Air-cooled refrigerator with low-humidity accurate adjusting function and control method of air-cooled refrigerator

Abstract

The invention discloses an air-cooled refrigerator with a low-humidity accurate adjusting function and a control method thereof, and the control method comprises the following steps: obtaining food moisture content detection data and food type information, and constructing a layered food state feature input set; acquiring environment parameter information, and constructing a humidity control auxiliary feature input set; inputting the input set into a pre-constructed AI humidity control decision model, and generating humidity control strategy output; executing dehumidification control operation according to the humidity control strategy output, and calculating a drying trend coefficient during execution; strategy offset correction is carried out on the AI humidity control decision model based on the drying trend coefficient; after strategy offset correction, the effectiveness of the humidity control strategy is further comprehensively judged based on the current environment and food state indexes, and if the judgment result is invalid, the control strategy is returned and regenerated; otherwise, the current humidity control strategy execution state is maintained. The humidity control precision, the fresh-keeping effect and the safety stability of the refrigeration area can be improved.

Description

Technical Field

[0001] This invention relates to the field of air-cooled refrigerator technology, and in particular to an air-cooled refrigerator with precise low humidity regulation function and its control method. Background Technology

[0002] With the improvement of living standards and the upgrading of food consumption patterns, consumers have placed higher demands on the preservation effect of refrigerated foods, especially in air-cooled refrigerators, where low humidity environments are widely used to inhibit bacterial growth, delay spoilage, and extend shelf life. However, in existing air-cooled refrigerators, low humidity control generally adopts dehumidification modes with fixed operating parameters or on / off control methods based on feedback from a single humidity sensor. This lacks the ability to adapt in real time to the different moisture-sensitive characteristics and moisture content changes of various foods, easily leading to two types of problems: on the one hand, excessive dehumidification accelerates the dehydration of food, causing it to harden, lose weight, and even degrade in flavor; on the other hand, insufficient dehumidification may cause localized high humidity, leading to mold growth or quality deterioration. Although some high-end air-cooled refrigerators have introduced humidity monitoring based on multiple sensors, their control strategies still rely on fixed parameter adjustments. They cannot self-correct their strategies based on the dynamic changes in food moisture content and lack real-time evaluation of the effectiveness of the humidity control strategy. Once external environmental conditions, user habits, or food types change, the original control strategy may quickly become ineffective, leading to the accumulation of humidity control deviations and affecting long-term preservation stability.

[0003] Furthermore, traditional air-cooled refrigerators typically employ a unified humidity control system across the entire compartment, using a single air duct system to supply air and dehumidify the entire refrigerator area. This approach ignores the differences in humidity requirements, moisture content variations, and other characteristics of food items stored in different compartments. For example, leafy vegetables are more sensitive to humidity and thrive in higher humidity conditions, while meats and dried goods require lower humidity levels to prevent condensation or spoilage. This whole-compartment control method struggles to meet the optimal humidity needs of all food items, often resulting in less than ideal preservation for some products.

[0004] Therefore, existing technologies urgently need a low-humidity precision control scheme that can combine the trend of food moisture content change, regional humidity differentiation adjustment and strategy adaptive correction, so as to ensure the advantages of low humidity while avoiding food from being too dry or too dry, thereby achieving more precise and long-term effective humidity control and preservation in different food types, storage locations and usage environments. Summary of the Invention

[0005] The purpose of this section is to outline some aspects of embodiments of the present invention and to briefly describe some preferred embodiments. Simplifications or omissions may be made in this section, as well as in the abstract and title of this application, to avoid obscuring the purpose of these documents; however, such simplifications or omissions should not be construed as limiting the scope of the invention.

[0006] To solve the above-mentioned technical problems, the present invention provides the following technical solution: a wind-cooled refrigerator with precise low humidity regulation function, mainly comprising: The box has an internal refrigeration compartment for preserving and storing food. The refrigeration compartment has multiple storage layers for storing different types of food. The enclosure is equipped with a humidity control system, which includes: The air duct system is used to realize air flow and local humidity regulation in the cold storage area; A dehumidification module is installed on the air circulation path of the air duct system in the cold storage area to regulate the humidity level in the cold storage area. The sensor assembly is used to collect data on the moisture content of food on each loading layer, information on the type of food, and information on environmental parameters of the cold storage area. The controller is electrically connected to the dehumidification module, the air duct system, and the sensor assembly. The controller has a built-in memory and processor. The memory pre-stores an artificial intelligence decision-making program for humidity regulation, and the processor runs this decision-making program to: Obtain the input set of food state features and the input set of humidity control auxiliary features based on sensor components; The food state feature input set and the humidity control auxiliary feature input set are input into the AI ​​humidity control decision model to output a humidity control strategy. The operation of the dehumidification module and the air duct system is controlled based on the humidity control strategy described above. During the dehumidification control process, real-time data on changes in food moisture content is collected, and combined with historical moisture content data, a drying trend coefficient is calculated. The humidity control strategy offset correction is performed based on the aforementioned drying trend coefficient, and the effectiveness of the control strategy is judged based on the current environmental and food status indicators. If it is ineffective, the humidity control strategy is regenerated.

[0007] As a preferred embodiment of the air-cooled refrigerator with precise low humidity regulation function described in this invention, the dehumidification module is disposed on the main circulation path of the air duct system. The air duct system includes a main air duct and multiple branch air ducts for connecting different load layers. Each branch air duct is provided with an independent air valve or wind speed adjustment mechanism for zoned regulation of humidity at different levels according to the humidity control strategy.

[0008] As a preferred embodiment of the air-cooled refrigerator with precise low humidity regulation function described in this invention, the controller integrates a data acquisition unit, a strategy generation unit, an execution control unit, and an effectiveness judgment unit. The data acquisition unit is used to collect and summarize the food status and environmental parameter data output by the sensor components; The strategy generation unit includes a processor and a pre-stored AI decision program, used to run the humidity control strategy generation logic; The execution control unit is used to control the dehumidification module and the air duct system according to the strategy output structure, and to perform strategy correction operations in real time. The validity judgment unit is used to judge the validity of the control strategy based on the current environmental and food status indicators. If it is invalid, the humidity control strategy will be regenerated.

[0009] The present invention also provides a control method for a frost-free refrigerator with precise low humidity regulation, applied to the aforementioned frost-free refrigerator with precise low humidity regulation, comprising the following steps: S1: Obtain food moisture content detection data and food type information on each loading layer in the cold storage area, and construct a layered food state feature input set; S2: Obtain environmental parameter information of the cold storage area and construct a humidity control auxiliary feature input set; S3: Input the food state feature input set and the humidity control auxiliary feature input set into the AI ​​humidity control decision model pre-built in the controller to generate humidity control strategy output; S4: Based on the output of the humidity control strategy, control the dehumidification module and the air duct system to perform dehumidification control operations, and continuously collect data on changes in food moisture content during execution, compare it with historical data, and calculate the drying trend coefficient; S5: Based on the drying trend coefficient, perform strategy offset correction on the AI ​​humidity control decision model; S6: After completing the strategy offset correction based on the drying trend coefficient, further make a comprehensive judgment on the effectiveness of the humidity control strategy based on the current environmental and food status indicators. If the judgment result is invalid, return to step S3 to regenerate the control strategy; otherwise, maintain the current humidity control strategy execution status.

[0010] As a preferred embodiment of the control method for the air-cooled refrigerator with precise low humidity regulation function described in this invention, the food moisture content detection data includes, but is not limited to, current moisture content, rate of change of moisture content, local moisture content distribution, and historical moisture content data; the food type information includes, but is not limited to, food type label, food morphological characteristics, humidity sensitivity level, recommended storage humidity range, and layer placement position; the environmental parameter information includes, but is not limited to, internal temperature, humidity, wind speed, air duct status, door opening and closing frequency, and external ambient temperature and humidity.

[0011] As a preferred embodiment of the control method for the air-cooled refrigerator with precise low humidity regulation function described in this invention, the construction of the AI ​​humidity control decision model includes the following steps: The constructed food state feature input set, the humidity control auxiliary feature input set, and the output label of the optimal humidity control strategy corresponding to the input set are combined to form a sample dataset for supervised learning. The sample dataset is divided according to a preset ratio to form a training set, a validation set, and a test set. Perform preprocessing operations on the sample data in the training set, validation set, and test set respectively; Select an AI model architecture suitable for predicting humidity control strategies, train the model using the training set, and adjust the model hyperparameters in real time using the validation set. The model performance was evaluated using the test set, and the model construction was completed.

[0012] As a preferred embodiment of the control method for the air-cooled refrigerator with precise low humidity regulation function described in this invention, the humidity control strategy output includes a target humidity setpoint, a humidity threshold for starting and stopping the dehumidification module, an operating frequency or power setting for the dehumidification module, a wind speed level setting, an air duct selection, and a control cycle length.

[0013] In a preferred embodiment of the control method for the air-cooled refrigerator with precise low humidity regulation function described in this invention, the formula for calculating the drying trend coefficient is as follows: ; in, This indicates the rate of change of current moisture content, in % / h. This indicates the rate of change in historical moisture content, in % / h. This represents the drying trend coefficient. To prevent small positive numbers with a denominator of zero; By calculating the drying trend coefficient, the actual impact of the current dehumidification strategy on the food dehydration rate can be reflected in real time.

[0014] As a preferred embodiment of the control method for the air-cooled refrigerator with precise low humidity regulation function described in this invention, the strategy offset correction includes: when When this occurs, it indicates that the current dehydration rate of the food is significantly higher than historical levels, posing a risk of over-drying and triggering the first correction order; when When this occurs, it indicates that the current food dehydration rate is significantly lower than historical levels, which may lead to insufficient humidity control and excessive humidity, triggering the second correction order; when When the humidity is stable, the humidity control strategy remains unchanged. in: The first correction instruction is: Reduce the operating frequency or power of the dehumidification module; Reduce the wind speed level of the corresponding cargo layer or air duct; Fine-tune the target humidity setting upwards; Increase the threshold for triggering dehumidification; The second correction instruction is: Increase the operating frequency or power of the dehumidification module; Enhance airflow circulation in the corresponding cargo layer; Lower the lower limit of the target humidity range; Lower the threshold for triggering dehumidification startup.

[0015] As a preferred embodiment of the control method for the air-cooled refrigerator with precise low humidity regulation function described in this invention, the effectiveness judgment of the humidity control strategy includes: judging whether the current actual humidity is within the target humidity tolerance range and whether the change in food moisture content is within the preset threshold range. If at least one of them is not satisfied, the humidity control strategy is deemed invalid.

[0016] The beneficial effects of this invention are: 1. This invention introduces the calculation of the drying trend coefficient R and combines it with the upper threshold. and lower threshold The dual-threshold judgment logic can trigger strategy offset corrections in different directions when the food dehydration rate is significantly higher or lower than the historical level. This effectively prevents excessive drying or insufficient humidity, and enables the humidity regulation process to dynamically adapt to the moisture-sensitive characteristics of different foods, thereby significantly improving the humidity control accuracy and preservation stability of the refrigerator's cold storage area.

[0017] 2. This invention sets up a main air duct and branch air ducts in the air duct system of the air-cooled refrigerator, and equips the branch air ducts with independent air valves or air speed adjustment mechanisms. Combined with the zoned humidity control strategy output by the AI ​​humidity control decision model, it can implement layered humidity adjustment according to the food types, humidity sensitivity levels and moisture content differences of different storage layers. Compared with the traditional whole-warehouse humidity control method, it can significantly reduce the fluctuation of food moisture content changes and extend the optimal preservation time of different types of food.

[0018] 3. This invention not only generates an initial humidity control strategy by collecting food and environmental characteristic data in real time through sensor components, but also judges the effectiveness of the strategy during operation and returns to the model decision-making stage to regenerate the strategy when the strategy is ineffective, thus realizing a closed-loop adaptive adjustment from data acquisition to strategy generation to execution to effect verification. It can ensure that the humidity control strategy can maintain continuous optimization and long-term effectiveness under different environmental conditions, food types or usage habits. Attached Figure Description

[0019] To more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings used in the description of the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort. Wherein: Figure 1 This is a schematic diagram of the internal structure of a wind-cooled refrigerator with precise low humidity regulation function according to the present invention.

[0020] Figure 2 This is a flowchart illustrating a control method for an air-cooled refrigerator with precise low humidity regulation according to the present invention.

[0021] In the diagram: 100, cabinet; 101, refrigerated compartment; 102, storage layer; 103, air duct system; 104, dehumidification module; 105, controller. Detailed Implementation

[0022] To make the above-mentioned objects, features and advantages of the present invention more apparent and understandable, the specific embodiments of the present invention will be described in detail below with reference to the accompanying drawings.

[0023] Many specific details are set forth in the following description in order to provide a full understanding of the invention. However, the invention may also be practiced in other ways different from those described herein, and those skilled in the art can make similar extensions without departing from the spirit of the invention. Therefore, the invention is not limited to the specific embodiments disclosed below.

[0024] Secondly, the term "one embodiment" or "embodiment" as used herein refers to a specific feature, structure, or characteristic that may be included in at least one implementation of the present invention. The phrase "in one embodiment" appearing in different places in this specification does not necessarily refer to the same embodiment, nor is it a single or selective embodiment that is mutually exclusive with other embodiments.

[0025] Secondly, the present invention is described in detail with reference to the schematic diagrams. When detailing the embodiments of the present invention, for ease of explanation, the cross-sectional views illustrating the device structure may be partially enlarged, not according to the usual scale. Furthermore, the schematic diagrams are merely examples and should not limit the scope of protection of the present invention. In addition, actual fabrication should include three-dimensional spatial dimensions of length, width, and depth.

[0026] Example 1, referring to Figure 1 The first embodiment of the present invention provides a wind-cooled refrigerator with precise low humidity regulation function, mainly comprising: The box 100 has a refrigeration compartment 101 inside for preserving and storing food. The refrigeration compartment 101 has multiple storage layers 102 for placing different types of food. The enclosure 100 is equipped with a humidity control system, which includes: Air duct system 103, the air duct system 103 is used to realize air flow and local humidity regulation in the cold storage area; Dehumidification module 104 is disposed on the air circulation path of the air duct system 103 of the cold storage area 101, for example, embedded at the air circulation path inlet (the first place the airflow passes through), and is used to regulate the humidity level in the cold storage area 101; Sensor assembly 105 is used to collect food moisture content data, food type information and environmental parameter information of the cold storage area on each loading layer 102; Controller 106, electrically connected to dehumidification module 104, air duct system 103, and sensor assembly 105, has a built-in memory and processor. The memory pre-stores an artificial intelligence decision-making program for humidity regulation, and the processor runs this decision-making program to: Acquire the food state feature input set and humidity control auxiliary feature input set based on sensor component 105; The food state feature input set and the humidity control auxiliary feature input set are input into the AI ​​humidity control decision model to output a humidity control strategy. The operation of the dehumidification module 104 and the air duct system 103 is controlled based on the humidity control strategy. During the dehumidification control process, real-time data on changes in food moisture content is collected, and combined with historical moisture content data, a drying trend coefficient is calculated. The humidity control strategy offset correction is performed based on the aforementioned drying trend coefficient, and the effectiveness of the control strategy is judged based on the current environmental and food status indicators. If it is ineffective, the humidity control strategy is regenerated.

[0027] Specifically, the dehumidification module 104 is installed on the main circulation path of the air duct system 103. The air duct system 103 includes a main air duct 103-1 and multiple branch air ducts 103-2 for connecting different carrying layers 102. The dehumidification module 104 is embedded at the inlet of the main air duct. Each branch air duct 103-2 is equipped with an independent air valve or wind speed adjustment mechanism for zoned control of humidity at different levels according to the humidity control strategy.

[0028] Specifically, the controller 106 integrates a data acquisition unit, a strategy generation unit, an execution control unit, and a validity judgment unit. The data acquisition unit is used to collect and summarize the food status and environmental parameter data output by the sensor component 105; The strategy generation unit includes a processor and a pre-stored AI decision program, used to run the humidity control strategy generation logic; The execution control unit is used to control the dehumidification module 104 and the air duct system 103 according to the strategy output structure, and to perform strategy correction operations in real time. The validity judgment unit is used to judge the validity of the control strategy based on the current environmental and food status indicators. If it is invalid, the humidity control strategy will be regenerated.

[0029] During operation, sensor components 105 first collect environmental parameters such as food moisture content, food type, and temperature and humidity of the cold storage zone 101 on each carrying layer 102 in real time, and transmit them to the data acquisition unit of controller 106 for aggregation and processing. The controller then calls a pre-stored AI humidity control decision model, inputting the food state feature input set and the humidity control auxiliary feature input set into the model to generate a humidity control strategy. The execution control unit of controller 106 drives the dehumidification module 104 and the air duct system 103 to operate collaboratively, achieving overall or zoned humidity adjustment of the cold storage zone 101. During humidity control operation, controller 106 continuously collects data on changes in food moisture content, compares it with historical data to obtain a drying trend coefficient, and determines whether to execute strategy offset correction to prevent excessive water loss or insufficient humidity in the food. After correction, controller 106 evaluates the strategy effect through an effectiveness judgment unit. If ineffective, it returns to the strategy generation step to re-output the humidity control strategy, achieving dynamic closed-loop adjustment. This effectively prevents excessive water loss in food, delays quality deterioration, and significantly improves food preservation.

[0030] Example 2, refer to Figure 2 This is a second embodiment of the present invention, which provides a control method for an air-cooled refrigerator with precise low humidity regulation, comprising the following steps: S1: Obtain food moisture content detection data and food type information on each carrying layer 102 in the cold storage area 101, and construct a layered food state feature input set.

[0031] It should be noted that the food moisture content detection data includes, but is not limited to: current moisture content, rate of change of moisture content, local moisture content distribution, and historical moisture content data. This data comprehensively reflects the drying state and moisture retention trend of food in different regions and over different time periods. It can be acquired through near-infrared sensors installed on the top wall of the chamber and at the bottom of the storage layer 102, corresponding to the area of ​​each storage plate. The near-infrared sensors achieve continuous monitoring of the food's moisture state through non-contact spectral reflectance analysis. The current moisture content represents the overall moisture content value of the food, reflecting its moisture retention state at the current moment.

[0032] Food type information includes, but is not limited to, food type label, food physical characteristics (such as cut / whole, packaged / uncovered), humidity sensitivity level, recommended storage humidity range, and layer placement location. This information can be obtained through user input, touch selection, or by cameras located on the top wall of the cabinet or at the bottom of the storage layer 102, corresponding to each storage layer 102 area.

[0033] By collecting the above-mentioned food moisture content detection data and food type information, an accurate set of stratified food state characteristics can be constructed and provided as input parameters to the AI ​​humidity control decision model. This will generate a humidity control strategy that matches the actual preservation needs of the ingredients in each of the 102 layers, thereby achieving differentiated and stratified humidity control.

[0034] Specifically, the steps for constructing a hierarchical food state feature input set are as follows: First, establish a one-to-one correspondence between the collected data and the specific carrying layer 102 to form hierarchical index labels, ensuring that each layer of data is independently identified. Second, standardize and denoise the original data, extract core features, and convert them into feature vector forms recognizable by the AI ​​model. Third, organize the feature vectors of each carrying layer 102 according to a hierarchical structure to form a food state feature input set containing multiple hierarchical subsets. The core features include, but are not limited to, current moisture content, rate of change of moisture content, humidity sensitivity level, and recommended humidity range. For example, the feature vector form is [65.2, -1.8, 2, (75, 85)], corresponding to the current moisture content %, rate of change % / h, humidity sensitivity level, and recommended lower and upper humidity limits, respectively.

[0035] It should be noted that by constructing an input set with layered food state characteristics, the above-mentioned method not only achieves accurate identification and data mapping of the differentiated needs of the 102 food items in each layer of the refrigeration area, but also provides refined input support for the AI ​​humidity control decision model, thereby improving the matching degree of subsequent humidity control strategies and the level of intelligent preservation.

[0036] S2: Obtain environmental parameter information for cold storage area 101 and construct a humidity control auxiliary feature input set.

[0037] It should be noted that environmental parameters include, but are not limited to, internal temperature, humidity, fan speed, air duct status, door opening and closing frequency, and external ambient temperature and humidity. By acquiring these parameters, the refrigerator's current operating conditions, sealing status, and external interference factors can be comprehensively reflected. This provides accurate boundary conditions and dynamic references for subsequent humidity control strategies, enhancing the AI ​​humidity control model's responsiveness to environmental changes and the adaptability of strategy adjustments.

[0038] Specifically, the steps for constructing the humidity control auxiliary feature input set are as follows: The acquired raw environmental parameter data is aligned with the collection timestamp and the system operation cycle, and labeled with the corresponding level or whole-enclosure environmental label. That is, if the environmental parameter has independent data for each layer, such as each layer has a separate temperature and humidity sensor, then it is necessary to label which layer the data belongs to (i.e., the corresponding level); if the environmental parameter is shared by the whole enclosure, such as the enclosure has only one air duct status, then it is necessary to label that this data is universal to the whole enclosure, i.e., whole-enclosure environmental label.

[0039] Normalization and outlier filtering are performed on various data to extract feature variables that are strongly correlated with humidity regulation, such as temperature and humidity fluctuation amplitude, door opening and closing frequency, average wind speed, and duct open / closed status. These feature variables are then combined to form a standardized feature vector (e.g., [4.5, 82.1, 1, 0, 3, 30.2, 65.0], representing wind speed (m / s), humidity (%RH), duct open / closed status, dehumidification status, door opening and closing frequency, external temperature, and external humidity, respectively), forming a humidity control auxiliary feature input set. This set serves as the external influence dimension input for the AI ​​humidity control decision model, enabling the model to have stronger environmental robustness when predicting control strategies, thereby generating humidity control strategy output.

[0040] It should be noted that the temperature, humidity and wind speed inside the chamber in this application document are environmental parameters set in layers, while the duct status, door opening and closing frequency and external ambient temperature and humidity are parameters for the entire chamber.

[0041] The S2 step described above accurately acquires and labels key environmental parameters such as internal temperature, humidity, and wind speed, as well as universal air duct status, door opening and closing frequency, and external ambient temperature and humidity. This constructs a standardized and normalized set of humidity control auxiliary features, which can comprehensively reflect the refrigerator's operating conditions and external disturbances. This greatly enhances the AI ​​humidity control decision model's ability to perceive and respond to environmental changes, thereby effectively improving the accuracy and adaptability of the humidity control strategy and ensuring the balanced preservation of multi-layered food in the refrigeration compartment.

[0042] S3: Input the food state feature input set and the humidity control auxiliary feature input set together into the pre-built AI humidity control decision model to generate humidity control strategy output; It should be noted that the humidity control strategy output includes, but is not limited to, combinations of parameters such as wind speed setting, air duct selection, dehumidification module 104 start / stop signal, and control cycle length.

[0043] The purpose of building an AI humidity control decision model is to use machine learning technology to integrate multi-dimensional food status and environmental parameter information, automatically generate humidity control strategies that match the actual preservation needs of each storage layer 102, achieve precise humidity control of the air-cooled refrigerator in changing environments, and improve the storage quality and preservation efficiency of various types of food.

[0044] Specifically, the steps for building an AI humidity control decision model are as follows: The food state feature input set and humidity control auxiliary feature input set that have been constructed in steps S1 and S2, as well as the output label of the optimal humidity control strategy corresponding to the above input sets, are used to form a sample dataset for supervised learning. The sample dataset is divided according to a preset ratio to form a training set, a validation set and a test set, which serve as the basis for training, tuning and performance evaluation of the AI ​​humidity control decision model. Perform preprocessing operations on the sample data in the training set, validation set, and test set respectively; Choose an AI model architecture (such as a deep neural network architecture) suitable for predicting humidity control strategies, train the model using the training set, and use the validation set to adjust the model hyperparameters in real time and prevent overfitting in order to improve the model's generalization ability. After training, the model's performance is evaluated using an independent test set to ensure that the model has stable accuracy and robustness, and meets the actual humidity control strategy prediction requirements, thereby completing the construction of the AI ​​humidity control decision model.

[0045] After the model is built, it can be deployed in the refrigerator control system to receive the food state feature input set and humidity control auxiliary feature input set generated in steps S1 and S2 in real time, and output humidity control strategy. At the same time, during the application of the model, the strategy is corrected by continuously collecting the drying trend coefficient, so as to realize the dynamic optimization and adaptive adjustment of the AI ​​humidity control decision model.

[0046] In one specific implementation, the preprocessing operation includes: The event sequence of food moisture content change rate data and door opening events in environmental parameters is aligned to establish an event correlation matrix. Based on the food's moisture sensitivity level, the moisture content characteristics within each group were normalized using Z-score.

[0047] It should be noted that the event correlation matrix is ​​a two-dimensional matrix structure used to describe the temporal relationship between door opening events and changes in food moisture content. Each row represents a door opening event, the columns correspond to the amount of moisture content change within a specific time window, and the matrix elements reflect the change value or trend indicator of food moisture content at a certain time interval after the door is opened. By constructing the event correlation matrix, causal features between "door opening-drying" can be extracted, enhancing the model's responsiveness to humidity fluctuations caused by human interference.

[0048] In one specific implementation, the optimal humidity control strategy output label is obtained in the following way: Based on the humidity control parameters that ensure food moisture retention rate is ≥90% in historical humidity control records; Based on expert experience; Weighted calibration was performed based on the optimal preservation humidity range for this type of food in a constant temperature test chamber.

[0049] It should be noted that by using historical humidity control parameters based on a food moisture retention rate of ≥90%, or by using expert experience and constant temperature experiment calibration results as the "optimal humidity control strategy output" label, it can be ensured that the humidity control strategy learned by the model directly corresponds to the conditions for optimal food preservation, thereby effectively improving the training effect and prediction accuracy of the AI ​​humidity control decision model.

[0050] Specifically, the humidity control strategy output includes the target humidity setpoint, the humidity threshold for starting and stopping the dehumidification module 104, the operating frequency or power setting of the dehumidification module 104, the wind speed level setting, the air duct selection, and the control cycle length. Among these: The target humidity setpoint is used to define the "desired humidity value" output by the AI ​​humidity control decision model, serving as a reference for closed-loop control. The operating frequency or power setting of the dehumidification module 104 can instruct the controller 106 to adjust the operating intensity of the dehumidification module 104.

[0051] Wind speed setting, that is, setting the wind speed level of each carrying layer 102 or the air outlet of the air duct, is used to adjust the air flow intensity, thereby affecting the rate of moisture evaporation or accumulation.

[0052] Duct selection refers to controlling the on / off state of air ducts, such as selecting to open or close air ducts on one or more levels to achieve zoned air supply and differentiated humidity control.

[0053] The humidity threshold for starting and stopping the dehumidification module 104 is used to define the ambient humidity trigger value when to start or stop the dehumidification behavior.

[0054] The control cycle length defines the execution cycle of the humidity control strategy, i.e. how long to maintain the current strategy or how long to re-evaluate and adjust it, affecting the frequency and stability of the adjustment response.

[0055] It should be noted that the above control strategies can be combined to drive the coordinated work of various hardware modules inside the refrigerator, ultimately achieving differentiated and refined humidity control to better match the preservation needs of different types and levels of food.

[0056] In summary, step S3 inputs the food state feature input set and the humidity control auxiliary feature input set into the pre-built AI humidity control decision model, and performs supervised learning training based on the optimal strategy output label with the food moisture content retention rate ≥90%. This enables the model to integrate multi-dimensional food attributes and environmental state information, achieve accurate identification and matching of humidity control requirements for each loading layer 102, and intelligently generate the optimal humidity control strategy under changing environmental conditions, significantly improving the humidity control accuracy and preservation effect of the air-cooled refrigerator.

[0057] S4: Execute dehumidification control operation according to the humidity control strategy output, and continuously collect food moisture content change data during the execution of the dehumidification control operation, compare it with historical data, and calculate the drying trend coefficient.

[0058] Specifically, the formula for calculating the drying trend coefficient is as follows: ; in, This indicates the rate of change of current moisture content, in % / h. This indicates the rate of change in historical moisture content, in % / h. This represents the drying trend coefficient. To prevent small positive numbers with a denominator of zero. For example, 0.01.

[0059] It should be noted that step S4 continuously collects data on changes in food moisture content during the dehumidification control operation and compares this data with historical data to calculate the drying trend coefficient. This not only reflects the actual impact of the current dehumidification strategy on the food dehydration rate in real time, but also dynamically monitors the trend of food moisture loss. This provides quantifiable feedback for subsequent corrections to the humidity control strategy, avoiding food drying due to excessive humidity control or abnormal humidity in the storage environment due to insufficient humidity control, thus meeting the preservation needs of various types of food.

[0060] S5: Based on the drying trend coefficient, perform strategy offset correction on the AI ​​humidity control decision model.

[0061] Specifically, policy offset correction includes: when When this occurs, it indicates that the current dehydration rate of the food is significantly higher than historical levels, posing a risk of over-drying and triggering the first correction order; when When this occurs, it indicates that the current food dehydration rate is significantly lower than historical levels, which may lead to insufficient humidity control and excessive humidity, triggering the second correction order; when When the humidity level is high, it indicates that the drying trend is stable and the humidity control strategy remains unchanged.

[0062] It should be noted that, The recommended value range is [0.05, 0.20], which indicates that the current dehydration rate is 5%-20% higher than the historical reference value. The recommended value range is [-0.05, -0.15], which means that the current dehydration rate is 5%-15% lower than the reference value. and The basis for this setting includes: Historical humidity control data statistical analysis, that is, by analyzing a large number of moisture content change samples, calculate the stable range of R value when the moisture content retention rate of food is ≥90%, and use this as the basis for setting the benchmark deviation range; The experimental results were verified by controlling different humidity levels in a constant temperature experimental chamber, recording the performance of food indicators such as water loss rate and air drying under different R value ranges, and selecting the fluctuation limit with the best preservation effect. Food moisture sensitivity classification, that is, foods with high moisture sensitivity should be classified into smaller categories. and For foods with low moisture sensitivity, the permissible deviation range can be appropriately expanded.

[0063] in: The first correction instruction is: Reduce the operating frequency or power of the dehumidification module 104; for example, the controller 106 instructs the dehumidification module 104 to reduce its operating frequency by 10%-30%, or to reduce its power by 50W-200W from the original setting, in order to slow down the dehumidification rate and control moisture loss.

[0064] Reduce the wind speed level of the corresponding cargo layer 102 or the air duct to slow down the trend of accelerated moisture evaporation by airflow; for example, adjust the wind speed control parameter to the next level, or directly reduce the wind speed by 15%-25%.

[0065] Slightly adjust the target humidity setting upwards to cause a slight increase in local humidity; for example, increase the current humidity control target value by 3%RH-5%RH to keep the local humidity in a higher range, alleviate the trend of excessive dryness in the local environment, and enhance the air's moisturizing ability.

[0066] Increase the threshold for triggering dehumidification. For example, raise the starting humidity threshold of the original dehumidification module 104 by 2%RH-4%RH, and only start it when the humidity is critically high, so as to avoid excessive drying caused by frequent start-ups.

[0067] The second amendment instruction is: Increase the operating frequency or power of the dehumidification module 104 to quickly reduce ambient humidity; for example, increase the operating frequency of the dehumidification module 104 by 15%-35%, or increase the dehumidification power by 50W-200W to improve moisture removal efficiency and accelerate the reduction of humidity.

[0068] Enhance the airflow circulation of the corresponding carrier layer 102 to facilitate faster removal of moisture; for example, increase the output wind speed of the air duct by one level or by 20%-30% to strengthen the airflow inside the refrigeration cavity and promote the rapid transfer of moisture to the dehumidification area.

[0069] Lowering the lower limit of the target humidity range enhances the overall dehumidification intent; for example, the control strategy lowers the lower limit of the humidity range by 3%RH-6%RH (RH is relative humidity, for example, 3%RH means that the relative humidity of the air in the corresponding layer of the refrigeration area is 3%), thereby expanding the dehumidification working range and enhancing the system's dehumidification intent.

[0070] Lowering the threshold for triggering dehumidification allows the system to respond to humidity increases earlier. For example, lowering the dehumidification start threshold by 2%RH-4%RH allows the dehumidification module 104 to intervene in humidity control at an earlier stage, enabling a faster response to humidity fluctuations.

[0071] It should be noted that step S5 uses dynamic analysis based on the drying trend coefficient R value, combined with empirically set upper and lower thresholds. and This triggers targeted first or second correction instructions, enabling the AI ​​humidity control decision model to perceive the food dehydration status and environmental humidity control deviation in real time during strategy execution. It then adjusts the dehumidification intensity, wind speed level, and control threshold in a timely manner, thereby improving the sensitivity and accuracy of humidity control response. This allows for continuous optimization and adaptive control of humidity regulation strategies under different food storage conditions, effectively preventing the risks of excessive dryness or insufficient humidity and enhancing the preservation capabilities of air-cooled refrigerators.

[0072] S6: After completing the strategy offset correction based on the drying trend coefficient, further make a comprehensive judgment on the effectiveness of the humidity control strategy based on the current environmental and food status indicators. If the judgment result is invalid, return to step S3 to regenerate the control strategy; otherwise, maintain the current execution status of the humidity control strategy.

[0073] In one specific implementation, the effectiveness judgment of the humidity control strategy includes: determining whether the current actual humidity is within the target humidity tolerance range and whether the change in food moisture content is within the preset threshold range. If at least one of them is not satisfied, the humidity control strategy is deemed invalid.

[0074] It should be noted that the target humidity tolerance range refers to the allowable fluctuation range of the target humidity value set by the system, that is, the tolerable deviation range between the actual ambient humidity value and the target humidity value. The food moisture content change threshold refers to the allowable range of change in food moisture content per unit time (e.g., per hour), used to determine whether the current humidity control strategy can effectively inhibit or control the food dehydration rate.

[0075] Preferably, the target humidity tolerance range is the set value ±3%RH. For example, if the target humidity is set to 75%RH, then allowing the actual humidity to fluctuate between 72%RH and 78%RH is still considered a valid strategy.

[0076] It should be noted that setting the above target humidity tolerance range helps to avoid increased energy consumption and strategy fluctuations caused by frequent adjustments. While ensuring the preservation effect, it provides a reasonable adjustment buffer, thereby improving the stability and adaptability of the humidity control strategy.

[0077] Preferably, the threshold range for changes in food moisture content is [-0.3% / h, +0.2% / h], which can be understood as: If the system detects that the rate of decrease in food moisture content exceeds 0.3% / h (i.e., below -0.3% / h) within a certain period, it indicates that the moisture loss is too rapid. If the rate of increase in food moisture content exceeds 0.2% / h (i.e., above +0.2% / h) within a certain period, it indicates that the risk of rehydration due to excessive humidity has increased. Only when the change value is within this range can the current humidity control strategy be judged to be effective in maintaining food moisture.

[0078] For example, if the moisture content of a certain vegetable food drops from 94.2% to 93.1% in the past 3 hours, with an average rate of change of -0.37%, exceeding the lower threshold, the system will determine that the humidity control strategy is invalid and trigger an adjustment.

[0079] It should be noted that the above-mentioned threshold for food moisture content change enables the system to perceive the change in food moisture status in real time during the execution of the humidity control strategy, quickly identify the shortcomings of the strategy, optimize the strategy, effectively ensure food freshness and avoid irreversible cracking, discoloration or deterioration in taste.

[0080] Therefore, step S6 achieves dynamic evaluation and decision feedback on the effectiveness of the humidity control strategy by setting the target humidity tolerance range and the food moisture content change threshold. The former is used to determine whether the environmental control precision meets the target setting, while the latter is used to evaluate the actual effect of the humidity control strategy on the food moisture retention. Together, they constitute the basis for judging the effectiveness of the strategy and can significantly improve the food storage and preservation capabilities.

[0081] In summary, this invention introduces the calculation of the drying trend coefficient R and combines it with the upper threshold. and lower threshold The dual-threshold judgment logic can trigger strategy offset corrections in different directions when the food dehydration rate is significantly higher or lower than historical levels, effectively preventing excessive drying or insufficient humidity. This allows the humidity regulation process to dynamically adapt to the moisture-sensitive characteristics of different foods, thereby significantly improving the humidity control accuracy and preservation stability of the refrigerator's cold storage compartment. This invention sets up a main air duct and branch air ducts in the air duct system of the air-cooled refrigerator, and equips the branch air ducts with independent air valves or wind speed adjustment mechanisms. Combined with the zoned humidity control strategy output by the AI ​​humidity control decision model, it can implement layered humidity regulation based on the food type, moisture sensitivity level, and moisture content differences in different storage layers. Compared with traditional whole-warehouse humidity control methods, it can significantly reduce the fluctuation of food moisture content changes and extend the optimal preservation time of different types of food. This invention not only generates an initial humidity control strategy by collecting real-time data on food and environmental characteristics using sensor components, but also judges the effectiveness of the strategy during operation and returns to the model decision-making stage to regenerate the strategy when the strategy is ineffective. This achieves a fully closed-loop adaptive adjustment from data acquisition to strategy generation, execution, and effect verification. It can ensure that the humidity control strategy can maintain continuous optimization and long-term effectiveness under different environmental conditions, food types, or usage habits.

[0082] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention, and all such modifications or substitutions should be covered within the scope of the claims of the present invention.

Claims

1. A frost-free refrigerator with precise low humidity control, characterized in that, include: The box (100) has a refrigeration area (101) inside for storing food fresh. The refrigeration area (101) has multiple storage layers (102) for placing different kinds of food. The housing (100) is equipped with a humidity control system, which includes: Air duct system (103), the air duct system (103) is used to realize air flow and local humidity regulation in the cold storage area; A dehumidification module (104) is provided on the air circulation path of the air duct system (103) in the cold storage area (101) to regulate the humidity level in the cold storage area (101); Sensor assembly (105) is used to collect food moisture content data, food type information and environmental parameter information of the cold storage area on each loading layer (102); A controller (106) is electrically connected to a dehumidification module (104), an air duct system (103), and a sensor assembly (105). The controller (106) has a built-in memory and a processor. The memory stores an artificial intelligence decision-making program for humidity regulation, and the processor runs the decision-making program to: Obtain the food state feature input set and humidity control auxiliary feature input set based on sensor component (105); The food state feature input set and the humidity control auxiliary feature input set are input into the AI ​​humidity control decision model to output a humidity control strategy. The operation of the dehumidification module (104) and the air duct system (103) is controlled based on the humidity control strategy. During the dehumidification control process, real-time data on changes in food moisture content is collected, and combined with historical moisture content data, a drying trend coefficient is calculated. The humidity control strategy offset correction is performed based on the aforementioned drying trend coefficient, and the effectiveness of the control strategy is judged based on the current environmental and food status indicators. If it is ineffective, the humidity control strategy is regenerated.

2. The air-cooled refrigerator with precise low humidity regulation function as described in claim 1, characterized in that: The dehumidification module (104) is located on the main circulation path of the air duct system (103). The air duct system (103) includes a main air duct (103-1) and multiple branch air ducts (103-2) for connecting different load layers (102). Each branch air duct (103-2) is equipped with an independent air valve or wind speed adjustment mechanism for zoned control of humidity at different levels according to the humidity control strategy.

3. The air-cooled refrigerator with precise low humidity regulation function as described in claim 1, characterized in that: The controller (106) integrates a data acquisition unit, a strategy generation unit, an execution control unit, and a validity judgment unit. The data acquisition unit is used to collect and summarize the food status and environmental parameter data output by the sensor assembly (105); The strategy generation unit includes a processor and a pre-stored AI decision program, used to run the humidity control strategy generation logic; The execution control unit is used to control the dehumidification module (104) and the air duct system (103) according to the strategy output structure, and to perform strategy correction operations in real time; The validity judgment unit is used to judge the validity of the control strategy based on the current environmental and food status indicators. If it is invalid, the humidity control strategy will be regenerated.

4. A control method for a frost-free refrigerator with precise low humidity control, applied to the frost-free refrigerator with precise low humidity control as described in any one of claims 1-3, characterized in that, Includes the following steps: S1: Obtain the food moisture content detection data and food type information on each loading layer (102) in the cold storage area (101) and construct a layered food state feature input set; S2: Obtain environmental parameter information of the cold storage area (101) and construct a humidity control auxiliary feature input set; S3: Input the food state feature input set and the humidity control auxiliary feature input set into the AI ​​humidity control decision model pre-built in the controller (106) to generate humidity control strategy output; S4: According to the output of the humidity control strategy, control the dehumidification module (104) and the air duct system (103) to perform dehumidification control operations, and continuously collect food moisture content change data during the execution, compare it with historical data, and calculate the drying trend coefficient; S5: Based on the drying trend coefficient, perform strategy offset correction on the AI ​​humidity control decision model; S6: After completing the strategy offset correction based on the drying trend coefficient, further make a comprehensive judgment on the effectiveness of the humidity control strategy based on the current environmental and food status indicators. If the judgment result is invalid, return to step S3 to regenerate the control strategy. Otherwise, maintain the current humidity control strategy.

5. The control method for a frost-cooled refrigerator with precise low humidity regulation function as described in claim 4, characterized in that: The food moisture content detection data includes, but is not limited to, current moisture content, rate of change of moisture content, local moisture content distribution, and historical moisture content data; the food type information includes, but is not limited to, food type label, food morphological characteristics, humidity sensitivity level, recommended storage humidity range, and layer placement location; the environmental parameter information includes, but is not limited to, internal temperature, humidity, wind speed, air duct status, door opening and closing frequency, and external ambient temperature and humidity.

6. The control method for a frost-cooled refrigerator with precise low humidity regulation function as described in claim 4, characterized in that: The steps for constructing the AI ​​humidity control decision model are as follows: The constructed food state feature input set, the humidity control auxiliary feature input set, and the output label of the optimal humidity control strategy corresponding to the input set are combined to form a sample dataset for supervised learning. The sample dataset is divided according to a preset ratio to form a training set, a validation set, and a test set. Perform preprocessing operations on the sample data in the training set, validation set, and test set respectively; Select an AI model architecture suitable for predicting humidity control strategies, train the model using the training set, and adjust the model hyperparameters in real time using the validation set. The model performance was evaluated using the test set, and the model construction was completed.

7. The control method for a frost-cooled refrigerator with precise low humidity regulation function as described in claim 4, characterized in that: The humidity control strategy output includes the target humidity setting value, the humidity threshold for starting and stopping the dehumidification module (104), the operating frequency or power setting of the dehumidification module (104), the wind speed level setting, the duct selection, and the control cycle length.

8. The control method for a frost-cooled refrigerator with precise low humidity regulation function as described in claim 4, characterized in that: The formula for calculating the drying trend coefficient is as follows: ； in, This indicates the rate of change of current moisture content, in % / h. This indicates the rate of change in historical moisture content, in % / h. This represents the drying trend coefficient. To prevent small positive numbers with a denominator of zero; By calculating the drying trend coefficient, the actual impact of the current dehumidification strategy on the food dehydration rate can be reflected in real time.

9. The control method for a frost-cooled refrigerator with precise low humidity regulation function as described in claim 4, characterized in that: The strategy offset correction includes: when When this occurs, it indicates that the current dehydration rate of the food is significantly higher than historical levels, posing a risk of over-drying and triggering the first correction order; when When this occurs, it indicates that the current food dehydration rate is significantly lower than historical levels, which may lead to insufficient humidity control and excessive humidity, triggering the second correction order; when When the humidity is stable, the humidity control strategy remains unchanged. in: The first correction instruction is: Reduce the operating frequency or power of the dehumidification module (104); Reduce the wind speed level of the corresponding cargo layer (102) or duct; Fine-tune the target humidity setting upwards; Increase the threshold for triggering dehumidification; The second correction instruction is: Increase the operating frequency or power of the dehumidification module (104); Enhance airflow circulation in the corresponding cargo layer (102); Lower the lower limit of the target humidity range; Lower the threshold for triggering dehumidification startup.

10. The control method for a frost-cooled refrigerator with precise low humidity regulation function as described in claim 4, characterized in that: The effectiveness judgment of the humidity control strategy includes: judging whether the current actual humidity is within the target humidity tolerance range, and whether the change in food moisture content is within the preset threshold range. If at least one of them is not met, the humidity control strategy is deemed invalid.

Images