A method and system for intermittent refrigeration of a refrigerated wine cabinet based on wine bottle storage and retrieval behavior recognition

Abstract

This invention discloses an intermittent cooling method and system for refrigerated wine cabinets based on wine bottle storage and retrieval behavior recognition. The method includes: acquiring several adjacent angular velocities of the refrigerated wine cabinet; marking the storage and retrieval behavior cycle on the time axis; calculating the instantaneous heat intrusion rate; calculating the cooling capacity compensation correction amount based on the instantaneous heat intrusion rate and the cabinet door opening angle; acquiring the minimum necessary cooling capacity of the wine cabinet under the current wine bottle storage and retrieval behavior; splicing a four-tuple storage and retrieval behavior sample; inputting the four-tuple storage and retrieval behavior sample into a pre-trained cooling capacity loss prediction model; outputting the predicted cooling capacity loss; comparing the minimum necessary cooling capacity and the cooling capacity loss; calculating the cooling deviation; and triggering an intermittent cooling control command if the cooling deviation exceeds a preset safety tolerance. This invention enables the actual cooling capacity of the wine cabinet to match the current heat load, achieving constant temperature control with micro-disturbance micro-compensation and large-disturbance large-compensation, avoiding temperature fluctuations caused by inaccurate cooling capacity of the wine cabinet.

Description

Technical Field

[0001] This invention relates to the field of intelligent cold chain storage equipment control, specifically to an intermittent refrigeration method and system for a refrigerated wine cabinet based on wine bottle storage and retrieval behavior recognition. Background Technology

[0002] The core function of a refrigerated wine cooler is to provide a constant low-temperature storage environment for wine, as temperature fluctuations can severely affect the quality of the wine. Currently, the cooling control of mainstream wine coolers mainly relies on feedback from internal temperature sensors: when the detected temperature exceeds a set threshold, the compressor starts; when it falls below the threshold, it stops (i.e., start-stop control), or the compressor speed is adjusted through a PID algorithm. Some high-end models have added door magnetic sensors to detect the opening and closing status of the cabinet door and force it into a rapid cooling mode when the door is opened.

[0003] However, existing technologies suffer from a fundamental technical flaw when dealing with the user's behavior of opening and retrieving items: the heat load compensation strategy is severely disconnected from the actual physical heat intrusion process, lacking the ability to quantitatively perceive the intensity of the retrieval behavior. Specifically, it treats cabinet door opening as a binary signal (0 or 1), completely ignoring the nonlinear influence of three key physical variables—door opening angle, opening duration, and the intensity of the behavior—on the heat exchange rate.

[0004] Furthermore, the difference between a user quickly retrieving wine by opening the cabinet door to only 5 degrees and selecting wine over a prolonged period by opening it to 60 degrees results in a difference of several times, or even tens of times, in the rate of cold air loss and the rate of hot air intrusion. Because of the lack of real-time monitoring of angular velocity and opening angle, these two drastically different behaviors cannot be distinguished, often leading to the application of the same fixed, strong cooling compensation logic. This results in overcooling under minor disturbances, causing significant temperature fluctuations and frequent compressor start-stop cycles, while under large opening shocks, insufficient compensation may lead to a delayed temperature recovery. Summary of the Invention

[0005] To address the shortcomings of existing technologies, this invention provides an intermittent cooling method and system for refrigerated wine cabinets based on wine bottle storage and retrieval behavior recognition, which solves the technical problems in the background art by introducing the quantification of the cabinet door opening angle.

[0006] To achieve the above objectives, the present invention provides the following technical solution:

[0007] In a first aspect, the present invention discloses an intermittent refrigeration method for a refrigerated wine cabinet based on wine bottle storage and retrieval behavior recognition, the method comprising:

[0008] S1. Obtain several adjacent angular velocities at the door hinge of the refrigerated wine cabinet;

[0009] S2. Based on the aforementioned adjacent angular velocities, mark the access behavior cycle on the time axis;

[0010] S3. During the access cycle, obtain the cabinet door opening angle and the ambient temperature gradient to calculate the instantaneous heat intrusion rate.

[0011] S4. Calculate the cooling compensation correction amount based on the instantaneous heat intrusion rate and the cabinet door opening angle;

[0012] S5. Obtain the basic cooling consumption benchmark of the wine cabinet in a static sealed state, and calculate the minimum necessary cooling capacity under the current wine bottle storage and retrieval behavior by combining the cooling capacity compensation correction amount and the storage and retrieval behavior cycle.

[0013] S6. Based on the basic cooling consumption benchmark, instantaneous heat intrusion rate, cabinet door opening angle and storage and retrieval behavior cycle, construct a quadruple storage and retrieval behavior sample.

[0014] S7. Input the quadruple access behavior samples into the pre-trained cold loss prediction model and output the predicted cold loss.

[0015] S8. Compare the minimum required cooling capacity and cooling capacity loss to calculate the refrigeration deviation;

[0016] S9. If the cooling deviation exceeds the preset safety tolerance, an intermittent cooling control command will be triggered.

[0017] In some specific embodiments, the access behavior cycle is marked on the time axis based on the plurality of adjacent angular velocities, including:

[0018] S2-1. Establish a behavior capture window that slides over time, with a fixed length and the end of the window always aligned with the current sampling time;

[0019] S2-2. Within the behavior capture window, collect several adjacent angular velocities at the cabinet door hinge with a set sampling step size;

[0020] S2-3. Perform a difference operation on several adjacent angular velocities to obtain the angular acceleration change sequence;

[0021] S2-4. When the angular acceleration change sequence abruptly changes from the zero value range to a positive peak value, mark this moment as the starting point for wine collection.

[0022] S2-5. After the starting point of wine collection, continuously monitor the angular acceleration change sequence until the signal returns from the negative peak to the zero range and remains stable. Mark this moment as the end point of wine collection.

[0023] S2-6. Define the time interval between the start point and the end point of wine retrieval as the storage and retrieval cycle.

[0024] In some specific embodiments, the calculation of the instantaneous heat intrusion rate includes:

[0025] S3-1, Predefined area exposure weight and door seal airtightness attenuation coefficient;

[0026] S3-2. Calculate the effective convection cross-sectional area based on the sinusoidal projection of the cabinet door opening angle and the regional exposure weight.

[0027] S3-3. Introduce the door seal airtightness attenuation coefficient, correct the failure rate of the sealing strip caused by the large angle opening of the cabinet door, and calculate the heat exchange efficiency.

[0028] S3-4. Multiply the effective convection cross-sectional area and heat exchange efficiency by the ambient temperature gradient to calculate the instantaneous heat intrusion rate.

[0029] In some specific embodiments, the cooling capacity compensation correction is calculated based on the instantaneous heat intrusion rate and the cabinet door opening angle, including:

[0030] S4-1, Preset critical opening angle and heat loss surge slope;

[0031] S4-2. Construct a heat loss function based on the cabinet door opening angle. When the opening angle is less than the critical opening angle, the heat loss increases linearly. When the opening angle exceeds the critical opening angle, the heat loss increases exponentially with the heat loss surge slope.

[0032] S4-3. The heat loss function and the instantaneous heat intrusion rate are weighted and fused together, and the cold energy compensation correction is calculated by exponential mapping.

[0033] In some specific embodiments, calculating the minimum required cooling capacity for the current bottle storage and retrieval behavior includes:

[0034] S5-1. Multiply the baseline cooling consumption by the duration of the access behavior cycle to obtain the static background cooling consumption.

[0035] S5-2. Multiply the instantaneous heat intrusion rate with the cold compensation correction amount to obtain the corrected heat intrusion rate;

[0036] S5-3. Multiply the corrected heat intrusion rate with the access behavior cycle to generate dynamic impact cooling loss.

[0037] S5-4. Calculate the sum of static background cooling loss and dynamic impact cooling loss to generate the minimum necessary cooling capacity under the current wine bottle storage and retrieval behavior.

[0038] In some specific embodiments, the pre-training step of the cooling loss prediction model includes:

[0039] S6-1. Extract M basic cooling consumption benchmarks, instantaneous heat intrusion rate, cabinet door opening angle, and storage and retrieval behavior cycle based on historical database;

[0040] S6-2. Perform feature processing on M basic cooling loss benchmarks, instantaneous heat intrusion rate, cabinet door opening angle and storage and retrieval cycle to generate corresponding basic cooling loss features, heat intrusion features, cabinet door opening features and storage and retrieval cycle features.

[0041] S6-3. Perform feature splicing on the basic cold loss characteristics, heat intrusion characteristics, cabinet door opening characteristics and storage and retrieval cycle characteristics to generate quadruple storage and retrieval behavior samples until M quadruple storage and retrieval behavior samples are obtained.

[0042] S6-4. Obtain the net cooling loss value corresponding to each quadruple access behavior sample and define it as the target label of the quadruple access behavior sample.

[0043] S6-5. Input the quadruple access behavior samples and target labels into the decision tree network for supervised training, learn the complex correspondence between wine bottle access behavior and actual cold loss, and generate a cold loss prediction model.

[0044] In some specific embodiments, the generation of corresponding basic cooling loss characteristics, heat intrusion characteristics, cabinet door opening characteristics, and storage / retrieval cycle characteristics includes:

[0045] S6-2-1. Select target type parameters based on the basic cooling consumption benchmark, instantaneous heat intrusion rate, cabinet door opening angle, and storage and retrieval cycle.

[0046] S6-2-2, Obtain the maximum and minimum parameters in the target type parameters;

[0047] S6-2-3. Calculate the parameter limit difference between the maximum and minimum parameters;

[0048] S6-2-4. In the target type parameters, anchor the current feature parameters one by one;

[0049] S6-2-5. Calculate the current parameter difference between the current feature parameter and the minimum parameter value;

[0050] S6-2-6. Perform a ratio operation between the current parameter difference and the parameter limit difference to obtain the target parameter features corresponding to the current feature parameter, until the M target parameter features corresponding to the target type parameter are obtained.

[0051] S6-2-7. Traverse the basic cooling loss baseline, instantaneous heat intrusion rate, cabinet door opening angle and access behavior cycle, and repeatedly calculate the target parameter characteristics until the corresponding basic cooling loss characteristics, heat intrusion characteristics, cabinet door opening characteristics and access cycle characteristics are generated.

[0052] In some specific embodiments, the execution of intermittent cooling control commands includes:

[0053] If the cooling deviation shows that the estimated cooling capacity loss is much greater than the minimum required cooling capacity, it is determined to be a high heat load scenario. The compressor is controlled to run continuously at a high duty cycle, and the opening of the electronic expansion valve is increased.

[0054] If the cooling deviation shows that the estimated cooling capacity loss is slightly greater than the minimum required cooling capacity, it is determined to be a normal access scenario, and pulse cooling is executed: the compressor runs for a preset time and then stops, and the residual cooling in the air duct is used for fine adjustment;

[0055] If the cooling deviation is within the safety tolerance, it is determined to be a minor disturbance scenario. Only the circulating fan is started to equalize the internal temperature, and the compressor is not started.

[0056] This invention provides an intermittent method for refrigerated wine cabinets based on wine bottle storage and retrieval behavior recognition, which has the following beneficial effects.

[0057] This invention accurately identifies the storage and retrieval cycle by collecting the angular velocity of the cabinet door hinge, and on this basis, calculates in parallel the theoretical minimum necessary cooling capacity based on temperature difference and angle, and the cooling capacity loss predicted based on the four-tuple sample model; further, by comparing the deviation between the minimum necessary cooling capacity and the cooling capacity loss, it directly triggers intermittent cooling control commands, dynamically adjusts the compressor duty cycle and the opening of the electronic expansion valve, so that the actual cooling capacity of the wine cabinet can match the current heat load, realizing constant temperature control with micro-disturbance micro-compensation and large disturbance large compensation, avoiding temperature fluctuations caused by inaccurate cooling capacity of the wine cabinet.

[0058] Secondly, this invention discloses an intermittent refrigeration system for a refrigerated wine cabinet based on wine bottle storage and retrieval behavior recognition, characterized in that it includes:

[0059] An angular velocity acquisition module is used to acquire several adjacent angular velocities of the refrigerated wine cabinet at the door hinge.

[0060] The behavior cycle marking module is used to mark the access behavior cycle on the time axis based on the several adjacent angular velocities;

[0061] The instantaneous heat calculation module is used to obtain the cabinet door opening angle and ambient temperature gradient during the access behavior cycle, and to calculate the instantaneous heat intrusion rate.

[0062] The cold energy compensation correction module is used to calculate the cold energy compensation correction amount based on the instantaneous heat intrusion rate and the cabinet door opening angle.

[0063] The minimum cooling capacity calculation module is used to obtain the basic cooling consumption benchmark of the wine cabinet in a static sealed state, and calculate the minimum necessary cooling capacity under the current wine bottle storage and retrieval behavior by combining the cooling capacity compensation correction amount and the storage and retrieval behavior cycle.

[0064] The behavior sample construction module is used to construct four-tuple access behavior samples based on the basic cooling consumption benchmark, instantaneous heat intrusion rate, cabinet door opening angle, and access behavior cycle.

[0065] The cooling loss output module is used to input the quadruple access behavior samples into the pre-trained cooling loss prediction model and output the predicted cooling loss.

[0066] The refrigeration deviation calculation module is used to compare the minimum required cooling capacity and the cooling capacity loss to calculate the refrigeration deviation.

[0067] The intermittent cooling control module is used to determine if the cooling deviation exceeds the preset safety tolerance, and then trigger the intermittent cooling control command.

[0068] Compared with the prior art, the beneficial effects of the intermittent refrigeration system for a refrigerated wine cabinet based on wine bottle access behavior recognition of the present invention are the same as the beneficial effects of the intermittent refrigeration method for a refrigerated wine cabinet based on wine bottle access behavior recognition described above, so they will not be repeated here. Attached Figure Description

[0069] Figure 1 This is a schematic diagram of the intermittent refrigeration method for a refrigerated wine cabinet based on wine bottle storage and retrieval behavior recognition according to the present invention.

[0070] Figure 2 This is a schematic diagram of the marking process for the access behavior cycle described in this invention;

[0071] Figure 3 This is a schematic diagram of the process for generating the minimum necessary cooling capacity described in this invention.

[0072] Figure 4 This is a schematic diagram of the process for generating the target parameter features described in this invention;

[0073] Figure 5 This is a structural block diagram of an intermittent refrigeration system for a refrigerated wine cabinet based on wine bottle storage and retrieval behavior recognition according to the present invention. Detailed Implementation

[0074] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0075] Please see Figures 1 to 4 This invention provides an intermittent refrigeration method for a refrigerated wine cabinet based on wine bottle storage and retrieval behavior recognition, the method comprising:

[0076] S1. Obtain several adjacent angular velocities at the door hinge of the refrigerated wine cabinet;

[0077] S2. Based on the aforementioned adjacent angular velocities, mark the access behavior cycle on the time axis;

[0078] S3. During the access cycle, obtain the cabinet door opening angle and the ambient temperature gradient to calculate the instantaneous heat intrusion rate.

[0079] S4. Calculate the cooling compensation correction amount based on the instantaneous heat intrusion rate and the cabinet door opening angle;

[0080] S5. Obtain the basic cooling consumption benchmark of the wine cabinet in a static sealed state, and calculate the minimum necessary cooling capacity under the current wine bottle storage and retrieval behavior by combining the cooling capacity compensation correction amount and the storage and retrieval behavior cycle.

[0081] S6. Based on the basic cooling consumption benchmark, instantaneous heat intrusion rate, cabinet door opening angle and storage and retrieval behavior cycle, construct a quadruple storage and retrieval behavior sample.

[0082] S7. Input the quadruple access behavior samples into the pre-trained cold loss prediction model and output the predicted cold loss.

[0083] S8. Compare the minimum required cooling capacity and cooling capacity loss to calculate the refrigeration deviation;

[0084] S9. If the cooling deviation exceeds the preset safety tolerance, an intermittent cooling control command will be triggered.

[0085] The intermittent cooling control command dynamically adjusts the compressor's duty cycle and the opening of the electronic expansion valve to ensure that the actual cooling capacity of the wine cabinet precisely matches the minimum necessary cooling capacity.

[0086] In this embodiment, the angular velocity of the cabinet door hinge is collected at high frequency to identify the start and end times of the user opening the door, thereby defining a complete storage and retrieval behavior cycle and distinguishing between different behaviors such as "quick wine retrieval" and "long-term selection".

[0087] Next, within this cycle, the system combines the cabinet door opening angle and the internal and external temperature difference to calculate the instantaneous heat intrusion rate and the required cooling capacity compensation correction in real time. This is then superimposed on the baseline cooling capacity under static conditions of the wine cabinet to obtain the theoretically minimum required cooling capacity under the current operating conditions. Simultaneously, the above behavioral characteristics are constructed into a four-tuple sample to output the predicted cooling capacity loss.

[0088] Finally, the theoretical threshold and the actual cooling deviation are compared. If the deviation exceeds the safe range, an intermittent cooling control command is triggered. By dynamically adjusting the compressor duty cycle and the opening of the electronic expansion valve, the actual cooling capacity of the wine cabinet is precisely matched with the current heat load, achieving constant temperature control with "micro-disturbance micro-compensation and large disturbance large compensation", effectively avoiding large temperature fluctuations and energy waste.

[0089] In this embodiment, step S2 specifically includes:

[0090] S2-1. Establish a behavior capture window that slides over time, with a fixed length and the end of the window always aligned with the current sampling time;

[0091] S2-2. Within the behavior capture window, collect several adjacent angular velocities at the cabinet door hinge with a set sampling step size;

[0092] Specifically, the aforementioned adjacent angular velocities refer to a set of values ​​continuously read by the gyroscope sensor within an extremely short time interval. Each value directly corresponds to the instantaneous speed at which the wine cabinet door rotates around the hinge. The changing trend of this set of values ​​intuitively reflects whether the user is forcefully pushing open the door, releasing it to let it slide down due to inertia, or the vibration frequency generated when the cabinet door hits the door frame at the moment of closing.

[0093] S2-3. Perform a difference operation on several adjacent angular velocities to obtain the angular acceleration change sequence;

[0094] Specifically, the angular acceleration change sequence is obtained by calculating the difference between the last reading and the previous reading in the above continuous values; the peaks and troughs in the sequence correspond to the instantaneous burst force when the user applies force to open the cabinet door, and the reverse deceleration generated by the compression and buffering of the sealing strip when the cabinet door is closed, respectively, which is used to accurately capture the sudden change point of mechanical motion.

[0095] S2-4. When the angular acceleration change sequence abruptly changes from the zero value range to a positive peak value, mark this moment as the starting point for wine collection.

[0096] Specifically, when the zero value range suddenly changes to a positive peak value, it means that the wine cabinet door, which was originally in a static state attracted by the magnetic strip, is instantly broken from its equilibrium by the pulling or pushing force applied by the user, resulting in obvious rotational acceleration. This marks the actual start of the user's effective access operation and eliminates the possibility of slight shaking of the cabinet door caused by environmental vibration.

[0097] S2-5. After the starting point of wine collection, continuously monitor the angular acceleration change sequence until the signal returns from the negative peak to the zero range and remains stable. Mark this moment as the end point of wine collection.

[0098] S2-6. Define the time interval between the start point and the end point of wine retrieval as the storage and retrieval cycle.

[0099] In this embodiment, the access behavior cycle specifically refers to the duration of the complete physical process from the moment the user exerts force to pull open the cabinet door, through opening the door to take out the wine, pushing the door to close, the cabinet door hitting the door frame and generating reverse vibration, until the cabinet door is completely closed and the internal mechanical vibration completely subsides and returns to a static state.

[0100] In this embodiment, step S3 specifically includes:

[0101] S3-1, Predefined area exposure weight and door seal airtightness attenuation coefficient;

[0102] S3-2. Calculate the effective convection cross-sectional area based on the sinusoidal projection of the cabinet door opening angle and the regional exposure weight.

[0103] The formula for calculating the effective convection cross-sectional area is:

[0104] ;

[0105] in, Indicates the effective convection cross-sectional area. Indicates the physical width of the wine cabinet doors. This indicates the physical height of the wine cabinet door; the product of the two represents the maximum geometric area of ​​the door. This indicates the opening angle of the cabinet door relative to its closed state at the current moment. This represents the actual opening width perpendicular to the airflow direction, used to convert the cabinet door's rotation angle into the actual opening width, simulating the physical fact that the wider the door opens, the wider the cold air outlet channel. This represents the regional exposure weight, used to correct for the obstruction coefficient to air convection caused by the different density of wine bottles in different wine storage areas (such as red wine and champagne).

[0106] S3-3. Introduce the door seal airtightness attenuation coefficient, correct the failure rate of the sealing strip caused by the large angle opening of the cabinet door, and calculate the heat exchange efficiency.

[0107] The formula for calculating the heat exchange efficiency is:

[0108] ;

[0109] in, This represents the heat exchange efficiency factor, characterizing how easily external heat can penetrate the interior when the device is currently open. A value closer to 1 indicates a better seal and less heat penetration. The door seal airtightness attenuation coefficient is a preset constant that represents the maximum proportion of the sealing strip that fails when the cabinet door is fully opened. This describes the process by which the magnetic sealing strips around the cabinet door gradually lose contact as the opening angle θ increases, resulting in a non-linear decrease in sealing performance. The efficiency is highest when θ=0, and decreases rapidly as the angle increases, simulating the phenomenon of cold air flowing out like a waterfall when the door is opened at a large angle.

[0110] S3-4. Multiply the effective convection cross-sectional area and heat exchange efficiency by the ambient temperature gradient to calculate the instantaneous heat intrusion rate;

[0111] The formula for calculating the instantaneous heat intrusion rate is:

[0112] ;

[0113] in, It represents the instantaneous heat intrusion rate, which is the total amount of heat entering the wine cabinet through the open door gap per unit time. The temperature gradient, calculated by subtracting the set temperature inside the wine cabinet from the real-time external temperature, represents the driving force for heat inflow. This formula indicates that the heat intrusion rate is determined by the "opening size" (…). ), good or bad sealing ( ), and "internal and external temperature difference" ( The three factors together determine the actual heat load caused by the user's door-opening behavior, which is accurately quantified.

[0114] In this embodiment, step S4 specifically includes:

[0115] S4-1, Preset critical opening angle and heat loss surge slope;

[0116] Specifically, the critical opening angle and the slope of the surge in heat loss are used to represent the angle threshold at which the wine cabinet door transitions from a "slightly open" state to a "fully open" state, and the acceleration factor of heat intrusion caused by a unit increase in angle after exceeding the threshold. The former corresponds to the inflection point angle at which the magnetic sealing strip begins to detach from the contact area over a large area, while the latter quantifies the degree of heat exchange deterioration when the cabinet door is fully open and cold air pours out like a waterfall.

[0117] S4-2. Construct a heat loss function based on the cabinet door opening angle. When the opening angle is less than the critical opening angle, the heat loss increases linearly. When the opening angle exceeds the critical opening angle, the heat loss increases exponentially with the heat loss surge slope.

[0118] The heat loss function is constructed by simulating the physical sealing characteristics of the wine cabinet door at different opening stages: when the door opening angle is small (below the critical opening angle), the door seal still maintains partial contact, and external heat mainly penetrates slowly through the gaps. Therefore, the heat loss shows a gentle linear upward trend with the increase of the angle. Once the opening angle exceeds the critical value, the door seal completely detaches, and the door opening forms a direct air convection channel. At this time, the heat loss no longer increases uniformly with the angle, but is controlled by the "heat loss surge slope", showing an explosive exponential growth, which accurately reflects the real working condition of rapid loss of cold air and sharp increase of heat load under large opening.

[0119] S4-3. The heat loss function and the instantaneous heat intrusion rate are weighted and fused together, and the cold energy compensation correction is calculated by exponential mapping.

[0120] The formula for calculating the cooling capacity compensation correction is:

[0121] ;

[0122] It is a multiplication factor greater than or equal to 1, used to guide how much additional cooling capacity the refrigeration system needs to output to offset the thermal shock caused by opening the door; It is a weighted fusion weighting coefficient used to balance the contribution ratio of "structural heat loss caused by door opening angle" and "instantaneous heat intrusion driven by ambient temperature difference" to the total heat load; That is, the output value of the heat loss function constructed in the previous step at the current opening angle θ. It quantifies the basic heat loss intensity caused by the degree of failure of the sealing structure due to different cabinet door opening angles. The instantaneous heat intrusion rate calculated above represents the actual rate of heat inflow driven by the temperature difference between the inside and outside per unit time. This formula, through exponential mapping, ensures that when the user keeps the cabinet door open for a long time, resulting in severe heat loss, the system can issue a powerful compensation command that far exceeds linearity, preventing the temperature inside the cabinet from running out of control.

[0123] In this embodiment, the cold air compensation correction amount is not a fixed value, but rather increases non-linearly and dynamically as the user's door opening angle increases. This is specifically used to quantify the degree of instantaneous heat flow deterioration caused by seal failure under large opening angles. For example, if a user only opens the door to 5 degrees to take out a bottle of wine, the heat loss is... With a small cooling capacity compensation correction value close to 1, the system requires almost no additional cooling; however, if the user turns on the 60-degree setting and stays there for 10 seconds, heat loss will occur. The rate of increase is rapid due to entering the exponential zone, coupled with the instantaneous heat intrusion rate. This could cause the cooling compensation correction to spike to over 3, requiring the system to amplify the instantaneous heat intrusion rate by this factor. As for the cumulative effect of the door opening duration on the total cooling capacity, the intensity correction and total amount calculation are then achieved by multiplying it by the access behavior cycle.

[0124] In this embodiment, step S5 specifically includes:

[0125] S5-1. Multiply the baseline cooling consumption by the duration of the access behavior cycle to obtain the static background cooling consumption.

[0126] S5-2. Multiply the instantaneous heat intrusion rate with the cold compensation correction amount to obtain the corrected heat intrusion rate;

[0127] Specifically, the cold energy compensation correction amount, as a nonlinear amplification factor, is applied to the instantaneous heat intrusion rate to reflect the sealing failure and intensified convection heat transfer effect caused by the increased opening angle of the cabinet door. When the opening angle of the cabinet door is small, the cold energy compensation correction amount is close to 1, and the corrected heat intrusion rate is approximately equal to the instantaneous heat intrusion rate. When the opening angle of the cabinet door exceeds the critical threshold, the cold energy compensation correction amount increases exponentially, making the corrected heat intrusion rate significantly higher than the original instantaneous heat intrusion rate, thereby quantifying the explosive characteristics of heat intrusion under large opening angles.

[0128] S5-3. Multiply the corrected heat intrusion rate with the access behavior cycle to generate dynamic impact cooling loss.

[0129] S5-4. Calculate the sum of static background cooling loss and dynamic impact cooling loss to generate the minimum necessary cooling capacity under the current wine bottle storage and retrieval behavior.

[0130] Specifically, the static background cooling loss represents the baseline heat that naturally penetrates due to environmental temperature differences under sealed conditions; the dynamic impact cooling loss quantifies the additional heat load caused by opening the cabinet door, i.e., the integral value of the instantaneous heat intrusion rate after being amplified by the cooling loss compensation correction, combined with the duration of the behavior. The minimum necessary cooling loss generated by the superposition of the two accurately defines the theoretical total cooling loss required to offset the total heat load and bring the cabinet temperature back to a steady state, taking into account both the needs of maintaining the basic temperature field and compensating for sudden thermal shocks.

[0131] In this embodiment, the determination of the threshold combines dynamic integration with real-time correction. When the cabinet door opening angle is small and the cycle is short, the cooling capacity compensation correction approaches the baseline, and the threshold is dominated by the static background cooling consumption, requiring only a small amount of cooling from the system. Conversely, when the opening angle exceeds the critical point, causing a sharp increase in the heat intrusion rate, the cooling capacity compensation correction increases exponentially, and the dynamic impact cooling consumption becomes dominant. The system then raises the upper limit of the threshold and forces high-frequency cooling to ensure that the temperature field quickly returns to a steady state after the door is closed, avoiding overshoot or lag.

[0132] In this embodiment, step S6 specifically includes:

[0133] S6-1. Extract M basic cooling consumption benchmarks, instantaneous heat intrusion rate, cabinet door opening angle, and storage and retrieval behavior cycle based on historical database;

[0134] S6-2. Perform feature processing on M basic cooling loss benchmarks, instantaneous heat intrusion rate, cabinet door opening angle and storage and retrieval cycle to generate corresponding basic cooling loss features, heat intrusion features, cabinet door opening features and storage and retrieval cycle features.

[0135] S6-3. Perform feature splicing on the basic cold loss characteristics, heat intrusion characteristics, cabinet door opening characteristics and storage and retrieval cycle characteristics to generate quadruple storage and retrieval behavior samples until M quadruple storage and retrieval behavior samples are obtained.

[0136] S6-4. Obtain the net cooling loss value corresponding to each quadruple access behavior sample and define it as the target label of the quadruple access behavior sample.

[0137] Specifically, the net cooling loss value refers to the difference between the total actual cooling loss inside the wine cabinet and the theoretically required cooling loss during a complete wine bottle storage and retrieval operation. It is calculated as follows: the total cooling energy consumed from the opening of the door until the temperature recovers to the set threshold after the door is closed, minus the baseline energy consumption naturally consumed by the system due to environmental infiltration and compressor start-stop during the same time period if there were no door opening operation. This value directly reflects the true impact of user behavior (such as door opening angle, duration, and frequency) on the heat load of the wine cabinet and is the core quantitative indicator for measuring "extra energy consumption caused by human interference".

[0138] S6-5. Input the quadruple access behavior samples and target labels into the decision tree network for supervised training, learn the complex correspondence between wine bottle access behavior and actual cold loss, and generate a cold loss prediction model.

[0139] Specifically, in this embodiment, the decision tree network prioritizes sampling the general model of XGBoost. Its network architecture is an ensemble learning structure based on gradient boosting decision trees (GBDT), consisting of multiple weak classifiers that iterate sequentially. Each tree focuses on correcting the residual error of the previous tree, and finally outputs the prediction result through weighted summation. Therefore, the features of XGBoost can efficiently handle complex feature combinations, automatically capture nonlinear interaction relationships between variables, and have strong anti-interference ability against abnormal door opening behaviors (such as doors not being closed for a long time or frequent opening and closing), avoiding overfitting while ensuring prediction accuracy. In this embodiment, the model uses quadruple samples as input vectors and net cold loss value as regression target. After cross-validation and parameter tuning, it can accurately predict the cold loss trend under different bottle storage and retrieval behaviors, providing a dynamic adjustment basis for subsequent intermittent refrigeration strategies.

[0140] Furthermore, step S6-2 also includes:

[0141] S6-2-1. Select target type parameters based on the basic cooling consumption benchmark, instantaneous heat intrusion rate, cabinet door opening angle, and storage and retrieval cycle.

[0142] S6-2-2, Obtain the maximum and minimum parameters in the target type parameters;

[0143] S6-2-3. Calculate the parameter limit difference between the maximum and minimum parameters;

[0144] S6-2-4. In the target type parameters, anchor the current feature parameters one by one;

[0145] S6-2-5. Calculate the current parameter difference between the current feature parameter and the minimum parameter value;

[0146] S6-2-6. Perform a ratio operation between the current parameter difference and the parameter limit difference to obtain the target parameter features corresponding to the current feature parameter, until the M target parameter features corresponding to the target type parameter are obtained.

[0147] S6-2-7. Traverse the basic cooling loss baseline, instantaneous heat intrusion rate, cabinet door opening angle and access behavior cycle, and repeatedly calculate the target parameter characteristics until the corresponding basic cooling loss characteristics, heat intrusion characteristics, cabinet door opening characteristics and access cycle characteristics are generated.

[0148] In this embodiment, physical parameters with different dimensions are mapped to a unified dimensionless interval. This eliminates the weight imbalance caused by differences in numerical scales (such as angle values ​​versus time values) of different features, preventing large numerical features from dominating model training. Through this processing, the decision tree network can more fairly evaluate the contribution of each dimension to cooling loss, significantly improving the model's generalization ability to sparse data and abnormal operating conditions, and ensuring the portability of prediction results across wine cabinets of different capacity specifications.

[0149] Specifically, in this embodiment, the triggering intermittent cooling control command in step S9 further includes:

[0150] If the cooling deviation shows that the estimated cooling capacity loss is much greater than the minimum required cooling capacity, it is determined to be a high heat load scenario. The compressor is controlled to run continuously at a high duty cycle, and the opening of the electronic expansion valve is increased.

[0151] If the cooling deviation shows that the estimated cooling capacity loss is slightly greater than the minimum required cooling capacity, it is determined to be a normal access scenario, and pulse cooling is executed: the compressor runs for a preset time and then stops, and the residual cooling in the air duct is used for fine adjustment;

[0152] If the cooling deviation is within the safety tolerance, it is determined to be a minor disturbance scenario. Only the circulating fan is started to equalize the internal temperature, and the compressor is not started.

[0153] In this embodiment, the duty cycle limitation is lifted under high heat load scenarios, and the maximum cooling power is used to quickly fill the energy gap; in normal scenarios, the residual cooling of the evaporator after the compressor stops is used for pulsed fine-tuning, balancing temperature control accuracy and start-stop lifespan; in perturbation scenarios, ineffective start-stop is avoided only by equalizing the temperature of the fan. This tiered strategy ensures that the system outputs the necessary cooling capacity only when necessary, significantly reducing energy waste caused by ineffective cooling.

[0154] This invention also provides an intermittent refrigeration system for a refrigerated wine cabinet based on wine bottle storage and retrieval behavior recognition. This system is used to implement the above-described method embodiments, and details already described will not be repeated. The terms "module," "unit," and "subunit" used below refer to combinations of software and / or hardware that perform a predetermined function. Although the system described in the following embodiments is preferably implemented in software, hardware implementation, or a combination of software and hardware, is also possible and contemplated.

[0155] Figure 5 This is a structural block diagram of an intermittent refrigeration system for a refrigerated wine cabinet based on wine bottle storage and retrieval behavior recognition according to the present invention. The system includes:

[0156] An intermittent refrigeration system for a refrigerated wine cabinet based on wine bottle storage and retrieval behavior recognition includes:

[0157] An angular velocity acquisition module is used to acquire several adjacent angular velocities of the refrigerated wine cabinet at the door hinge.

[0158] The behavior cycle marking module is used to mark the access behavior cycle on the time axis based on the several adjacent angular velocities;

[0159] The instantaneous heat calculation module is used to obtain the cabinet door opening angle and ambient temperature gradient during the access behavior cycle, and to calculate the instantaneous heat intrusion rate.

[0160] The cold energy compensation correction module is used to calculate the cold energy compensation correction amount based on the instantaneous heat intrusion rate and the cabinet door opening angle.

[0161] The minimum cooling capacity calculation module is used to obtain the basic cooling consumption benchmark of the wine cabinet in a static sealed state, and calculate the minimum necessary cooling capacity under the current wine bottle storage and retrieval behavior by combining the cooling capacity compensation correction amount and the storage and retrieval behavior cycle.

[0162] The behavior sample construction module is used to construct four-tuple access behavior samples based on the basic cooling consumption benchmark, instantaneous heat intrusion rate, cabinet door opening angle, and access behavior cycle.

[0163] The cooling loss output module is used to input the quadruple access behavior samples into the pre-trained cooling loss prediction model and output the predicted cooling loss.

[0164] The refrigeration deviation calculation module is used to compare the minimum required cooling capacity and the cooling capacity loss to calculate the refrigeration deviation.

[0165] The intermittent cooling control module is used to determine if the cooling deviation exceeds the preset safety tolerance, and then trigger the intermittent cooling control command.

[0166] In the above system, several adjacent angular velocities are acquired through an angular velocity acquisition module; the storage and retrieval behavior cycle is marked through a behavior cycle marking module; the instantaneous heat intrusion rate is calculated through an instantaneous heat calculation module; the cold energy compensation correction amount is calculated through a cold energy compensation correction module; the minimum necessary cold energy is calculated through a minimum cold energy calculation module; a behavior sample construction module is used to construct a four-tuple storage and retrieval behavior sample; the predicted cold energy loss is output through a cold energy loss output module; the refrigeration deviation is calculated through a refrigeration deviation calculation module; and an intermittent refrigeration control module determines whether an intermittent refrigeration control command is triggered if the refrigeration deviation exceeds a preset safety tolerance. This solves the problem of over-refrigeration under small disturbances and insufficient compensation under large opening impacts.

[0167] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any changes or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application.

Claims

1. A method for intermittent refrigeration of a refrigerated wine cabinet based on wine bottle access behavior recognition, characterized by, include: S1. Obtain several adjacent angular velocities at the door hinge of the refrigerated wine cabinet; S2. Based on the aforementioned adjacent angular velocities, mark the access behavior cycle on the time axis; S3. During the access cycle, obtain the cabinet door opening angle and the ambient temperature gradient to calculate the instantaneous heat intrusion rate. S4. Calculate the cooling compensation correction amount based on the instantaneous heat intrusion rate and the cabinet door opening angle; S5. Obtain the basic cooling consumption benchmark of the wine cabinet in a static sealed state, and calculate the minimum necessary cooling capacity under the current wine bottle storage and retrieval behavior by combining the cooling capacity compensation correction amount and the storage and retrieval behavior cycle. S6. Based on the basic cooling consumption benchmark, instantaneous heat intrusion rate, cabinet door opening angle and storage and retrieval behavior cycle, construct a quadruple storage and retrieval behavior sample. S7. Input the quadruple access behavior samples into the pre-trained cold loss prediction model and output the predicted cold loss. S8. Compare the minimum required cooling capacity and cooling capacity loss to calculate the refrigeration deviation; S9. If the cooling deviation exceeds the preset safety tolerance, an intermittent cooling control command will be triggered.

2. The intermittent refrigeration method of a cold wine cabinet based on wine bottle access behavior recognition according to claim 1, characterized in that, Based on the aforementioned adjacent angular velocities, the access behavior cycle is marked on the time axis, including: S2-1. Establish a behavior capture window that slides over time, with a fixed length and the end of the window always aligned with the current sampling time; S2-2. Within the behavior capture window, collect several adjacent angular velocities at the cabinet door hinge with a set sampling step size; S2-3. Perform a difference operation on several adjacent angular velocities to obtain the angular acceleration change sequence; S2-4. When the angular acceleration change sequence abruptly changes from the zero value range to a positive peak value, mark this moment as the starting point for wine collection. S2-5. After the starting point of wine collection, continuously monitor the angular acceleration change sequence until the signal returns from the negative peak to the zero range and remains stable. Mark this moment as the end point of wine collection. S2-6. Define the time interval between the start point and the end point of wine retrieval as the storage and retrieval cycle.

3. The intermittent refrigeration method of a cold wine cabinet based on wine bottle access behavior recognition according to claim 1, characterized in that, The calculation of the instantaneous heat intrusion rate includes: S3-1, Predefined area exposure weight and door seal airtightness attenuation coefficient; S3-2. Calculate the effective convection cross-sectional area based on the sinusoidal projection of the cabinet door opening angle and the regional exposure weight. S3-3. Introduce the door seal airtightness attenuation coefficient, correct the failure rate of the sealing strip caused by the large angle opening of the cabinet door, and calculate the heat exchange efficiency. S3-4. Multiply the effective convection cross-sectional area and heat exchange efficiency by the ambient temperature gradient to calculate the instantaneous heat intrusion rate.

4. The intermittent refrigeration method of a cold wine cabinet based on wine bottle access behavior recognition according to claim 1, characterized in that, Based on the instantaneous heat intrusion rate and the cabinet door opening angle, calculate the cooling capacity compensation correction, including: S4-1, Preset critical opening angle and heat loss surge slope; S4-2. Construct a heat loss function based on the cabinet door opening angle. When the opening angle is less than the critical opening angle, the heat loss increases linearly. When the opening angle exceeds the critical opening angle, the heat loss increases exponentially with the heat loss surge slope. S4-3. The heat loss function and the instantaneous heat intrusion rate are weighted and fused together, and the cold energy compensation correction is calculated by exponential mapping.

5. The intermittent refrigeration method of a cold wine cabinet based on wine bottle access behavior recognition according to claim 4, characterized in that, Calculate the minimum required cooling capacity for the current bottle storage and retrieval behavior, including: S5-1. Multiply the baseline cooling consumption by the duration of the access behavior cycle to obtain the static background cooling consumption. S5-2. Multiply the instantaneous heat intrusion rate with the cold compensation correction amount to obtain the corrected heat intrusion rate; S5-3. Multiply the corrected heat intrusion rate with the access behavior cycle to generate dynamic impact cooling loss. S5-4. Calculate the sum of static background cooling loss and dynamic impact cooling loss to generate the minimum necessary cooling capacity under the current bottle storage and retrieval behavior.

6. The intermittent refrigeration method for a refrigerated wine cabinet based on wine bottle storage and retrieval behavior recognition according to claim 1, characterized in that, The pre-training steps of the cooling loss prediction model include: S6-1. Extract M basic cooling consumption benchmarks, instantaneous heat intrusion rate, cabinet door opening angle, and storage and retrieval behavior cycle based on historical database; S6-2. Perform feature processing on M basic cooling loss benchmarks, instantaneous heat intrusion rate, cabinet door opening angle and storage and retrieval cycle to generate corresponding basic cooling loss features, heat intrusion features, cabinet door opening features and storage and retrieval cycle features. S6-3. Perform feature splicing on the basic cold loss characteristics, heat intrusion characteristics, cabinet door opening characteristics and storage and retrieval cycle characteristics to generate quadruple storage and retrieval behavior samples until M quadruple storage and retrieval behavior samples are obtained. S6-4. Obtain the net cooling loss value corresponding to each quadruple access behavior sample and define it as the target label of the quadruple access behavior sample. S6-5. Input the quadruple access behavior samples and target labels into the decision tree network for supervised training, learn the complex correspondence between wine bottle access behavior and actual cold loss, and generate a cold loss prediction model.

7. The intermittent refrigeration method for a refrigerated wine cabinet based on wine bottle storage and retrieval behavior recognition according to claim 6, characterized in that, The generation of corresponding basic cooling loss characteristics, heat intrusion characteristics, cabinet door opening characteristics, and storage / retrieval cycle characteristics includes: S6-2-1. Select target type parameters based on the basic cooling consumption benchmark, instantaneous heat intrusion rate, cabinet door opening angle, and storage and retrieval cycle. S6-2-2, Obtain the maximum and minimum parameters in the target type parameters; S6-2-3. Calculate the parameter limit difference between the maximum and minimum parameters; S6-2-4. In the target type parameters, anchor the current feature parameters one by one; S6-2-5. Calculate the current parameter difference between the current feature parameter and the minimum parameter value; S6-2-6. Perform a ratio operation between the current parameter difference and the parameter limit difference to obtain the target parameter features corresponding to the current feature parameter, until the M target parameter features corresponding to the target type parameter are obtained. S6-2-7. Traverse the basic cooling loss baseline, instantaneous heat intrusion rate, cabinet door opening angle and access behavior cycle, and repeatedly calculate the target parameter characteristics until the corresponding basic cooling loss characteristics, heat intrusion characteristics, cabinet door opening characteristics and access cycle characteristics are generated.

8. The intermittent refrigeration method for a refrigerated wine cabinet based on wine bottle storage and retrieval behavior recognition according to claim 7, characterized in that, Execute intermittent cooling control commands, including: If the cooling deviation shows that the estimated cooling capacity loss is much greater than the minimum required cooling capacity, it is determined to be a high heat load scenario. The compressor is controlled to run continuously at a high duty cycle, and the opening of the electronic expansion valve is increased. If the cooling deviation shows that the estimated cooling capacity loss is slightly greater than the minimum required cooling capacity, it is determined to be a normal access scenario, and pulse cooling is executed: the compressor runs for a preset time and then stops, and the residual cooling in the air duct is used for fine adjustment; If the cooling deviation is within the safety tolerance, it is determined to be a minor disturbance scenario. Only the circulating fan is started to equalize the internal temperature, and the compressor is not started.

9. An intermittent refrigeration system for a refrigerated wine cabinet based on wine bottle storage and retrieval behavior recognition, characterized in that, include: An angular velocity acquisition module is used to acquire several adjacent angular velocities of the refrigerated wine cabinet at the door hinge. The behavior cycle marking module is used to mark the access behavior cycle on the time axis based on the several adjacent angular velocities; The instantaneous heat calculation module is used to obtain the cabinet door opening angle and ambient temperature gradient during the access behavior cycle, and to calculate the instantaneous heat intrusion rate. The cold energy compensation correction module is used to calculate the cold energy compensation correction amount based on the instantaneous heat intrusion rate and the cabinet door opening angle. The minimum cooling capacity calculation module is used to obtain the basic cooling consumption benchmark of the wine cabinet in a static sealed state, and calculate the minimum necessary cooling capacity under the current wine bottle storage and retrieval behavior by combining the cooling capacity compensation correction amount and the storage and retrieval behavior cycle. The behavior sample construction module is used to construct four-tuple access behavior samples based on the basic cooling consumption benchmark, instantaneous heat intrusion rate, cabinet door opening angle, and access behavior cycle. The cooling loss output module is used to input the quadruple access behavior samples into the pre-trained cooling loss prediction model and output the predicted cooling loss. The refrigeration deviation calculation module is used to compare the minimum required cooling capacity and the cooling capacity loss to calculate the refrigeration deviation. The intermittent cooling control module is used to determine if the cooling deviation exceeds the preset safety tolerance, and then trigger the intermittent cooling control command.

Images