuously records the residual sequence and establishes a residual statistical database. The dynamic threshold Th = μ + k·σ, where μ is the mean residual value under normal conditions over the past 24 hours, σ is the standard deviation, and k is the sensitivity coefficient. This threshold is updated daily to adapt to seasonal changes and refrigerator aging.

[0112] Step 4: Fusion Logic Determination: The system combines the gate magnetic state D(t) and the residual moving average R_avg calculated at the current time to make the determination: State 1 (Normal Closed): If D=0 and R_avg ≤ Th, it means the door is closed and the actual temperature change matches the model's expectations (compressor starts and stops normally or defrosts), which is determined to be a normal state, and monitoring continues.

[0113] State 2 (Door not closed tightly / chronic leakage): If D=0 and R_avg>Th and this condition persists for more than 3 minutes (to prevent instantaneous disturbances), then the door is considered not closed tightly. In this case, although the door appears to be closed, due to the presence of a tiny gap, cold air continues to leak, causing the actual temperature to deviate from the normal insulation curve predicted by the model.

[0114] State 3 (Normal door opening): If D=1 and R_avg ≤ Th, it means that the user opens the door to retrieve items but the action is quick, and the temperature change inside the box is within the acceptable range predicted by the model. Only the door opening event is recorded, and no alarm is triggered.

[0115] State 4 (Long-term door open): If D=1 and R_avg>Th, it means that the door has been open for a long time, resulting in a large loss of cold air, triggering a regular door open timeout alarm, which is different from the specific prompt sound of the door not being closed properly.

[0116] Step 5: Alarms and Interactions When the alarm is detected as not being properly closed, the alarm generation module controls the buzzer to emit a specific warning sound of three short beeps followed by one long beep. Simultaneously, the alarm information (including compartment name, time of occurrence, predicted temperature deviation, etc.) is uploaded to the cloud server via the network module. The cloud server then pushes the information to the user's linked mobile app. The app displays a message stating "The refrigerator door may not be properly closed; please check the door seal," and provides access to historical anomaly records.

[0117] Based on the system and method of the above embodiments, the embodiments of this application have the following technical effects: 1. Detection of hidden faults: It can detect problems such as door ajar and door seal aging that cannot be detected by traditional door magnetic solutions, and provide early warning before obvious temperature abnormalities, effectively saving energy and protecting food; 2. Reduction of false alarms and disturbances: By understanding the normal thermodynamic behavior of the refrigerator through the model, it can distinguish between long-term normal opening for food preparation and leakage due to the door not being closed tightly, avoiding unnecessary alarms; 3. Low-power real-time detection: Utilizing edge computing, all inference is completed locally in the refrigerator without relying on the cloud, resulting in fast response speed and protection of user privacy; 4. Self-learning and self-adaptation: Through incremental learning, the refrigerator can adapt to changes in thermal characteristics under different seasons and loads, and long-term use does not require manual calibration of thresholds.

[0118] Based on the refrigerator door state detection method of the above embodiments, the following presents various embodiments of the controller, refrigerator, computer-readable storage medium and computer program product of this application.

[0119] like Figure 10 As shown, Figure 10 This is a schematic diagram of a controller for performing a refrigerator door status detection method according to an embodiment of this application. The controller 400 implemented in this application includes: a processor 410, a memory 420, and a computer program stored in the memory 420 and executable on the processor 410, wherein... Figure 10 The example uses a processor 410 and a memory 420.

[0120] The processor 410 and the memory 420 can be connected via a bus or other means. Figure 10 Taking the example of a connection between China and Israel via a bus.

[0121] Memory 420, as a non-transitory computer-readable storage medium, can be used to store non-transitory software programs and non-transitory computer-executable programs. Furthermore, memory 420 may include high-speed random access memory, and may also include non-transitory memory, such as at least one disk storage device, flash memory device, or other non-transitory solid-state storage device. In some embodiments, memory 420 may optionally include remotely located memories 420 relative to processor 410, which can be connected to controller 400 via a network. Examples of such networks include, but are not limited to, the Internet, intranets, local area networks, mobile communication networks, and combinations thereof.

[0122] Those skilled in the art will understand that Figure 10 The device structure shown does not constitute a limitation on the controller 400 and may include more or fewer components than shown, or combine certain components, or have different component arrangements.

[0123] exist Figure 10 In the controller 400 shown, the processor 410 can be used to call the control program stored in the memory 420 to implement the refrigerator door state detection method described above. Specifically, the non-transient software program and instructions required to implement the refrigerator door state detection method of the above embodiment are stored in the memory 420. When executed by the processor 410, the refrigerator door state detection method of the above embodiment is executed.

[0124] It is worth noting that since the controller 400 of this application embodiment can execute the refrigerator door state detection method of any of the above embodiments, the specific implementation method and technical effects of the controller 400 of this application embodiment can refer to the specific implementation method and technical effects of the refrigerator door state detection method of any of the above embodiments.

[0125] Furthermore, one embodiment of this application also provides a refrigerator that includes the controller described in the above embodiment.

[0126] It is worth noting that, since the refrigerator of this application embodiment includes the controller of the above embodiments, and the controller of the above embodiments is capable of executing the refrigerator door state detection method of any of the above embodiments, the specific implementation method and technical effects of the refrigerator of this application embodiment can refer to the specific implementation method and technical effects of the refrigerator door state detection method of any of the above embodiments.

[0127] Furthermore, one embodiment of this application also provides a computer-readable storage medium storing computer-executable instructions for performing the aforementioned refrigerator door status detection method. Exemplarily, the above-described method is executed... Figures 1 to 7 and Figure 9 The methods and steps in the text.

[0128] It is worth noting that, since the computer-readable storage medium of this application embodiment can execute the refrigerator door state detection method of any of the above embodiments, the specific implementation and technical effects of the computer-readable storage medium of this application embodiment can be referred to the specific implementation and technical effects of the refrigerator door state detection method of any of the above embodiments.

[0129] Furthermore, one embodiment of this application also provides a computer program product, including a computer program or computer instructions, which are stored in a computer-readable storage medium. A processor of a computer device reads the computer program or computer instructions from the computer-readable storage medium and executes the computer program or computer instructions, causing the computer device to perform the aforementioned refrigerator door status detection method. Exemplarily, the above-described method is executed... Figures 1 to 7 and Figure 9 The methods and steps in the text.

[0130] It is worth noting that, since the computer program product of this application embodiment can execute the refrigerator door state detection method of any of the above embodiments, the specific implementation method and technical effect of the computer program product of this application embodiment can refer to the specific implementation method and technical effect of the refrigerator door state detection method of any of the above embodiments.

[0131] It will be understood by those skilled in the art that all or some of the steps and systems in the methods disclosed above can be implemented as software, firmware, hardware, and suitable combinations thereof. Some or all of the physical components can be implemented as software executed by a processor, such as a central processing unit, digital signal processor, or microprocessor, or as hardware, or as an integrated circuit, such as an application-specific integrated circuit. Such software can be distributed on a computer-readable medium, which can include computer storage media (or non-transitory media) and communication media (or transient media). As is known to those skilled in the art, the term computer storage media includes volatile and non-volatile, removable and non-removable media implemented in any method or technology for storing information (such as computer-readable instructions, data structures, program modules, or other data). Computer storage media includes, but is not limited to, RAM, ROM, EEPROM, flash memory or other memory technologies, CD-ROM, digital versatile disc (DVD) or other optical disc storage, magnetic cartridges, magnetic tape, disk storage or other magnetic storage devices, or any other medium that can be used to store desired information and is accessible to a computer. Furthermore, as is known to those skilled in the art, communication media typically include computer-readable instructions, data structures, program modules, or other data in modulated data signals such as carrier waves or other transmission mechanisms, and may include any information delivery medium.

[0132] It should be understood that in this application, "at least one (item)" means one or more, and "more than" means two or more. "And / or" is used to describe the relationship between related objects, indicating that three relationships can exist. For example, "A and / or B" can represent three cases: only A exists, only B exists, and both A and B exist simultaneously, where A and B can be singular or plural. The character " / " generally indicates that the preceding and following related objects are in an "or" relationship. "At least one (item) of the following" or similar expressions refer to any combination of these items, including any combination of single or plural items. For example, at least one (item) of a, b, or c can represent: a, b, c, "a and b", "a and c", "b and c", or "a and b and c", where a, b, and c can be single or multiple.

[0133] In the several embodiments provided in this application, it should be understood that the disclosed systems, instruments, and methods can be implemented in other ways. For example, the instrument embodiments described above are merely illustrative; for instance, the division of units is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the shown or discussed mutual couplings, direct couplings, or communication connections may be through some interfaces; indirect couplings or communication connections between instruments or units may be electrical, mechanical, or other forms. Units described as separate components may or may not be physically separate, and components shown as units may or may not be physical units, i.e., they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.

[0134] It should also be understood that the various implementation methods provided in this application can be combined arbitrarily to achieve different technical effects.

[0135] The above provides a detailed description of the preferred embodiments of this application. However, this application is not limited to the above-described embodiments. Those skilled in the art can make various equivalent modifications or substitutions without departing from the spirit of this application. All such equivalent modifications or substitutions are included within the scope defined by the claims of this application.

Claims

1. A method for detecting the status of a refrigerator door, characterized in that, The method includes: Collect the historical operating parameters of the refrigerator for the first time period before the current moment, preprocess the historical operating parameters, and obtain the first parameter time series; The first parameter time series is input into a pre-trained prediction model, and the prediction model outputs the second parameter time series for the second time period after the current time. The prediction model is pre-trained based on the normal operating parameters of a standard refrigerator under various working conditions. Collect the actual parameter time series of the refrigerator during the second time period, determine the residual sequence based on the actual parameter time series and the second parameter time series, and calculate the residual characteristic value of the residual sequence; The door magnetic switch signal of the refrigerator door is obtained by a door magnetic sensor, and the door opening and closing state of the refrigerator door is determined based on the door magnetic switch signal and the residual characteristic value.

2. The method according to claim 1, characterized in that, The preprocessing of the historical operating parameters to obtain the time series of the first parameter includes: The historical operating parameters are subjected to median filtering based on a preset window length to remove abnormal peak parameters. The filtered historical operating parameters are normalized to a preset numerical range to obtain a standardized first parameter time series.

3. The method according to claim 1, characterized in that, The step of inputting the first parameter time series into a pre-trained prediction model and outputting the second parameter time series for a second time period after the current time through the prediction model includes at least one of the following: When the first parameter time series is the first temperature time series of the compartment, the first temperature time series is input into the pre-trained compartment temperature prediction model, and the compartment temperature prediction model outputs the second temperature time series of the second time period after the current time. Here, the first temperature time series and the second temperature time series are multiple compartment temperature parameters arranged in chronological order. When the first parameter time series is the first frequency time series of the compressor, the first frequency time series is input into the pre-trained compressor frequency prediction model, and the compressor frequency prediction model outputs the second frequency time series of the second time period after the current moment. The first frequency time series and the second frequency time series are multiple compressor frequency parameters arranged in chronological order. When the first parameter time series is the third frequency time series of the wind turbine, the third frequency time series is input into the pre-trained wind turbine frequency prediction model, and the wind turbine frequency prediction model outputs the fourth frequency time series of the second time period after the current time. The third frequency time series and the fourth frequency time series are multiple wind turbine frequency parameters arranged in chronological order.

4. The method according to claim 1, characterized in that, Based on the actual parameter time series and the second parameter time series, a residual sequence is determined, and the residual characteristic values ​​of the residual sequence are calculated, including: For each time point, the residual value is calculated based on the actual parameter time series and the second parameter time series; The residual values ​​are arranged in chronological order to obtain a residual sequence; The moving average of all the residual values ​​in the residual sequence is obtained by performing a moving average process.

5. The method according to claim 1, characterized in that, The step of determining the door opening / closing state of the refrigerator door based on the door magnetic switch signal and the residual characteristic value includes: The target dynamic threshold is obtained at each first preset time interval; The residual feature value is compared with the target dynamic threshold to obtain a numerical comparison result; Determine the level type of the door magnetic switch signal, and determine the door opening / closing state of the refrigerator door based on the comparison result of the level type and the value.

6. The method according to claim 5, characterized in that, The target dynamic threshold is obtained through the following steps: Determine the target time period during which the refrigerator is operating normally without any abnormal alarms; Obtain the mean and standard deviation of the residuals of the refrigerator's operating parameters over the target time period; The target dynamic threshold is determined based on the residual mean, the standard deviation, and the preset sensitivity coefficient, wherein the target dynamic threshold is the sum of the product of the standard deviation and the preset sensitivity coefficient and the residual mean.

7. The method according to claim 5, characterized in that, Determining the opening / closing state of the refrigerator door based on the level type and the numerical comparison result includes at least one of the following: When the level type is the first level type and the numerical comparison result indicates that the residual characteristic value is less than or equal to the target dynamic threshold, it is determined that the refrigerator door is in a normally closed state; When the level type is the first level type, and the numerical comparison result indicates that the residual characteristic value is greater than the target dynamic threshold and remains for a second preset time, it is determined that the refrigerator door is in a leaky state where the door is not closed tightly. When the level type is the second level type, and the numerical comparison result indicates that the residual characteristic value is less than or equal to the target dynamic threshold, and it is determined that no new food has been placed in the compartment, the refrigerator door is determined to be in a normal open state. When the level type is the second level type, and the numerical comparison result indicates that the residual characteristic value is greater than the target dynamic threshold, and it is determined that no new food has been placed in the compartment, it is determined that the refrigerator door is in a long-term open state.

8. The method according to claim 7, characterized in that, After acquiring the door magnetic switch signal of the refrigerator door via the door magnetic sensor, the method further includes one of the following: When the level type is the second level type and it is determined that new food is placed in the compartment, the incremental operating parameters of the refrigerator are collected during the third time period after the food is placed in. The end time of the third time period is taken as the second moment. Based on the incremental operating parameters, all or part of the historical operating parameters, and the new actual parameter time series collected by the refrigerator during the fourth time period after the second moment, the door opening and closing status of the refrigerator is determined. When the level type is the second level type and it is determined that new food is placed in the compartment, the determination of the door opening and closing status of the refrigerator door based on the door magnetic switch signal and the residual characteristic value is stopped.

9. The method according to claim 3, characterized in that, When the prediction model includes a compartment temperature prediction model, a compressor frequency prediction model, and a fan frequency prediction model, the method further includes: The first door opening / closing state determined by the compartment temperature prediction model, the second door opening / closing state determined by the compressor frequency prediction model, and the third door opening / closing state determined by the fan frequency prediction model are all determined. Determine the first weighting coefficient corresponding to the compartment temperature prediction model, the second weighting coefficient corresponding to the compressor frequency prediction model, and the third weighting coefficient corresponding to the fan frequency prediction model; A weighted average value is calculated based on the first weighting coefficient, the first door opening / closing state, the second weighting coefficient, the second door opening / closing state, the third weighting coefficient, and the third door opening / closing state. The weighted average value is compared with the preset value to obtain the final door opening and closing state determined after the fusion of multiple prediction models.

10. The method according to claim 1, characterized in that, The method also includes one of the following: If it is determined that the refrigerator door is in a leaking state due to not being closed properly or has been left open for an extended period of time, a prompt message will be generated via a speaker and / or a display screen. If it is determined that the refrigerator door is in a leaky state due to not being closed properly or has been left open for an extended period of time, an alarm message will be sent to the mobile terminal. The alarm message includes at least one of the following: refrigerator compartment name, time of occurrence, predicted temperature deviation, and prompt message.

11. The method according to claim 1, characterized in that, The prediction model is trained through the following steps: Auxiliary variables of a standard refrigerator under various operating conditions are obtained, wherein the auxiliary variables include at least one of the following: sensor data, operating status of the refrigeration equipment, and behavioral events; Obtain the dominant variables of the standard refrigerator, wherein the dominant variables include at least one of the following: compartment volume temperature and food surface temperature; The auxiliary variable and the dominant variable are input into the prediction model for training to obtain the trained prediction model.

12. The method according to claim 11, characterized in that: The operating conditions include at least one of the following: no-load operating condition, half-load operating condition, and full-load operating condition; The sensor data includes at least one of the following: room internal temperature, ambient temperature, and evaporator temperature; The operating status of the refrigeration equipment includes at least one of the following: compressor status, fan speed, valve opening, defrosting status, and heater power; The behavioral events include at least one of the following: door open / closed status, cumulative door open duration, and door open frequency.

13. A controller, characterized in that, include: The refrigerator door status detection method as described in any one of claims 1 to 12 is provided in the memory, the processor, and the computer program stored in the memory and executable on the processor.

14. A refrigerator, characterized in that, Includes the controller as described in claim 13.

15. A computer-readable storage medium, characterized in that: The refrigerator door is stored with computer-executable instructions for performing the refrigerator door status detection method as described in any one of claims 1 to 12.

16. A computer program product, comprising a computer program or computer instructions, characterized in that, The computer program or the computer instructions are stored in a computer-readable storage medium. The processor of the computer device reads the computer program or the computer instructions from the computer-readable storage medium and executes the computer program or the computer instructions, causing the computer device to perform the refrigerator door status detection method as described in any one of claims 1 to 13.

Images