Refrigeration appliance and load detection method thereof

Abstract

The invention discloses a refrigeration appliance and a load detection method thereof. The method comprises the following steps: obtaining refrigeration data of a target chamber; the refrigeration data are input into a preset machine learning model to obtain a load prediction value, and the load prediction value is used for representing space parameters of a real-time load in the target chamber; and sending and / or displaying the load prediction value. Through the scheme of the invention, real-time space load detection can be accurately realized under the condition that the hardware cost is not increased.

Description

TECHNICAL FIELD

[0001] Embodiments of the present application relate to the technical field of refrigeration appliances, in particular to a refrigeration appliance and a load detection method thereof. BACKGROUND

[0002] In daily life, people will use refrigeration appliances such as refrigerators to store perishable food such as vegetables and meat. The refrigeration appliance forms a low-temperature sealed space to prolong the shelf life of the food. Since the refrigerator door needs to be filled with foaming agent or other thermal insulation materials to achieve the cold preservation effect, the door of the existing refrigerator is usually opaque. This results in that after the user puts the food into the refrigerator, the user must open the refrigerator door to directly view the actual situation in the refrigerator chamber (for example, the refrigerating chamber), such as the space load of the refrigerating chamber. However, frequent opening and closing of the refrigerator door inevitably leads to loss of cold quantity in the refrigerating chamber, affecting the cold preservation effect of the refrigerator.

[0003] At present, there is a method of installing a camera in the refrigerating chamber and installing a display screen on the outer surface of the refrigerator door to display the shooting content of the camera, so that the user can know the space load condition in the refrigerating chamber without opening the refrigerator door. However, the camera is easily blocked by something such as a large plastic bag, so the way of reflecting the real-time space load of the refrigerating chamber by shooting a live image of the refrigerating chamber through the camera has low accuracy, and once the camera is blocked, the user still needs to manually open the refrigerator door to obtain the real and accurate space load information. In addition, the camera placed in the refrigerator needs to consider the performance of waterproofing and cold protection, resulting in high cost of a suitable camera, which is not conducive to popularization and application. SUMMARY

[0004] An object of embodiments of the present application is to provide an improved refrigeration appliance and a load detection method thereof.

[0005] Therefore, embodiments of the present application provide a load detection method of a refrigeration appliance, comprising: obtaining refrigeration data of a target compartment; inputting the refrigeration data into a preset machine learning model to obtain a load prediction value, wherein the load prediction value is used to represent a space parameter of a real-time load in the target compartment; and sending and / or displaying the load prediction value.

[0006] The prior art generally realizes the compartment load detection by adding a camera, but the camera is easily blocked, affecting the detection accuracy and having high implementation cost. In comparison, the present embodiment utilizes machine learning combined with refrigeration data to analyze the space usage in the target compartment, which can accurately realize real-time space load detection without increasing the hardware cost.

[0007] Further, the obtained load prediction value is displayed to the user, so that the user can timely and accurately know the real-time load in the target compartment without checking the target compartment by himself / herself, and the user experience is optimized. Further, the load prediction value can be sent to the user's mobile phone when the user is shopping, which helps the user to reasonably make a shopping plan.

[0008] Optionally, the refrigeration appliance includes at least one compartment, and the target compartment is a compartment in which the time interval from the last detected door opening / closing operation exceeds a first threshold. Thus, the load detection is performed after a period of time from the door opening / closing and after the refrigeration state in the compartment is stable, which is beneficial to exclude the influence of the cold / hot exchange between the inside and outside of the compartment caused by the door opening / closing and the recently put cold / hot food on the detection result, and improve the accuracy of the detection result.

[0009] Optionally, the obtaining the refrigeration data of the target compartment includes: receiving a detection instruction, the detection instruction being used to instruct to detect the load of the target compartment; obtaining the time of the last door opening / closing of the target compartment; judging whether the time interval from the time of the last door opening / closing exceeds a first threshold; and in response to the judgment result indicating that the time interval exceeds the first threshold, obtaining the refrigeration data of the target compartment. Thus, the user can operate on the interactive interface of the refrigerator or trigger remotely through the mobile phone to detect the load of the target compartment. Further, the load detection is performed in response to the user instruction on the premise that the time interval from the last door opening / closing of the target compartment is long enough, which is beneficial to ensure the accuracy of the provided detection result.

[0010] Optionally, the refrigeration appliance includes at least one compartment, and the number of the preset machine learning models is at least one and corresponds to the at least one compartment one by one, and each preset machine learning model is obtained based on the load test data of the corresponding compartment. The inputting the refrigeration data into the preset machine learning model to obtain the load prediction value includes: inputting the refrigeration data into the preset machine learning model corresponding to the target compartment. Different compartments have great differences in specific values of refrigeration data, for example, the refrigeration data (such as temperature) of the refrigerating compartment and the freezing compartment naturally have great differences. If a preset machine learning model is trained based on the refrigeration data of all compartments, it may not accurately represent the space use of each compartment. Therefore, the present embodiment trains a special preset machine learning model for different compartments, and the corresponding preset machine learning model is called to obtain a more accurate load detection result when the load of a specific target compartment is detected.

[0011] Optionally, the training process of the preset machine learning model comprises: obtaining a data set, wherein the data set comprises a plurality of groups of refrigeration data obtained when the target compartment is in a plurality of loading states, and different spatial parameters of the load in each loading state; training the preset machine learning model based on the data set until the prediction accuracy of the preset machine learning model reaches a second threshold. Thus, the parameters of the model are adjusted to the optimal by model training to obtain better prediction accuracy.

[0012] Optionally, the different spatial parameters of the load in the plurality of loading states are extracted at preset intervals. By extracting the spatial parameters of the load in different loading states at preset intervals, the data set is expanded and the model prediction accuracy is improved.

[0013] Optionally, a plurality of groups of the refrigeration data are collected in each loading state, wherein the types / quantities / qualities of the items placed in the target compartment corresponding to different groups of the refrigeration data are at least partially different. Thus, the sample data diversity is enriched and the model prediction accuracy is improved.

[0014] Optionally, the refrigeration data is selected from at least one of the following: the temperature, the humidity of the target compartment, the rotation speed of the fan associated with the target compartment, the opening time and the running time length. The cooling parameters of the target compartment can be obtained through the related parameters of the fan, and the space usage of the target compartment is analyzed in combination with the temperature and humidity information.

[0015] Optionally, the spatial parameters of the load are selected from: the proportion of the load in the total volume of the target compartment, the occupied volume of the load and the empty volume of the target compartment.

[0016] Therefore, the embodiment of the present application also provides a refrigeration appliance, comprising: a target compartment; at least one component arranged in the refrigeration appliance to adjust and / or monitor the refrigeration state of the target compartment; a control module in communication with the at least one component to obtain refrigeration data of the target compartment, wherein the control module executes the load detection method described above and sends and / or displays the load prediction value.

[0017] Compared with the prior art which needs to add a camera inside the refrigeration appliance to realize the compartment space load detection, the present embodiment utilizes the original components of the refrigerator to obtain refrigeration data, and utilizes machine learning to analyze the space usage in the target compartment based on the refrigeration data, so that the real-time space load in the target compartment can be accurately detected without additional hardware cost.

[0018] Optionally, the at least one component is selected from a temperature sensor, a humidity sensor, and a fan. The cooling parameter of the target compartment can be obtained based on the operating parameter of the fan, and the temperature and humidity information provided by the temperature sensor and the humidity sensor are combined to provide the control module with accurate analysis of the space usage of the target compartment.

[0019] Optionally, the refrigeration appliance further comprises a display module arranged on the outer surface of the refrigeration appliance, and the control module displays the load prediction value on the display module. In this way, the user can timely and accurately obtain the real-time load in the target compartment without opening the door, and the user operation is more convenient, thereby optimizing the user experience.

[0020] Optionally, the refrigeration appliance further comprises a communication module in communication with the control module, and the control module sends the load prediction value to a smart device through the communication module.

[0021] Optionally, the refrigeration appliance further comprises a refrigeration control module configured to control the cooling process of the target compartment based on the load prediction value. In this way, the refrigeration appliance can trigger a reasonable refrigeration cycle according to the load prediction value, which is conducive to saving the energy consumption of the refrigeration appliance while ensuring a better refrigeration effect.

[0022] Optionally, the refrigeration appliance further comprises a storage module configured to store the preset machine learning model, and the control module communicates with the storage module to call the preset machine learning model.

[0023] Optionally, the preset machine learning model comprises a multilayer perceptron neural network, and / or a decision tree, and / or a support vector machine. BRIEF DESCRIPTION OF DRAWINGS

[0024] Figure 1 is a flowchart of a load detection method of a refrigeration appliance according to an embodiment of the present application;

[0025] Figure 2 is Figure 1 a flowchart of one specific embodiment of step S101 in

[0026] Figure 3 is a schematic diagram of a refrigeration appliance according to an embodiment of the present application;

[0027] In the drawings:

[0028] 1 - refrigeration appliance; 10 - target compartment; 101 - freezing chamber; 102 - refrigerating chamber; 11 - door; 111 - temperature sensor; 112 - humidity sensor; 113 - fan; 12 - control module; 13 - storage module; 14 - communication module; 15 - display module. DETAILED DESCRIPTION

[0029] To make the above-mentioned objects, features and advantages of the present invention more apparent and understandable, specific embodiments of the present invention will be described in detail below with reference to the accompanying drawings.

[0030] Figure 1 This is a flowchart of a load detection method for a refrigeration appliance according to an embodiment of the present invention.

[0031] This implementation plan can be applied to scenarios involving the detection of space load within refrigeration appliances, specifically detecting the space utilization of the appliance's compartments. A compartment can be, for example, a freezer, refrigerator, or variable temperature compartment. For ease of description, the compartment being tested will be referred to as the target compartment. Load or load can refer to the stored items placed in the target compartment, such as food, flowers, cosmetics, etc., items whose shelf life needs to be extended by the temperature provided by the target compartment.

[0032] Specifically, refer to Figure 1 The load detection method for refrigeration appliances described in this embodiment may include the following steps:

[0033] S101, acquire the cooling data of the target room;

[0034] S102, the cooling data is input into a preset machine learning model to obtain a load prediction value, wherein the load prediction value is used to characterize the spatial parameters of the real-time load in the target room;

[0035] S103, send and / or display the load prediction value.

[0036] More specifically, the cooling data can be selected from at least one of the following: the temperature and humidity of the target room, the speed, operating time, and duration of the fan associated with the target room. The temperature of the target room can characterize the heating or cooling status of the target room in real time. The humidity of the target room can characterize the dryness of the target room in real time. The cooling appliance may include multiple rooms, each of which is equipped with a fan to adjust or control the cooling status of that room. The cooling parameters of the target room can be obtained through the relevant parameters of the fan, and the space utilization of the target room can be analyzed in combination with the temperature and humidity information.

[0037] Furthermore, the spatial parameters of the load can be selected from: the percentage of the target room's total volume occupied by the load, the volume occupied by the load, and the remaining volume of the target room. Assuming the total volume of the target room is 1 liter (L), and the volume occupied by the load within the target room is 0.5L, then the percentage of the target room's total volume occupied by the load is 50%, the volume occupied by the load is 0.5L, and the remaining volume of the target room is 1L - 0.5L = 0.5L.

[0038] In some embodiments, user instructions may be received and displayed and / or sent using the spatial parameters of the payload indicated by the user instructions.

[0039] Specifically, the display module of the refrigeration appliance can list the candidate items of the spatial parameter of the load, i.e., “the proportion of the load in the total volume of the target compartment”, “the volume occupied by the load”, and “the empty volume of the target compartment”. The user can directly select the desired candidate item on the display module and click OK to trigger the user instruction. In response to receiving the user instruction, the refrigeration appliance shows the load prediction value of the target compartment according to the spatial parameter of the load corresponding to the user-selected candidate item.

[0040] For example, assuming that the total volume of the target compartment is 1 liter (L), the volume occupied by the load in the target compartment is 0.1 L, and the current load prediction value of the target compartment displayed by the refrigeration appliance is 10%. The user operates on the display module to indicate that the empty volume of the target compartment is used to represent the load prediction value of the target compartment. In response to receiving the user instruction, the display module of the refrigeration appliance adjusts the displayed load prediction value of the target compartment to 0.9 L.

[0041] Alternatively, the candidate items of the spatial parameter of the load can be displayed on the smart device of the user, and the user can send the user instruction through the smart device. In response to receiving the user instruction sent by the smart device, the refrigeration appliance shows the load prediction value of the target compartment according to the spatial parameter of the load corresponding to the user-selected candidate item.

[0042] In one specific implementation, the refrigeration appliance can include at least one compartment. For example, the refrigeration appliance can include multiple compartments such as a freezer compartment, a refrigerator compartment, a fresh-keeping compartment, a variable-temperature compartment, etc. The refrigerator compartment can further include multiple independent compartments.

[0043] Further, the target compartment is the compartment in the at least one compartment for which the time interval from the detection time of the last detected opening / closing door operation reaches or exceeds a first threshold. In other words, in the actual application stage, the refrigeration data of the target compartment is collected after a period of time (i.e., the first threshold) from the last opening / closing door operation for the target compartment, and the load prediction is performed based on the preset machine learning model according to the refrigeration data.

[0044] For example, the opening / closing door operations for each compartment can be monitored. In response to detecting the opening / closing door operation for the freezer compartment at time t1, the control module of the refrigeration appliance can start timing. If the timing result shows that no opening / closing door operation for the freezer compartment is detected during the period of time of the first threshold from time t1, the control module collects the refrigeration data in the freezer compartment at time t2 (the time interval from time t1 is not less than the first threshold), and performs step S102 to obtain the load prediction value of the freezer compartment.

[0045] If the switch-door operation for the freezing chamber is detected again before the timer expires (the duration of the timer is the first threshold), the control module restarts the timer until the switch-door operation for the freezing chamber is not detected again before the timer expires.

[0046] In some embodiments, the first threshold can be from 1.5 to 2.5 hours. For example, the first threshold can be 2 hours.

[0047] In this way, the load detection is performed after a certain time interval after the door is switched, and after the refrigeration state in the compartment is stable, which is beneficial to exclude the influence of the cold and hot exchange between the inside and outside of the compartment caused by switching the door and the recently put cold / hot food on the detection result, and improve the accuracy of the detection result.

[0048] In some embodiments, different compartments can correspond to different first thresholds. Considering that different compartments have different requirements for the refrigeration state, and the speed of the compartment to reach a stable refrigeration state again after the door switching operation is also different, the first threshold can be set for each compartment. For example, the first threshold corresponding to the refrigerating chamber can be less than the first threshold corresponding to the freezing chamber.

[0049] In some embodiments, for the same compartment, the corresponding first threshold can be determined according to the time of the last door switching operation (i.e., the time interval between detecting the door opening operation and detecting the door closing operation). The longer the time of the last door switching operation, the greater the first threshold.

[0050] For example, after detecting the door opening operation for the refrigerating chamber, the door closing operation for the same refrigerating chamber is detected within a short time interval (e.g., 10 seconds), at which time it can be considered that the refrigeration state of the refrigerating chamber is not greatly affected by the door switching, and a smaller first threshold can be used. In this way, real-time load detection can be performed more timely for the refrigerating chamber after the door is closed (which means that the user may have updated the load in the refrigerating chamber).

[0051] For another example, if the door closing operation for the freezing chamber is detected after a long time interval (e.g., 1 minute) after the door opening operation for the freezing chamber is detected, it indicates that the cold quantity loss in the freezing chamber is relatively large at this time, and the freezing chamber may need to take a long time to reach a relatively stable refrigeration state again. Therefore, a larger first threshold can be used to ensure that the load prediction value obtained in S102 can accurately represent the space usage in the freezing chamber.

[0052] In one specific implementation, with reference to Figure 2 S101 can specifically include the following steps:

[0053] S1011, receiving a detection instruction, the detection instruction being used to instruct to detect the load of a target compartment;

[0054] Step S1012, acquiring a latest door opening / closing time of the target compartment;

[0055] Step S1013, determining whether a time interval from the latest door opening / closing time exceeds a first threshold value.

[0056] In response to a positive determination result of step S1013, i.e., the time interval exceeds the first threshold value, step S1014 is performed to acquire refrigeration data of the target compartment.

[0057] In response to a negative determination result of step S1013, i.e., the time interval does not exceed the first threshold value, step S1015 is performed to call a camera device arranged in the refrigeration appliance to capture a picture of the target compartment and display the picture.

[0058] Specifically, the detection instruction can be triggered by a user's operation on the display module or the smart device. The detection instruction can include a target compartment, for example, the user clicks an icon of the target compartment on the display module to trigger a load detection operation on the target compartment. Thus, the user can trigger the detection of the load condition of the target compartment on the interactive interface of the refrigerator or remotely through a mobile phone.

[0059] Further, in step S1012, the time point of the latest door opening / closing operation of the target compartment can be acquired and determined as the latest door opening / closing time.

[0060] Further, by performing the determination logic of step S1013, the load detection is performed in response to the user instruction on the premise that the latest door opening / closing time of the target compartment is long enough, which is beneficial to ensure the accuracy of the provided detection result.

[0061] If the determination result of step S1013 is negative, i.e., the time length from the latest door opening / closing operation of the target compartment is less than the first threshold value, the real-time picture of the target compartment can be displayed directly by calling the camera device to assist the user in understanding the space usage of the target compartment for the refrigeration appliance that has already arranged the camera device. Thus, the existing hardware in the refrigeration appliance can be fully utilized to meet the user's demand for understanding the space load of the target compartment from all angles. Further, the control module can continue to count time and execute step S1014 when the timer expires to provide the user with an accurate load prediction value.

[0062] In one specific implementation, the number of preset machine learning models can be at least one and correspond to at least one compartment one by one. For example, the refrigeration appliance includes two compartments, i.e., a refrigeration compartment and a freezer compartment, and accordingly, two preset machine learning models can be arranged in the refrigeration appliance, which correspond to the refrigeration compartment and the freezer compartment, respectively.

[0063] Further, each preset machine learning model is trained based on load test data of the corresponding compartment. That is, for each compartment, a dedicated dataset is constructed to train the preset machine learning model corresponding to the compartment. The load test data in the dataset can include a plurality of sets of corresponding refrigeration data obtained when the compartment is in a plurality of loading states (corresponding to the input of the preset machine learning model), and different spatial parameters of the load in each loading state (corresponding to the output of the preset machine learning model).

[0064] Different compartments have large differences in specific values of refrigeration data. For example, there are large differences in refrigeration data (such as temperature) between the refrigeration compartment and the freezer compartment. If a preset machine learning model is trained based on refrigeration data of all compartments, it may not accurately represent the space usage of each compartment. Therefore, the present embodiment trains a dedicated preset machine learning model for each different compartment, and invokes the corresponding preset machine learning model when performing load detection on a specific target compartment to obtain a more accurate load detection result.

[0065] Correspondingly, step S102 can include the step of inputting the refrigeration data into the preset machine learning model corresponding to the target compartment. For example, before / while / after obtaining the refrigeration data of the target compartment in step S101, the preset machine learning model corresponding to the target compartment is determined, and the obtained refrigeration data is input into the preset machine learning model to obtain a load prediction value.

[0066] In one specific implementation, the training process of the preset machine learning model can include the following steps: obtaining a dataset, wherein the dataset includes a plurality of sets of corresponding refrigeration data obtained when the target compartment is in a plurality of loading states, and different spatial parameters of the load in each loading state; training the preset machine learning model based on the dataset until the prediction accuracy of the preset machine learning model reaches a second threshold. In this way, the parameters of the model are adjusted to the optimal by model training to obtain better prediction accuracy.

[0067] In the dataset construction phase, different loading states and a list of items placed in the refrigeration appliance can be defined. The loading state can be represented by a loading ratio, such as 10%, 20%, 30%,..., 100%. The loading ratio can also be understood as the proportion of the load to the total volume of the target compartment. After placing the items in the target compartment of the refrigeration appliance according to different loading ratios, the refrigeration data is collected. After all tests are completed, the refrigeration data and the loading ratio are combined into a dataset for training the preset machine learning model.

[0068] Further, the spatial parameters of the load under the plurality of loading states can be extracted at preset intervals, for example, the loading ratio is extracted at 10% intervals from 0% (no load, also known as empty load) to 100% (full load) as the spatial parameters of the load under the plurality of loading states. By extracting the spatial parameters of the load under different loading states at preset intervals, the data set can be expanded, and the prediction accuracy of the model can be improved.

[0069] For example, the operating parameters of each component (such as temperature data of the temperature sensor, humidity data of the humidity sensor, rotation speed of the fan, opening time of the fan, and running time of the fan, etc.) can be collected when the freezing chamber is empty, and then the loading ratio of the freezing chamber is gradually increased to full load, during which the operating parameters of each component at each stage are collected, and then the data set is constructed.

[0070] Further, a plurality of sets of refrigeration data are collected under each loading state, wherein different sets of refrigeration data correspond to at least partially different types / quantities / qualities of items placed in the target chamber. For example, the refrigeration data of the refrigeration chamber before and after adjusting the quantity of the same type of items placed in the refrigeration chamber can be collected while keeping the 55% loading ratio unchanged. In this way, the diversity of sample data can be enriched, and the prediction accuracy of the model can be improved.

[0071] In the model training phase, the data set can be divided into a training set and a test set, a preset machine learning model is trained based on the training set, and the trained preset machine learning model is tested using the test set. The difference between the load prediction value obtained by inputting the refrigeration data in the test set into the preset machine learning model and the spatial parameter of the load corresponding to the refrigeration data recorded in the test set is used to determine the prediction accuracy of the preset machine learning model. The smaller the difference, the higher the prediction accuracy of the preset machine learning model.

[0072] Further, the second threshold value can be from 90% to 95%. For example, the second threshold value can be 93%.

[0073] In some embodiments, if the prediction accuracy of the preset machine learning model trained based on the existing data set cannot reach the second threshold value no matter how the parameters of the model are adjusted, new data can be collected to construct a new data set. By repeatedly training and collecting data until the prediction accuracy of the preset machine learning model reaches the target.

[0074] In some embodiments, after obtaining the dataset, the load test data in the dataset can be normalized, cleaned, and the like, before training the preset machine learning model based on the dataset, to achieve a denoising effect. For example, abnormal data in the obtained data can be removed, such as removing the temperature and the spatial parameters of the load corresponding to the loading state when the value of the temperature is obviously outside the normal temperature range of the target chamber. For another example, error data can be removed, such as data with packet loss during transmission.

[0075] In one specific implementation, the preset machine learning model can be constructed based on a machine learning algorithm. The machine learning algorithm can be selected from at least one of the following: decision tree, Naive Bayes classification, least squares regression, logistic regression, support vector machine, neural network, deep learning, multilayer perceptron (MLP), random forest algorithm, Extreme Gradient Boosting Decision Tree (XGBoost), and K-NearestNeighbor (KNN) classification algorithm, etc.

[0076] For example, the preset machine learning model can include a multilayer perceptron neural network. In the model training stage, three preset machine learning models can be respectively constructed based on the multilayer perceptron neural network, the decision tree, or the support vector machine, and then the three preset machine learning models are trained using the same dataset, and finally the preset machine learning model with the highest prediction accuracy is selected as the final application model.

[0077] For another example, the preset machine learning model can include a multilayer perceptron neural network and a decision tree. The multilayer perceptron neural network and the decision tree can be cascaded, and the former inputs the load prediction value and the refrigeration data to the latter, so that the latter verifies and corrects the load prediction value output by the former according to the refrigeration data.

[0078] For another example, the preset machine learning model can include a multilayer perceptron neural network, a decision tree, and a support vector machine. The three machine learning architectures can be cascaded, and the former inputs the load prediction value and the refrigeration data to the latter, so that the latter verifies and corrects the load prediction value output by the former according to the refrigeration data.

[0079] In summary, by using the machine learning to analyze the space usage in the target compartment in combination with the refrigeration data, the real-time space load detection can be accurately realized without increasing the hardware cost. Further, the obtained load prediction value is displayed to the user, so that the user can timely and accurately know the real-time load in the target compartment without checking the target compartment by himself / herself, thereby optimizing the user experience. Further, the load prediction value can be sent to the user's mobile phone when the user shops, which is helpful for the user to reasonably make a shopping plan.

[0080] Figure 3 is a schematic view of a refrigeration appliance 1 according to an embodiment of the present application. For more clearly showing the internal structure of the refrigeration appliance 1, Figure 3 The area surrounded by the door 11 of the refrigeration appliance 1 is exemplarily framed by a dashed line, and the specific situation of the internal compartment of the refrigeration appliance 1 is shown in a perspective form.

[0081] Specifically, referring to Figure 3 The refrigeration appliance 1 according to the embodiment can include a target compartment 10. For example, the refrigeration appliance 1 can include a freezing compartment 101 and a refrigerating compartment 102. For example, the refrigeration appliance 1 can include a refrigerator, a refrigerating cabinet, a freezing cabinet, etc.

[0082] Further, the refrigeration appliance 1 can include at least one component arranged in the refrigeration appliance 1 to adjust and / or monitor the refrigeration state of the target compartment 10. In some embodiments, the at least one component can be selected from: a temperature sensor 111, a humidity sensor 112, and a fan 113. For example, the top of the freezing compartment 101 can be provided with the temperature sensor 111 and the humidity sensor 112 for collecting the temperature and humidity of the freezing compartment 101, and the fan 113 can be arranged at the back wall of the freezing compartment 101. A valve can be further arranged in front of the fan 113 to control whether the cold energy delivered by the compressor of the refrigeration appliance 1 can enter the freezing compartment 101. For another example, the top of the refrigerating compartment 102 is provided with the temperature sensor 111, and the side wall is provided with the humidity sensor 103. The fan 113 corresponding to the refrigerating compartment 102 can be arranged at the back plate of the air duct behind the refrigerating compartment 102 and thus invisible in the figure.

[0083] Further, the refrigeration appliance 1 can further include a control module 12 in communication with the at least one component to obtain the refrigeration data of the target compartment 10. In response to the refrigerating compartment 102 being the target compartment 10, the control module 12 can obtain the operating parameters of the temperature sensor 111, the humidity sensor 112, and the fan 113 arranged in the refrigerating compartment 102 to obtain the refrigeration data.

[0084] Further, in response to obtaining the refrigeration data of the target compartment 10, the control module 12 can perform the above-mentioned Figure 1 and Figure 2The load detection method and the load prediction value are sent and / or displayed. Specifically, the cooling parameter of the target room can be obtained based on the operating parameter of the fan 113, and the temperature and humidity information provided by the temperature sensor 111 and the humidity sensor 112 are combined to provide the control module 12 with accurate analysis of the space usage of the target room 10.

[0085] The control module 12 can include a main control board / single-chip microcomputer of the refrigeration appliance 1. Alternatively, the control module 12 can be a module specially used for performing the above-mentioned Figure 1 and Figure 2 the method.

[0086] In a specific implementation, the refrigeration appliance 1 can further include a storage module 13 for storing a preset machine learning model, and the control module 12 communicates with the storage module 13 to call the preset machine learning model.

[0087] In some embodiments, the storage module 13 can be integrated in the refrigeration appliance 1, such as being integrated with the control module 12 or being independent of the control module 12.

[0088] For example, the storage module 13 can be integrated in the control module 12, and after the preset machine learning model is trained, the model can be deployed to the control module 12. The control module 12 is coupled to at least one component to obtain refrigeration data, and inputs the data into the preset machine learning model to obtain the load prediction value.

[0089] In a variant, the storage module 13 can be an external storage such as a cloud, and the control module 12 calls the preset machine learning model by accessing the storage module 13.

[0090] In a specific implementation, the refrigeration appliance 1 can further include a communication module 14 in communication with the control module 12, and the control module 12 sends the load prediction value to a smart device through the communication module 14.

[0091] Specifically, the smart device can be a mobile terminal such as a mobile phone or an IPAD, and can also be another smart home appliance in the same local area network as the refrigeration appliance 1.

[0092] For example, the communication module 14 can communicate with the smart device of the user to send the load prediction value to the smart device for the user to browse at any time and at random.

[0093] In some embodiments, the communication module 14 can communicate with the smart device through wireless communication technologies such as Wireless Fidelity (WIFI) and Near Field Communication (NFC).

[0094] Further, the control module 12 can also communicate with the at least one component through the communication module 14.

[0095] In one implementation, the refrigeration appliance 1 can further comprise a display module 15 disposed on an outer surface of the refrigeration appliance 1, and the control module 12 displays the load prediction value on the display module 15.

[0096] Specifically, the display module 15 can be disposed on a front surface of the door 11, i.e. the surface of the door 11 facing the user when the door 11 closes the target compartment 10. Alternatively, the display module 15 can be disposed on an outer surface of a cabinet of the refrigeration appliance 1, and the cabinet is adapted to form the target compartment 10.

[0097] Further, the refrigeration appliance 1 can comprise an interactive panel to realize human-computer interaction. The interactive panel can comprise a touchpad, a display screen, etc. The interactive panel can comprise the display module 15. In some embodiments, the display module 15 can be a User Interface Module (UIM) to display setting parameters or states of the refrigeration appliance 1, the load prediction value and / or receive a control instruction input by a user.

[0098] Further, the interactive panel can further comprise a cover plate (not shown in the figure) disposed in front of the display module 15 to at least play a protective role. The cover plate can be, for example, a front panel of the door 11, or can be, for example, a glass plate independently provided in front of the UIM. In some embodiments, the display module 15 can be closely attached to a rear side of the cover plate, so as to, for example, receive an input signal input by a user via touching the cover plate. The display module 15 can comprise a light-emitting piece to make light pass through a corresponding area of the cover plate.

[0099] In some embodiments, the control module 12 can directly display the load prediction value on the display module 15. Further, the control module 12 can receive a detection instruction input by a user through the display module 15 to detect a space usage of the target compartment 10 specified by the user.

[0100] For example, the display module 15 can be a touch screen, and the user triggers the detection instruction through a touch operation.

[0101] For another example, the interactive panel can comprise a gesture sensing area, and the user triggers the detection instruction through a gesture operation.

[0102] In this way, the user can timely and accurately know the real-time load in the target compartment 10 without opening the door 11 to check the target compartment 10, and the user operation is more convenient, thereby optimizing the user experience.

[0103] It should be noted that, Figure 3The possible setting positions of the control module 12, the at least one component, the communication module 14 and the display module 15 on the refrigeration appliance 1 are only exemplary, and in actual applications, the mutual position relationship and the specific setting position on the refrigeration appliance 1 of each module can be adjusted as needed. Each module can be independent of each other, or integrated in the same chip or integrated into the same functional module. For example, the control module 12 and the communication module 14 can be integrated together.

[0104] In one specific implementation, the refrigeration appliance 1 can further include a refrigeration control module (not shown in the figure) for controlling the cooling process of the target compartment 10 based on the load prediction value.

[0105] For example, the refrigeration control module can include the aforementioned valve for controlling the amount of cold delivered to the target compartment 10 via the fan 113. In this way, the degree of refrigeration in the target compartment 10 can be adjusted.

[0106] For another example, the refrigeration control module can include components in the refrigeration system of the refrigeration appliance 1. The control module 12 adjusts the operating parameters of the components according to the load prediction value to control the intensity, time, etc. of the cooling process for the target compartment 10.

[0107] In this way, the refrigeration appliance 1 can trigger a reasonable refrigeration cycle according to the load prediction value, which is conducive to saving the energy consumption of the refrigeration appliance 1 while ensuring better refrigeration effect.

[0108] In one typical application scenario, continuing to refer to Figure 3 , it is assumed that the freezer compartment 101 and the refrigerator compartment 102 are both target compartments 10, and the control module 12 performs the above Figure 1 and Figure 2 described solutions to obtain the space load conditions in the freezer compartment 101 and the refrigerator compartment 102, respectively.

[0109] For the freezer compartment 101, the control module 12 can determine whether the last opening and closing time of the door 11 of the freezer compartment 101 has exceeded a first threshold value. In response to the determination result indicating that the last time the door 11 was closed has exceeded the first threshold value, the control module 12 obtains the temperature data of the temperature sensor 111, the humidity data of the humidity sensor 112, and the operating parameters of the fan 113, respectively. The control module 12 accesses the storage module 13 to call the preset machine learning model A corresponding to the freezer compartment 101, and inputs the obtained refrigeration data into the preset machine learning model A to obtain the load prediction value.

[0110] Assuming the load prediction value is 10%, i.e., the load (such as the illustrated durians) accounts for 10% of the total volume of the freezer compartment 101, the control module 12 displays the following information on the display module 15: "Freezer real-time space load: 10%". Optionally, the control module 12 controls the operating parameters of the refrigeration control module related to the freezer compartment 101, taking into account that the load in the freezer compartment 101 is small, and only a light cooling process is needed to keep the freezing state of the freezer compartment 101 within the preset refrigeration range. Further, the freezing state of the freezer compartment 101 can be maintained at the upper limit of the preset refrigeration range to further reduce the energy consumption of the refrigeration appliance 1, saving energy and protecting the environment. Among them, the preset refrigeration range is the numerical range of various parameters (such as temperature, humidity) required for the freezer compartment 101 to achieve the freezing function. The larger the value, the worse the freezing effect.

[0111] For the refrigeration compartment 102, after the user opens the door 11 at time t1 and puts a plate of hot roast chicken into the refrigeration compartment 102, the user operates the mobile phone at time t2 to trigger the load detection of the refrigeration compartment 102. In response to receiving the detection instruction via the communication module 14, the control module 12 can determine whether the time interval between time t1 and time t2 exceeds the first threshold. In response to the determination result indicating that the time interval between time t1 and time t2 has exceeded the first threshold, the control module 12 respectively acquires the temperature data of the temperature sensor 111, the humidity data of the humidity sensor 112, and the operating parameters of the fan 113. The control module 12 accesses the storage module 13 to call the preset machine learning model B corresponding to the refrigeration compartment 102, and inputs the acquired foregoing refrigeration data into the preset machine learning model B to obtain the load prediction value.

[0112] Assuming the load prediction value is 15%, i.e., the load (such as the illustrated fish and roast chicken) accounts for 15% of the total volume of the refrigeration compartment 102, the control module 12 displays the following information on the display module 15: "Refrigeration compartment real-time space load: 15%". Optionally, the control module 12 controls the operating parameters of the refrigeration control module related to the refrigeration compartment 102, taking into account that the load in the refrigeration compartment 102 is large, and the cooling process for the refrigeration compartment 102 can be appropriately enhanced to ensure that the refrigeration state of the refrigeration compartment 102 remains within the preset refrigeration range that meets the refrigeration requirements.

[0113] Assuming that the user views the display module 15 and finds that the load space parameter display method of "Refrigeration compartment real-time space load: 15%" is not liked, the user can select to change the display method of the load space parameter on the display module 15. For example, after adjustment, the control module 12 displays the following information on the display module 15: "Refrigeration compartment empty volume: 0.8L".

[0114] From the above, by using the original components of the refrigeration appliance 1 to obtain the refrigeration data, and by using machine learning to analyze the space usage in the target chamber 10 based on the refrigeration data, the real-time space load in the target chamber 10 can be accurately detected without additional hardware costs.

[0115] Although specific embodiments have been described above, these embodiments are not intended to limit the scope of the present disclosure, even if a single embodiment is described in relation to only a particular feature. The feature examples provided in the present disclosure are intended to be illustrative rather than restrictive and unless otherwise specified the examples in one aspect of the disclosure can be applied to any other aspect of the disclosure, in particular though not exclusively, by substituting technical features thereof. In particular, the technical features of one or more dependent claims can be combined with the technical features of the independent claims and the technical features from corresponding independent claims can be combined in any appropriate manner rather than only the specific combinations listed in the claims.

[0116] Although the present application has been disclosed in its currently best embodiment with reference to the drawings, it will be apparent that modifications and / or deformations (especially regarding the dimensions of the elements), equivalent elements, and / or improvements thereof deemed by those skilled in the art, can be made therein without departing from the spirit and scope of the application and that the scope of the application should be limited only by the following claims.

Claims

1. A load detection method of a refrigeration appliance, characterized by, The method comprises: obtaining refrigeration data of a target compartment; inputting the refrigeration data into a preset machine learning model to obtain a load prediction value, wherein the load prediction value is used to represent a spatial parameter of a real-time load in the target compartment; sending and / or displaying the load prediction value.

2. The load detection method according to claim 1, wherein The refrigeration appliance comprises at least one compartment, and the target compartment is a compartment in which a time interval from a last detected door opening operation to the present time exceeds a first threshold.

3. The load detection method according to claim 1, wherein The method of obtaining refrigeration data of a target compartment comprises: receiving a detection instruction for instructing to detect a load of a target compartment; obtaining a last door opening time of the target compartment; determining whether a time interval from the last door opening time to the present time exceeds a first threshold; in response to a determination result indicating that the time interval exceeds the first threshold, obtaining refrigeration data of a target compartment.

4. The load detection method according to claim 1, wherein The refrigeration appliance comprises at least one compartment, and the number of the preset machine learning models is at least one and corresponds to the at least one compartment one by one, and each of the preset machine learning models is trained based on load test data of a corresponding compartment, and the inputting of the refrigeration data into a preset machine learning model to obtain a load prediction value comprises: inputting the refrigeration data into a preset machine learning model corresponding to the target compartment.

5. The load detection method according to claim 1, wherein The training process of the preset machine learning model comprises: obtaining a data set, wherein the data set comprises a plurality of sets of refrigeration data obtained when the target compartment is in a plurality of loading states, and different spatial parameters of a load in each loading state; and training the preset machine learning model based on the data set until a prediction accuracy of the preset machine learning model reaches a second threshold.

6. The load detection method according to claim 5, wherein The different spatial parameters of the load in the plurality of loading states are obtained at a preset interval; and / or, a plurality of sets of the refrigeration data are collected in each loading state, wherein the types / quantities / qualities of the items put into the target compartment corresponding to different sets of the refrigeration data are at least partially different.

7. The load detection method according to claim 1, wherein The refrigeration data is selected from at least one of the following: a temperature, a humidity, a rotation speed of a fan associated with the target compartment, an opening time, and a running time length of the target compartment.

8. The load detection method according to claim 1, wherein The spatial parameter of the load is selected from: a proportion of the load in a total volume of the target compartment, a volume occupied by the load, and a remaining volume of the target compartment.

9. A refrigeration appliance characterized in that, The method comprises: a target compartment (10); at least one component arranged in the refrigeration appliance to adjust and / or monitor a refrigeration state of the target compartment (10); a control module (12) in communication with the at least one component to obtain refrigeration data of the target compartment (10), wherein the control module (12) performs the load detection method of any one of claims 1 to 8, and sends and / or displays the load prediction value.

10. The refrigeration appliance of claim 9, wherein, The at least one component is selected from: a temperature sensor (111), a humidity sensor (112), and a fan (113).

11. The refrigeration appliance of claim 9, wherein, The method further comprises: a display module (15) arranged on an outer surface of the refrigeration appliance, wherein the control module (12) displays the load prediction value on the display module (15); and / or ​ a communication module (14) in communication with the control module (12), the control module (12) sending the load prediction value to a smart device through the communication module (14); and / or a refrigeration control module controlling a cooling process of the target inter-room (10) based on the load prediction value.

12. The refrigeration appliance of claim 9, wherein, Further comprising: a storage module (13) for storing the preset machine learning model, the control module (12) being in communication with the storage module (13) to call the preset machine learning model.

13. The refrigeration appliance of any of claims 9-12, wherein, The preset machine learning model comprises a multi-layer perceptron neural network, and / or a decision tree, and / or a support vector machine.

Images