Intelligent defrosting control method and system for refrigerator

Abstract

The invention discloses an intelligent defrosting control method and system for a refrigerator, and relates to the technical field of refrigeration equipment operation control and intelligent energy-saving management. The device comprises a goods temperature sensor array module, an evaporator state sensor, a business information acquisition module, a controller, a remote cloud server and a heat energy storage module. According to the control method, based on multi-source operation state data, intelligent decision making of defrosting opportunities and modes is achieved by constructing a multi-objective optimization function, dynamically adjusting weight coefficients and optimizing strategies in a closed-loop iteration mode. The problems that a traditional defrosting mode is rigid, energy consumption is high, goods temperature fluctuation is large, and the business experience is affected are solved, through data-driven self-adaptive optimization, energy saving, goods temperature stability and low business interference are considered, the method is suitable for various refrigerator types and application scenes, and the intelligent level and comprehensive benefits of refrigerator operation are remarkably improved.

Description

Technical Field

[0001] This invention relates to the field of refrigeration equipment operation control and intelligent energy-saving management technology, and in particular to an intelligent defrosting control method and system for freezers. Background Technology

[0002] During long-term operation, frost gradually accumulates on the evaporator surface of a freezer, leading to decreased heat exchange efficiency, increased evaporation temperature, and increased energy consumption. To maintain cooling performance, traditional freezers typically use timed defrosting or temperature-controlled defrosting methods, but these methods have the following problems: Defrosting is timed in a fixed manner and consumes a lot of energy. The fixed defrosting cycle does not take into account environmental conditions, cargo temperature, or changes in customer flow, which can easily lead to situations where defrosting is performed before frost is formed or after severe frost buildup, increasing energy consumption and shortening compressor life. It also lacks a cargo temperature protection mechanism, causing the evaporator to stop cooling during defrosting, which can easily cause the cargo temperature to rise. If not properly controlled, this could lead to food temperatures exceeding limits, affecting food quality and safety. Furthermore, it cannot be combined with business information and electricity price scheduling. The existing system does not take into account electricity price fluctuations or customer flow distribution. Defrosting during peak hours will affect the pickup experience, while defrosting during periods of high electricity prices will increase operating costs.

[0003] Therefore, an intelligent defrosting scheduling system is proposed that can integrate information on cargo temperature, frost layer status, passenger flow, and electricity price to achieve synergistic optimization of energy saving and food safety. Summary of the Invention

[0004] This invention aims to overcome the shortcomings of existing technologies and provide a method and system for intelligent defrosting control of freezers. By collecting multi-source data through multi-module collaboration, a dynamic optimization model is constructed to achieve precise decision-making on the timing and method of defrosting, taking into account energy saving, stable product temperature and low business interference, thereby improving the intelligence and economy of freezer operation.

[0005] The technical solution adopted by the present invention to solve the above problems is as follows: A method for intelligent defrosting control of a freezer includes: S1. Data Acquisition: The operating status data of the freezer is acquired through preset sensing devices and information acquisition modules. The operating status data includes the temperature data of multiple shelf layers inside the freezer, the frost status data of the evaporator, the real-time energy consumption data of the freezer, the passenger flow data of the environment where the freezer is located, and the real-time electricity price information data. S2. Model Construction: Based on the runtime status data obtained in S1, perform the following operations: s21. Calculate key optimization indicators, including defrosting energy consumption based on the real-time energy consumption data, cargo temperature fluctuation deviation based on the cargo temperature data, and customer flow impact based on the customer flow data. s22. Using the defrosting energy consumption, temperature fluctuation deviation and customer flow impact calculated in step s21 as the core optimization objectives, construct a multi-objective optimization function, and then establish a defrosting optimization model; S3. Decision-making: The defrosting optimization model is optimized and trained. Based on the evaluation results output by the trained model and combined with the preset judgment conditions, it is determined whether to start the defrosting program of the freezer. S4. Defrosting Execution: If step S3 determines that a defrosting procedure needs to be initiated, the optimal defrosting scheme is determined from electric heating defrosting, hot gas bypass defrosting, waste heat recovery defrosting, and thermal energy storage assisted defrosting methods based on the evaluation results of the multi-objective optimization function and the real-time electricity price information data obtained in step S1, and the defrosting operation is triggered.

[0006] Compared with existing technologies, the advantages of this approach include: constructing a multi-objective optimization model based on multi-dimensional operational status data to achieve precise and dynamic defrosting decisions, solving the problems of energy waste and large temperature fluctuations in existing technologies; combining real-time electricity prices and evaporator frost status to reduce defrosting energy consumption and electricity costs, and extend equipment lifespan; balancing temperature deviation control across multiple shelf layers inside the freezer to improve product preservation and reduce spoilage rates; adapting to changes in customer traffic to adjust defrosting timing, balancing freezer refrigeration efficiency and customer shopping experience; and enhancing the scientific nature of defrosting control through multi-objective optimization decisions, thus helping freezer operations achieve cost reduction and efficiency improvement.

[0007] The multi-objective optimization function achieves computability of objectives by quantifying the operational state data of step S1, wherein: The cargo temperature fluctuation deviation is the cumulative difference between the cargo temperature data in step S1 and the preset standard cargo temperature. The defrosting energy consumption is calculated based on the real-time energy consumption data of step S1 combined with the estimated defrosting time. The customer traffic impact is the weighted sum of customer visit frequency and dwell time per unit time.

[0008] Compared with existing technologies, the beneficial effects are as follows: by clearly quantifying and defining the impact of cargo temperature fluctuation deviation, defrosting energy consumption, and customer flow, the input parameters of the multi-objective optimization model are calculable and accurate, avoiding the distortion of optimization results caused by parameter ambiguity, ensuring that defrosting decisions can accurately anchor the core objectives of stable cargo temperature, reduced energy consumption, and customer flow adaptation, and improving the scientificity and reliability of model decisions.

[0009] The multi-objective optimization function is: ,in These are the weighting coefficients. For defrosting energy consumption, To account for temperature fluctuations during defrosting, Impact on customer traffic.

[0010] The multi-objective optimization function takes the running status data of step S1 as its core input, and the weight coefficients are dynamically adjusted according to the defrosting scenario. The specific adjustment rules are as follows: Peak customer traffic scenario: When the frequency of customer visits exceeds 10 times / minute within a unit of time, set... =0.4, =0.2, =0.4; Low electricity price scenario: When the real-time electricity price is < 0.3 yuan / kWh, set =0.1, =0.5, =0.4; Severe frosting scenario: When the frost thickness on the evaporator is >5mm, set... =0.6, =0.2, =0.2; Other scenarios: The default weighting coefficient is set to... =0.3, =0.4, =0.3.

[0011] Compared with existing technologies, the advantages of this approach are as follows: By constructing a clear functional expression, the three core optimization objectives of defrosting energy consumption, cargo temperature fluctuation deviation, and customer flow impact are quantitatively coupled. At the same time, the weight coefficients of each objective are dynamically adjusted based on different actual operating scenarios such as peak customer flow, low electricity prices, and severe frost. This avoids the shortcomings of fixed weights that cannot adapt to complex scenarios, ensuring that defrosting decisions in different scenarios can prioritize core needs, improving the adaptability and flexibility of the model, and thus achieving the optimal balance of defrosting operations in multiple dimensions such as energy consumption, cargo temperature, and customer experience.

[0012] The evaporator frost status data includes the frost thickness and heat exchange efficiency of the evaporator coil, and the customer traffic data is the frequency of customer visits to the freezer or the length of stay per unit time.

[0013] Compared with existing technologies, the advantages are as follows: By clearly defining the evaporator frost status data to include frost thickness and heat exchange efficiency, the actual degree of frost on the evaporator and its impact on refrigeration performance can be reflected more comprehensively and accurately, avoiding misjudgments of defrosting timing caused by relying solely on a single frost thickness parameter; by clearly defining customer traffic data as the frequency of visits or duration of stay per unit time, a clear and collectable basic indicator is provided for the quantitative calculation of the impact of customer traffic, ensuring the effective implementation of customer traffic factors in defrosting decisions and further improving the accuracy and rationality of defrosting control.

[0014] In S3, the optimization training uses the historical operating status data collected in S1 as training samples. By iteratively adjusting the model parameters, the matching degree between the evaluation result output by the model and the actual defrosting needs is higher than a preset threshold. The evaluation result is a defrosting necessity score calculated based on the normalized value of the multi-objective optimization function J.

[0015] Compared with existing technologies, the advantages of using historical operating data as training samples to iteratively adjust model parameters allow the defrosting optimization model to fully learn the defrosting demand patterns under different operating conditions of the freezer, improving the model's adaptability. By using the normalized value of the multi-objective optimization function J as a defrosting necessity score, a quantitative evaluation of multi-dimensional optimization objectives is achieved. Combined with a preset threshold to determine whether to start the defrosting program, errors from subjective experience judgments are avoided, ensuring that the model's output evaluation results match the actual defrosting needs as expected, further improving the accuracy and reliability of defrosting decisions.

[0016] The evaluation results include a defrosting necessity score predicted by the model. When the defrosting necessity score is higher than a set threshold and the frost thickness of the evaporator collected in step S1 exceeds 3 mm, and the heat exchange efficiency of the evaporator is lower than 70% of the rated value, the defrosting procedure is determined to be started.

[0017] Compared with existing technologies, the advantages of this technology are as follows: by coupling the defrosting necessity score with two key hardware operating parameters, namely evaporator frost thickness and heat exchange efficiency, multiple conditions are coupled to determine the defrosting necessity. This avoids the risk of misjudging defrosting based on a single score or a single hardware parameter. It ensures that the defrosting program is only started when the model determines that defrosting is necessary and the actual frost on the evaporator has affected the refrigeration performance. This effectively eliminates ineffective defrosting operations, reduces defrosting energy consumption, and ensures the refrigeration efficiency of the freezer and the preservation effect of the goods.

[0018] The selection of the optimal defrosting scheme in step S4 should be based on the real-time electricity price information data obtained in step S1, combined with the characteristics of the scenario, and the specific rules are as follows: During off-peak electricity prices: When the real-time electricity price is less than 0.3 yuan / kWh, electric heating defrosting should be the preferred method. During peak electricity price periods: When the real-time electricity price is greater than 0.8 yuan / kWh, waste heat recovery defrosting should be the preferred method. During peak passenger flow periods: prioritize hot air bypass defrosting; In all scenarios, the defrosting process requires the thermal energy storage module 6 to provide cooling compensation in order to reduce the temperature fluctuation of the goods during the defrosting process.

[0019] Compared with existing technologies, the advantages of this approach are as follows: By combining real-time electricity prices with the differentiated selection of defrosting methods based on scenario characteristics, electric heating defrosting is prioritized during off-peak electricity periods to control electricity costs, waste heat recovery defrosting is used during peak electricity periods to reduce energy consumption, and hot air bypass defrosting is used during peak customer traffic periods to shorten defrosting time and reduce the impact on the customer shopping experience. At the same time, in all scenarios, the thermal energy storage module provides cold compensation, effectively suppressing temperature fluctuations during the defrosting process and preventing damage to the freshness of goods. This achieves synergistic optimization of defrosting costs, operational efficiency, and the freshness of goods, further improving the economy and reliability of refrigerated display case defrosting control.

[0020] It also includes: the freezer controller synchronizing the real-time operating status data collected in step S1 to a remote cloud server; the remote cloud server generating an optimization strategy based on historical operating status data and real-time data; the optimization strategy including weight coefficient adjustment parameters and threshold adjustment parameters; the remote cloud server distributing the optimization strategy to the controller; and the controller updating the scheduling parameters based on the optimization strategy to achieve adaptive optimization of defrosting control.

[0021] Compared with existing technologies, the advantages are as follows: By synchronizing data between the freezer controller and the remote cloud server, and relying on the powerful computing power of the cloud to conduct in-depth analysis of historical and real-time operating status data, optimization strategies for scheduling parameters such as weight coefficients and judgment thresholds are generated and sent to the controller. This allows the parameters of the defrosting optimization model to be dynamically updated according to the long-term operating conditions of the freezer, avoiding the problem of decreased adaptability caused by parameter fixation in the traditional local control mode, and realizing adaptive iterative optimization of defrosting control.

[0022] The intelligent defrosting scheduling system for freezers includes: The cargo temperature sensor array module is configured to collect cargo temperature data from multiple shelf layers inside the refrigerated cabinet in step S1, providing a data basis for calculating cargo temperature fluctuation deviation in step S21. An evaporator status sensor is configured to detect the frost status data of the evaporator in step S1, and output the detection results to the controller to provide support for the training of the defrosting optimization model. The business information collection module is configured to acquire the passenger flow data and real-time electricity price information data in step S1, and synchronously upload the acquired data to the controller and remote cloud server. The controller is configured to perform the defrosting optimization model construction and training in step S22, parameter adjustment and actuator linkage control during the defrosting process, defrosting determination in step S3, and defrosting mode selection in step S4; the remote cloud server is configured to establish a communication connection with the controller, store historical operating status data collected in step S1, provide data support for the defrosting optimization model training in step S22, and send optimization strategies to the controller. A thermal energy storage module is configured to provide cooling compensation during defrosting, reducing temperature fluctuations of the goods during the defrosting process. The module includes a phase change energy storage material and a cooling release device. By pre-storing cooling capacity before defrosting and releasing it during defrosting, it maintains a stable internal temperature for the goods, minimizing temperature fluctuations during defrosting. The temperature should be controlled within ±1℃.

[0023] Compared with existing technologies, the system offers the following advantages: By configuring a cargo temperature sensor array module, an evaporator status sensor, and a business information acquisition module, it achieves accurate collection of multi-dimensional operational data, including refrigerated cargo temperature, evaporator frost status, customer flow, and electricity prices, providing comprehensive and reliable data support for defrosting optimization model construction and decision-making. Through a collaborative architecture between the controller and a remote cloud server, the system not only relies on the controller to complete local defrosting model construction, judgment, and execution control, but also leverages the massive historical data storage and analysis capabilities of the cloud server to generate optimization strategies, enabling adaptive updates of defrosting control parameters. By adding a thermal energy storage module composed of phase change energy storage materials and a cold energy release device, it provides cold energy compensation during the defrosting process, controlling cargo temperature fluctuations within ±1℃, effectively ensuring the freshness of goods. The system's modules have clear division of labor and work together, significantly improving the intelligence and precision of defrosting control, achieving multiple goals of reduced energy consumption, stable cargo temperature, and efficient operation.

[0024] Compared with the prior art, the present invention has the following advantages: By collecting multi-dimensional data, constructing multi-objective optimization models, and co-optimizing in the cloud, the system achieves precise and dynamic defrosting decisions. It selects defrosting methods based on the characteristics of different scenarios such as electricity prices and customer traffic, and uses thermal energy storage modules to compensate for cold air, effectively reducing defrosting energy consumption and electricity costs, strictly controlling temperature fluctuations, and ensuring the freshness of goods. By determining the defrosting program based on multiple conditions, it eliminates ineffective defrosting operations, balances refrigeration efficiency and customer shopping experience, and ultimately achieves a triple improvement in the intelligence, economy, and reliability of freezer defrosting control. Attached Figure Description

[0025] Figure 1 This is a schematic diagram of the control method of the present invention; Figure 2 This is a schematic diagram of the overall structure of the present invention. Figure 3 This is a schematic diagram of the system of the present invention. The following are the labels in the diagram: 1. Cargo temperature sensor array module; 2. Evaporator status sensor; 3. Business information acquisition module; 4. Controller; 5. Remote cloud server; 6. Thermal energy storage module. Detailed Implementation

[0026] The following are specific embodiments of the present invention, and the technical solutions of the present invention will be further described in conjunction with the accompanying drawings. However, the present invention is not limited to these embodiments.

[0027] This invention discloses an intelligent defrosting control method and system for freezers. Using a comprehensive objective optimization function as the core decision-making basis, it solves the problems of rigid timing, singular objectives, and poor fault tolerance in traditional defrosting methods through a closed-loop design involving precise multi-source data acquisition, dynamic model training, and adaptive matching of defrosting strategies. This achieves synergistic optimization of energy saving, stable product temperature, and user experience. It is applicable to commercial upright freezers, refrigerated display cases, and supermarket cold chain terminal equipment. The following details the implementation process of each step and system component in conjunction with specific application scenarios. Example 1

[0028] like Figure 1-3As shown, the present invention discloses an intelligent defrosting control method and system for refrigerated display cases, comprising a preset sensing device and an information acquisition module. This allows for comprehensive and high-precision acquisition of the refrigerated display case's operational status data, providing comprehensive and reliable data support for subsequent optimization decisions. Specifically, a distributed cargo temperature sensor array is employed, with at least three high-precision temperature sensors deployed in different shelf layers and areas within the refrigerated display case. The sensor measurement accuracy is ≤ ±0.1℃, and the sampling frequency is 1 time / 30 seconds. Real-time acquisition of cargo temperature data at each monitoring point is recorded with timestamps, forming a cargo temperature spatiotemporal distribution dataset. A combination of infrared temperature sensors, differential pressure sensors, and capacitive frost thickness sensors is configured to simultaneously acquire evaporator coil temperature difference, inlet and outlet air pressure difference, and frost thickness data (measurement range 0-10mm, accuracy ≤ ±0.1mm), with a sampling frequency of 1 time / 1 minute. After data fusion algorithms remove anomalies, the frost thickness value and heat exchange efficiency attenuation rate are output. A 0.5-level high-precision power metering module is installed in the main power supply circuit of the refrigerated display case, with a sampling frequency of 1 time / 10 seconds. The system collects data on total energy consumption, compressor energy consumption, and other sub-items at a frequency of seconds to clarify the energy consumption baseline related to defrosting; it collects data on customer numbers, dwell time, and pickup frequency through complementary infrared customer flow sensors and video image analysis modules to generate customer flow time-series curves and peak-valley time period divisions; and it connects to the electricity information service platform via an IoT module to obtain real-time electricity prices, time period types, and adjustment warning information at a frequency of once every 15 minutes, comprehensively covering five core data dimensions: shelf temperature data, evaporator frost status data, real-time energy consumption data of the freezer, customer flow data of the freezer's environment, and real-time electricity price information.

[0029] Based on the collected multi-dimensional operational status data, key optimization index calculations, multi-objective optimization function construction, and defrosting optimization model establishment were carried out. In the index calculation stage, an energy consumption decomposition algorithm was used to eliminate energy consumption under non-defrosting conditions. Combined with the frosting status, a linear regression and neural network fusion model was used to predict the energy consumption values ​​(including main and auxiliary energy consumption) for different defrosting methods. Using 0-4℃ as the safe temperature benchmark for goods, the cumulative time and standard deviation of goods temperatures exceeding 4℃ were statistically analyzed. Weights were assigned according to the type of goods (e.g., the cumulative time of exceeding the temperature limit for fresh produce had a weight of 0.6), and the weighted summation yielded the temperature fluctuation deviation. Based on the customer flow time-series curve, peak, valley, and flat intervals were divided, and corresponding influence coefficients were set (valley 0.1-0.3, flat peak 0.4-0.6, peak 0.7-0.9). These were then combined with adjustments based on the pickup frequency to obtain the customer flow influence.

[0030] Multi-objective optimization function This is the core mathematical model for achieving intelligent decision-making in refrigerator defrosting. Essentially, it constructs a unified quantitative evaluation standard by weightedly integrating three key influencing factors in the defrosting process, providing a scientific basis for selecting the timing and method of defrosting. The technical definitions, quantitative logic, and weighting rules for each parameter are as follows: (Defrosting energy consumption) is a core indicator for measuring the energy consumption of defrosting operations. It's not simply a statistical calculation of the total power consumption during defrosting, but rather a refined quantitative value that combines the operating conditions of the freezer with the characteristics of the defrosting method. Specifically, its quantification is based on real-time energy consumption data. An energy consumption decomposition algorithm separates the defrosting-specific energy consumption (excluding energy consumption from non-defrosting conditions such as cooling, standby, and lighting). This energy consumption is then combined with the evaporator's frost condition (frost thickness, heat exchange efficiency) to estimate the defrosting time—for example, with a frost thickness of 3-5mm, electric defrosting is estimated to take 8-10 minutes; with a frost thickness of over 8mm, the estimated time extends to 18-22 minutes. The final time is calculated according to… =Defrosting equipment power (kW) × estimated defrosting time (h) × real-time electricity price (yuan / kWh) is calculated. At the same time, it is necessary to distinguish the energy consumption characteristics of different defrosting methods such as electric heating, hot gas bypass, and waste heat recovery. For example, the energy loss of refrigerant circulation needs to be taken into account for hot gas bypass defrosting, and the energy consumption deduction brought by waste heat recovery efficiency needs to be calculated to ensure the accuracy of energy consumption quantification and the adaptability of operating conditions.

[0031] (Defrost Temperature Fluctuation Deviation) focuses on assessing the impact of defrosting on cargo temperature stability. Its quantification is based on the safe storage temperature of the cargo, combined with spatiotemporal distribution data of cargo temperature for refined calculation. First, a standard cargo temperature range is preset according to the type of stored cargo: 2-4℃ for refrigerated goods (such as fresh produce and dairy products), and -21 to -15℃ for frozen goods (such as frozen foods and meat). Then, time-series data collected by a distributed cargo temperature sensor array is extracted, and the instantaneous deviation between the cargo temperature at each monitoring point and the standard cargo temperature benchmark is calculated in 1-minute increments. If the cargo temperature exceeds the safe range, an over-limit penalty coefficient of 1.5 is multiplied by the absolute value of the instantaneous deviation to highlight the harm of exceeding the limit temperature to cargo quality. Finally, all instantaneous deviations throughout the entire defrosting cycle are accumulated to obtain... The quantitative value (unit: ℃・min) directly reflects the cumulative degree of deviation of the cargo temperature from the ideal state. The smaller the value, the smaller the disturbance of the cargo temperature during the defrosting process.

[0032] (Customer Traffic Impact) is used to quantify the degree of interference of defrosting operations on the customer shopping experience. Its quantification logic combines customer behavior data and scenario characteristics to achieve objective evaluation. Using a 10-minute time window, customer visit frequency (total number of times approaching the freezer, opening the door, and retrieving items) and dwell time (cumulative dwell time within a 1.5-meter radius of the freezer) are collected through traffic sensors and video analysis modules. Weighting coefficients are then set according to the freezer usage scenario: in commercial retail scenarios (supermarkets, convenience stores), the customer visit frequency weighting coefficient is 0.6, and the dwell time weighting coefficient is 0.4, highlighting the direct impact of defrosting on retrieving behavior; in non-customer-facing scenarios (cold chain warehousing), the visit frequency weighting coefficient is reduced to 0.2, and the dwell time weighting coefficient is 0.1, reducing the influence weight of non-core factors. Finally, the impact is assessed using the formula… = (Visit frequency × corresponding weighting coefficient) + (Dwell time / 60 × corresponding weighting coefficient)” The calculation shows that the larger the value, the more frequent the customer activity during the defrosting period, and the higher the risk of interference with the shopping experience.

[0033] As a weighting coefficient, its core function is to dynamically balance the priorities of the three optimization objectives, and satisfy the following: + + The constraint condition of =1 allows for flexible adjustment of weight values ​​based on the usage scenario and operational needs of the refrigerated display case. For example, in food cold chain warehousing scenarios, the temperature stability of goods is a core requirement. It can be set to 0.4-0.5. Set it to 0.3-0.4. Set at 0.1-0.2; in commercial retail scenarios such as supermarkets and convenience stores, customer experience and operating costs are equally important. and All values ​​can be set to 0.35-0.4. Set to 0.2-0.3; in scenarios where energy saving is the priority (such as cold chain transfer freezers operating at night), It can be set to 0.4-0.5. Set it to 0.3-0.4. The weighting coefficient is set to 0.1-0.2; furthermore, the weighting coefficient can be automatically optimized by the system based on historical operating data. For example, when the cargo temperature exceeds the limit multiple times, the system will automatically increase the weighting coefficient. The weight of the cargo temperature protection should be strengthened.

[0034] The technical role of this multi-objective optimization function is reflected in three core aspects: First, it enables unified quantitative evaluation of multi-dimensional objectives, transforming three originally scattered and incomparable indicators—energy cost, product temperature stability, and customer experience—into a single comprehensive objective value J through weighted fusion, thus providing a directly comparable quantitative basis for the merits of different defrosting solutions. Second, it supports dynamic adjustment of decision priorities, adapting to the core needs of different scenarios through flexible setting of weight coefficients, avoiding the disadvantages caused by single-objective optimization (such as pursuing energy saving at the expense of food quality, or focusing solely on customer experience while increasing energy costs). Third, it provides a clear solution direction for the defrosting optimization model. The model iteratively calculates the J value under different combinations of defrosting timing and methods, selecting the optimal decision solution with the smallest J value, ensuring that the defrosting operation achieves comprehensive optimization across the three objectives.

[0035] From a technical perspective, the application of this function has enabled a shift in defrosting decisions from "experience-driven" to "data-driven": in terms of energy consumption control, it achieves precise quantification. Combined with weighted adjustments, energy consumption is reduced by 25%-35% compared to traditional defrosting methods. Especially during periods of high electricity prices, increased weighting can avoid high-cost defrosting, significantly reducing operating costs. Regarding cargo temperature protection... The refined quantification and weight allocation reduce temperature fluctuations by 40%-50% and decrease food spoilage and loss due to defrosting by 12%-18%, effectively ensuring food quality and safety; in terms of customer experience, The quantitative assessment enables the system to accurately identify and initiate defrosting during off-peak hours, reducing customer complaints about defrosting operations by over 90% and improving shopping experience satisfaction by 5%–35%. Simultaneously, the dynamic adaptability of the weighting coefficients allows the function to be flexibly applied to various equipment such as commercial upright freezers, refrigerated display cases, and cold chain storage freezers, as well as multiple scenarios including retail, warehousing, and transit, demonstrating strong versatility and scalability. Furthermore, the unified quantitative model increases the solution speed of the defrosting optimization model by 40%–50%, enabling multi-solution evaluation and selection within 10 seconds, ensuring real-time and efficient decision-making. Example 2

[0036] This embodiment, based on Embodiment 1, performs systematic training and defrosting initiation determination on the defrosting optimization model. Historical operational data from the past 6 months is divided into training, validation, and test sets in a 7:2:1 ratio. Model parameters (such as crossover and mutation probabilities in genetic algorithms) are iteratively adjusted with the goal of minimizing prediction bias. Generalization ability is monitored using the validation set until the prediction error is ≤5%, at which point training stops. Real-time collected operational status data is input into the trained model to obtain a comprehensive evaluation value and the frost constraint satisfaction status: if the frost constraint is satisfied and ≤ a preset threshold (e.g., 0.6 for commercial display cabinets), defrosting is initiated; if the frost constraint is satisfied but > a threshold (e.g., during peak customer traffic and electricity price periods), a 1-2 hour delay suggestion is output, with re-evaluation every 15 minutes; if the frost constraint is not satisfied, the process returns to the data collection stage for continuous monitoring.

[0037] If the defrosting procedure is initiated, the optimal solution is selected from four defrosting methods and dynamically executed based on the evaluation results of the multi-objective optimization function and real-time electricity price information. During periods of low electricity price and when the predicted temperature fluctuation is ≤0.8℃, electric heating defrosting (8-12 minutes, easy to operate) is preferred; during periods of medium electricity price and when the heat exchange efficiency is moderately reduced (50%-70%), hot gas bypass defrosting is selected (energy consumption is 30%-40% lower than electric heating); when the freezer is running continuously and there is residual heat, waste heat recovery defrosting is used (energy consumption is close to zero, realizing energy cascade utilization); in temperature-sensitive scenarios (such as fresh food and dairy product freezers), thermal energy storage-assisted defrosting is preferred, combined with other methods to achieve cold load compensation. After the defrosting operation is triggered, the cargo temperature, frost thickness, and energy consumption data are collected every 5 minutes and fed back to the model for dynamic updates: if the cargo temperature fluctuates by ≥1.5℃, the thermal energy storage compensation is strengthened; if the frost layer is still ≥1mm after 25 minutes, the defrosting is switched to a combination of electric heating and hot gas bypass; when the frost layer is ≤0.5mm and the heat exchange efficiency recovers to more than 90% of the rated value, the defrosting ends and the system switches to normal cooling mode.

[0038] This method utilizes multi-dimensional data collection and precise modeling to scientifically determine the timing of defrosting, solving the problem of ineffective defrosting in traditional timed defrosting and extending the lifespan of core freezer components by 15%–20%. The multi-objective optimization function considers energy consumption, product temperature, and customer flow demands, reducing energy consumption by 25%–35%, product temperature fluctuation by 40%–50%, and the impact of peak customer flow by over 60% compared to traditional methods. The adaptive selection of four defrosting methods adapts to different types of freezers and application scenarios, significantly improving versatility. Dynamic feedback and closed-loop adjustment during the defrosting process effectively avoid incomplete defrosting or temperature control failure, and the model continuously optimizes over time. Simultaneously, by saving energy, extending equipment lifespan, and reducing food waste (by 10%–15%), it significantly reduces operating costs. Waste heat recovery and thermal energy storage auxiliary methods comply with energy conservation and emission reduction policies, achieving both economic and social benefits.

[0039] The core value of the multi-objective optimization function lies in transforming the collected, dispersed operational status data into indicators that can be accurately calculated and quantified for comparison. Through explicit quantitative logic, it provides scientific support for defrosting decisions. The specific quantitative implementation methods of the three core objectives are as follows: The quantification of cargo temperature fluctuation deviation focuses on "reflecting the impact of defrosting on cargo temperature stability." First, based on the type of goods stored in the refrigerated container, tiered and preset standard cargo temperatures are defined: refrigerated goods (such as vegetables and dairy products) have a standard temperature baseline of 3℃, with a safe fluctuation range of 2℃-4℃; frozen goods (such as meat and frozen foods) have a standard temperature baseline of -18℃, with a safe fluctuation range of -21℃ to -15℃. These baseline values ​​can be manually adjusted or automatically optimized by the system based on historical storage data. Then, time-series data from the distributed cargo temperature sensor array is extracted, and the instantaneous difference between the cargo temperature at each monitoring point and the corresponding standard baseline is calculated in 1-minute increments. If the instantaneous cargo temperature is within the safe fluctuation range, the absolute value of the difference is taken; if it exceeds the safe range, the absolute value is multiplied by a 1.5 times over-limit penalty coefficient to strengthen the weighting of the impact of over-limit temperatures on cargo quality. Finally, the instantaneous differences of all statistical units within a single defrosting cycle are summed to obtain a quantitative value of cargo temperature fluctuation deviation in ℃・min. The smaller the value, the better the cargo temperature stability.

[0040] The quantification of defrosting energy consumption focuses on "accurately predicting the energy consumption of the entire defrosting process." The first step, based on real-time energy consumption data from a high-precision electricity metering module, uses an energy consumption decomposition algorithm to eliminate energy consumption from non-defrosting conditions such as cooling, standby, and lighting. It then extracts the energy consumption baseline for defrosting-related equipment, including electric heating elements, bypass valves, and thermal energy storage modules, and optimizes the baseline accuracy by referencing historical defrosting energy consumption data under similar frost conditions. The second step determines the estimated defrosting time based on the collected evaporator frost status data: 8-10 minutes for electric heating defrosting when the frost thickness is 3-5mm and the heat exchange efficiency is 60%-70%; 12-15 minutes for frost thickness is 5-8mm and the heat exchange efficiency is 50%-60%; and 18-22 minutes for frost thickness ≥8mm and the heat exchange efficiency ≤50%. Simultaneously, the unit energy cost corresponding to the real-time electricity price time period (peak, flat, and valley) is recorded. The third step is to calculate the quantitative value according to "Defrosting energy consumption = energy consumption baseline (kWh / minute) × estimated defrosting time (minutes) × unit energy cost (yuan / kWh)". If it is a composite defrosting method, the weighted summation of the energy consumption ratio of each component is used to ensure that the calculation is comprehensive and accurate.

[0041] The quantification of customer traffic impact aims to "objectively reflect the interference of defrosting on the customer shopping experience." Using a 10-minute time window, it counts the total number of times customers approach the freezer, open the door, or retrieve goods (i.e., customer visit frequency) and the cumulative dwell time (in seconds) of all customers within a 1.5-meter radius of the freezer. Weighting coefficients are set based on the freezer's usage scenario: in commercial retail scenarios (supermarkets, convenience stores), the customer visit frequency weighting coefficient is 0.6, and the dwell time weighting coefficient is 0.4, highlighting the direct impact on retrieving behavior; in non-customer-facing scenarios such as cold chain warehousing, the visit frequency weighting coefficient is 0.2, and the dwell time weighting coefficient is 0.1, reducing the impact of non-core factors. Quantification is achieved using the formula "Customer Traffic Impact = (Customer Visit Frequency × Corresponding Weighting Coefficient) + (Customer Dwell Time / 60 × Corresponding Weighting Coefficient)" (dwell time is converted to minutes to ensure consistency). A higher value indicates more frequent customer activity during the defrosting period and a higher risk of interference.

[0042] This quantitative approach provides precise and objective data support for multi-objective optimization functions, transforming the previously ambiguous impact of product temperature, energy costs, and customer experience into calculable indicators with unified dimensions. This avoids subjective biases in decision-making and provides a quantitative basis for the trade-offs among the three core objectives. The optimal solution can be directly selected by comparing the sum of indicators from different defrosting solutions. The quantified indicator data can also serve as training samples for the defrosting optimization model, facilitating dynamic adjustment and iterative upgrades of model parameters. Furthermore, decision priorities can be dynamically determined based on the quantification results. For example, when product temperature fluctuations approach a safety threshold, a solution with lower product temperature impact can be prioritized; when the impact on customer traffic exceeds a threshold, defrosting can be postponed to off-peak hours.

[0043] In terms of effectiveness, the refined quantification of temperature fluctuation deviation upgrades temperature monitoring from "qualitative judgment" to "quantitative management." Food spoilage and loss due to defrosting are reduced by 12%-18%, and the cumulative time of exceeding temperature limits is shortened by more than 60%. The quantified defrosting energy consumption value deviates from the actual energy consumption by ≤5%, accurately avoiding high-cost defrosting during periods of high electricity prices, reducing operating costs by 20%-30%. Quantifying the impact of customer traffic accurately identifies off-peak periods, reducing customer complaints by more than 90% and improving shopping experience satisfaction by 25%-35%. Simultaneously, the quantified data increases the speed of optimization model solving by 40%-50%, completing solution evaluation and selection within 10 seconds. Furthermore, the model maintains a decision accuracy rate of over 92% under new operating conditions. Standardized quantification rules also eliminate reliance on human experience, allowing for rapid replication and application to different types of freezers. This facilitates managers' intuitive understanding of defrosting influencing factors, providing data support for operational optimization. Example 3

[0044] This embodiment, based on Embodiment 1, uses the collected refrigerated display case operating status data (including cargo temperature data, evaporator frost status data, real-time energy consumption data, customer flow data, and real-time electricity price information) as the core input for its multi-objective optimization function. This function precisely quantifies defrosting energy consumption. ), cargo temperature fluctuation deviation ( ) and the impact of customer traffic ( These three indicators form the basis of a mathematical model for comprehensively evaluating the merits of defrosting solutions, with weighting coefficients... , , Dynamic adjustment is key to achieving precise adaptation of optimization goals in different scenarios. Its core logic is to adjust the priority weight of each indicator according to the core requirements of the scenario, ensuring that defrosting decisions are highly matched with actual operational needs. The specific adjustment rules and technical details are as follows: The determination of peak customer traffic scenarios is based on customer traffic data acquired by the business information collection module 3. A specific trigger condition is defined as "customer visit frequency > 10 times / minute within a unit of time." Here, "customer visit frequency" refers to the cumulative number of times per minute a customer approaches the refrigerated display case, opens the door, retrieves goods, or stays within 1.5 meters of the display case for more than 3 seconds. This data is collected and statistically analyzed in real time by an infrared customer flow sensor and a video image analysis module, with a sampling frequency of 1 time / 10 seconds. When the visit frequency meets the requirement of > 10 times / minute for three consecutive sampling periods, the system determines that a peak customer traffic scenario has been entered. The core requirement in this scenario is to minimize the interference of defrosting on the customer shopping experience while ensuring stable goods temperature. Therefore, the following settings are implemented: =0.4 (Customer traffic volume has the highest impact factor) =0.4 (Cargo temperature fluctuation deviation is equally important) =0.2 (defrosting energy consumption weight is the lowest). The weight allocation guides the optimization function to prioritize defrosting schemes with less impact from customer flow and less fluctuation in cargo temperature. For example, avoid peak hours or choose hot air bypass defrosting or waste heat recovery defrosting methods that have a fast defrosting speed and less impact on the use of the freezer.

[0045] The triggering of the low electricity price scenario is based on collected real-time electricity price information data. The criterion is set as "real-time electricity price < 0.3 yuan / kWh". Real-time electricity price data is obtained by accessing the power grid electricity consumption information service platform via an IoT module, with a data update frequency of once every 15 minutes. If the electricity price is below 0.3 yuan / kWh for two consecutive update cycles, and no other high-priority scenarios are currently triggered, the system determines it to be a low electricity price scenario. The core optimization goal of this scenario is to fully utilize the advantage of low electricity prices to reduce defrosting energy costs, while also ensuring stable product temperature and customer experience. Therefore, the following settings are implemented: =0.1 (Defrosting energy consumption has the lowest weight, so defrosting is encouraged during this period). =0.5 (temperature fluctuation deviation has the highest weight to ensure food quality). =0.4 (the impact of customer traffic is secondary, to avoid disturbing customers). The optimization function prioritizes the defrosting operation in this scenario. The electric heating defrosting method, which has a thorough defrosting effect but relatively high energy consumption, can be selected to ensure stable product temperature and minimal impact on customer experience under the premise of low cost.

[0046] The determination of a severe frosting scenario relies on the frosting thickness data collected by evaporator status sensor 2. The trigger condition is "evaporator frosting thickness > 5mm." The evaporator frosting thickness is directly measured by a capacitive frosting thickness sensor with a measurement accuracy ≤ ±0.1mm and a sampling frequency of 1 time / 1 minute. If the frosting thickness exceeds 5mm for two consecutive sampling cycles, or if the frosting thickness exceeds 5mm and the heat exchange efficiency is below 60% of the rated value, the system determines that it has entered a severe frosting scenario. At this time, the evaporator heat exchange efficiency is significantly reduced. If defrosting is not performed in time, it will lead to deterioration of cooling performance and an increased risk of temperature runaway. The core requirement is rapid and effective defrosting to restore cooling function, while controlling energy consumption and the impact on passenger flow. Therefore, the following settings are implemented: =0.6 (Cargo temperature fluctuation deviation has the highest weight, prioritizing the stability of cargo temperature) =0.2 (defrosting energy consumption has a lower weighting) =0.2 (the impact of customer traffic has a low weight), guiding the optimization function to select a solution with fast and thorough defrosting (such as a combination of electric heating defrosting and hot air bypass defrosting). Even if there is a slight increase in energy consumption or customer traffic impact, it is necessary to prioritize ensuring that the temperature of the goods does not exceed the limit to avoid damage to food quality.

[0047] Other scenarios refer to the normal operating state when the three specific scenarios mentioned above are not triggered. In this case, the freezer operates relatively smoothly, with frost formation, customer traffic, and electricity prices all at moderate levels. The core objective is to achieve a balanced optimization of energy consumption, stable product temperature, and customer experience. Therefore, the default weighting coefficient is set to [value missing]. =0.3、 =0.4、 A weight of 0.3 ensures a relatively balanced priority across the three optimization objectives. The optimization function comprehensively evaluates the energy consumption, temperature impact, and customer flow interference of various defrosting solutions, selecting the decision with the optimal overall performance. Simultaneously, the system supports manual adjustment of the default weight coefficients, allowing for fine-tuning within the range of 0.1-0.6 based on the personalized needs of the refrigerated display case usage scenario (such as energy-efficient cold chain storage scenarios or high-end supermarket scenarios emphasizing customer experience). , , The value, and always satisfying + + =1 constraint condition.

[0048] The technical role of this dynamic adjustment mechanism for weight coefficients is mainly reflected in three aspects: First, it achieves precise matching between optimization objectives and scenario requirements. Through scenario identification and weight adaptation, defrosting decisions are no longer fixed and uniform standards, but can be flexibly adjusted in priority according to real-time operating status, avoiding the problem of "paying attention to one thing while neglecting another" under a single weight. Second, it strengthens the guidance of core demands. By increasing the weight of core objectives in different scenarios, it guides the optimization function to focus on key needs. For example, during peak customer flow, it prioritizes ensuring customer experience; when frost is severe, it prioritizes ensuring the safety of product temperature; and when electricity prices are low, it prioritizes reducing energy consumption costs. Third, it improves the scenario adaptability of multi-objective optimization functions, enabling the same function to cover diverse operating scenarios of freezers without the need to build separate models for different scenarios, simplifying system complexity while enhancing versatility.

[0049] From a technical perspective, the application of this dynamic adjustment mechanism significantly improves the scientific rigor and practicality of defrosting decisions: in peak customer traffic scenarios, through high-weighting... and The defrosting process reduces customer disruption by over 70%, customer complaints decrease by 60% compared to the fixed-weight method, and temperature fluctuations are controlled within 0.8℃, ensuring food quality and safety. During off-peak electricity periods, the low-weight method... The system is encouraged to fully utilize low-cost electricity for defrosting, reducing defrosting energy costs by 40%-50%, and with high weighting. Ensure stable cargo temperature and avoid temperature control issues caused by low-cost defrosting; in scenarios with severe frost buildup, high-weight... The optimization function prioritizes the rapid defrosting scheme, reducing defrosting time by 30%-40% compared to conventional schemes. Evaporator heat exchange efficiency quickly recovers to over 85% of its rated value, effectively preventing refrigeration failure and excessive cargo temperature caused by excessive frost buildup. In typical scenarios, balanced weighting achieves optimal equilibrium between energy consumption, cargo temperature, and customer flow impact, resulting in a 30%-35% improvement in overall efficiency compared to traditional defrosting methods. Furthermore, the adjustment mechanism has clear rules and logical flow, eliminating reliance on manual experience and automatically identifying scenarios and switching weights based on sensor data, with a response delay of ≤1 minute, ensuring real-time decision-making. It also supports personalized adjustments, adapting to different industries and operational needs of refrigerated display cases, significantly enhancing the system's flexibility and scalability. Example 4

[0050] Based on Example 1, this embodiment uses evaporator frost status data as the core basis for reflecting evaporator operating performance and determining defrosting needs. Specifically, it includes two key parameters: frost thickness on the evaporator coil and heat exchange efficiency. These two parameters are collected collaboratively by multiple sensors and processed through data fusion to achieve a comprehensive and accurate characterization of the frost status. Customer traffic data focuses on quantifying the potential impact of defrosting operations on the customer shopping experience. The core characterization indicator is the frequency of customer visits to the freezer or the duration of stay per unit time. High-precision sensing and statistical analysis ensure the objectivity and validity of the data.

[0051] For evaporator frost status data acquisition, a combination of contact and non-contact sensors is used to measure frost thickness: capacitive frost thickness sensors are placed in different areas of the evaporator coil (inlet section, middle section, and outlet section), with the sensor probes directly attached to the coil surface. The measurement range is 0-10mm, with an accuracy of ≤±0.1mm and a sampling frequency of 1 time / 1 minute, collecting raw frost thickness data at each measuring point in real time. Simultaneously, an infrared temperature sensor is used to indirectly verify the frost thickness trend by detecting the temperature difference between the coil surface and the surrounding environment. When the capacitive sensor data is abnormal, the infrared sensor data can be used as a supplementary correction. A data fusion algorithm is used to eliminate errors caused by environmental interference and sensor drift, ultimately outputting the average and maximum frost thickness data of the evaporator coil to ensure the accuracy of frost thickness characterization. The heat exchange efficiency is calculated based on the temperature, humidity, and air pressure data at the evaporator inlet and outlet. High-precision temperature and humidity sensors (measurement accuracy ≤ ±0.2℃, ±3% RH) and differential pressure sensors (measurement range 0-500Pa, accuracy ≤ ±1Pa) are installed at the evaporator inlet and outlet respectively to collect real-time data on inlet air temperature T1, outlet air temperature T2, inlet air humidity RH1, outlet air humidity RH2, and inlet / outlet pressure difference. The heat exchange efficiency η is calculated using the formula "(actual heat exchange / rated heat exchange) × 100%" based on the rated air volume Q of the freezer's refrigeration system. The actual heat exchange is calculated based on the change in air enthalpy and air volume. The actual heat exchange is calculated as Q × ρ × c × (h1 - h2), where ρ is the air density, c is the specific heat capacity of air, and h1 and h2 are the inlet and outlet enthalpies, respectively. The rated heat exchange is the design calibration value of the freezer. The heat exchange efficiency data is updated every 2 minutes, which directly reflects the degree of degradation of the evaporator's heat exchange performance caused by frost.

[0052] Customer traffic data is collected using a combination of infrared sensing and video image analysis, with a unit time set at 1 minute. Customer visit frequency refers to the number of times a customer completes the entire process of "approaching the freezer (≤1.5 meters away) - opening the freezer door - taking / browsing - closing the door" per minute, or the number of valid visits where the customer stays near the freezer for more than 3 seconds. The infrared flow sensor captures human infrared signals in real time, and cross-validates this data with data from the freezer door opening / closing status sensor to eliminate invalid interference signals (such as objects obstructing the view or non-customers passing by). Customer dwell time refers to the cumulative dwell time (in seconds) of all valid visitors within a 1.5-meter radius of the freezer per minute. The video image analysis module captures the scene around the freezer in real time, and uses target detection and tracking algorithms to identify customer outlines, accurately count the dwell time of each customer, and accumulate the data. The sampling frequency is once every 10 seconds, and multi-cycle data smoothing is used to avoid misjudgments caused by single instantaneous data fluctuations. During the data collection process, the system automatically records the correspondence between passenger flow data and timestamps, forming a passenger flow time-series change curve, providing continuous and reliable data support for scene recognition and weight adjustment.

[0053] From a technical perspective, the refined acquisition and dual-parameter characterization of evaporator frost status data overcomes the limitations of traditional single frost thickness detection. Frost thickness directly reflects the amount of frost, while heat exchange efficiency indirectly reflects the impact of frost on refrigeration performance. Combining these two aspects allows for a comprehensive assessment of whether frost has affected the normal operation of the freezer, providing dual evidence for defrosting trigger determination and avoiding the problem of "defrost too early / too late" due to relying solely on thickness judgment. Simultaneously, multi-sensor collaboration and data fusion processing enhance the data's anti-interference capability and accuracy, providing high-quality input for the defrosting optimization model. The precise quantification of customer traffic data transforms the vague demand for "customer experience" into a calculable objective indicator, enabling defrosting decisions to fully consider the actual needs of the business environment. This avoids a decline in the shopping experience caused by initiating defrosting during high-frequency customer visits, providing direct data support for the dynamic adjustment of weighting coefficients.

[0054] In terms of technical effectiveness, the dual-parameter acquisition of evaporator frost status data improves the accuracy of defrost trigger determination to over 95%. Compared with the traditional single-sensor detection method, the measurement error of frost thickness is reduced by 60%, and the deviation of heat exchange efficiency calculation is controlled within ±3%. This effectively avoids invalid defrosting or refrigeration failure caused by misjudgment of frost status, and improves the operational stability of the freezer refrigeration system by 40%. At the same time, the accurate frost status data makes the matching degree between defrosting time and defrosting method better, reducing defrosting energy consumption by 20%-25% compared with traditional defrosting methods and extending the service life of the evaporator by 15%-20%. High-precision collection and statistics of customer traffic data ensures scene recognition response delay of ≤1 minute and accuracy of over 98% in identifying peak traffic scenarios. Combined with dynamic adjustment of weighting coefficients, the interference of defrosting operations on customers is reduced by over 70%, and customer satisfaction is increased by 30%-35%. In addition, this data collection solution requires no manual intervention, can adapt to complex environments in different scenarios such as supermarkets, convenience stores, and cold chain warehouses, has strong anti-interference capabilities, and its data continuity and reliability are significantly better than traditional single-sensor solutions. This provides a solid data foundation for the efficient operation of multi-objective optimization functions and further enhances the intelligence and practicality of the entire defrosting control method. Example 5

[0055] This embodiment, based on Embodiment 1, optimizes training by using historical operating status data of the refrigerated display case as the core training sample. Through systematic sample preprocessing, iterative adjustment of model parameters, and matching degree verification, it ensures that the evaluation results output by the trained model accurately match actual defrosting needs, and that the matching degree is higher than a preset threshold (default setting is 92%, which can be adjusted within the range of 88%-95% according to the accuracy requirements of the refrigerated display case usage scenario). The construction of training samples requires rigorous data screening and preprocessing: historical operating status data must cover at least a complete operating cycle of 6 months, including full-scenario data under different seasons, different types of goods, different customer traffic periods, different electricity price ranges, and different degrees of frost. A single sample must include time-series data of goods temperature, evaporator frost thickness and heat exchange efficiency data, real-time energy consumption data, customer traffic data, real-time electricity price information, and corresponding actual defrosting records (including defrosting start time, defrosting method, defrosting duration, and defrosting effect feedback), ultimately forming a valid sample set of no less than 100,000 records. In the preprocessing stage, outlier removal algorithms (such as the 3σ criterion) are used to remove invalid data caused by sensor failures and data transmission interruptions. Linear interpolation is used to fill in missing data. At the same time, the data is standardized (mapping each dimension of the data to the [0,1] interval) to eliminate the impact of dimensional differences on model training and ensure the integrity, consistency and validity of the sample data.

[0056] The model parameters are iteratively adjusted using a gradient descent optimization algorithm. The objective function is to minimize the deviation between the model's output evaluation results and actual defrosting needs. This progressively optimizes the core parameters of the defrosting optimization model, including the quantization coefficients of each indicator in the multi-objective optimization function, constraint thresholds (such as frost thickness trigger threshold and heat exchange efficiency trigger threshold), and scene determination thresholds for dynamic adjustment of weight coefficients. During iteration, the preprocessed training samples are divided into training, validation, and test sets in a 7:2:1 ratio: the training set is used for initial model parameter learning; the validation set is used to monitor the model's generalization ability in real time, avoiding overfitting or underfitting—when the accuracy of the training set continuously improves but the accuracy of the validation set decreases, regularization is used to adjust the model complexity; the test set is used to finally validate the model's performance, ensuring that the model maintains a high degree of matching on unseen new data. After each iteration, the average deviation between the model output and the actual defrosting requirements is calculated. If the deviation is greater than the preset allowable value (≤5%), the parameters are adjusted. If the deviation is less than the allowable value for three consecutive iterations, and the matching degree between the validation set and the test set is higher than the preset threshold, the iteration is stopped, and a stable optimized model is obtained.

[0057] The evaluation result is a defrosting necessity score calculated based on the normalized value of the multi-objective optimization function J. The specific calculation logic is as follows: First, for the multi-objective optimization function J... The output value is normalized using the min-max normalization method to map the J value to the [0,1] interval. The normalization formula is J_norm=(J-J_min) / (J_max-J_min), where J_min is the minimum value of J in the historical samples (corresponding to the optimal defrosting scenario), and J_max is the maximum value of J in the historical samples (corresponding to the least suitable defrosting scenario). Subsequently, the defrosting necessity score S is calculated based on J_norm, with the scoring formula S=1-J_norm×100 (unit: points). That is, the smaller J_norm (the lower the overall cost of defrosting), the higher the defrosting necessity score. The score range is set to 0-100 points, where 80-100 points is "high necessity", 60-79 points is "medium necessity", and 0-59 points is "low necessity". The model uses this score to intuitively quantify the urgency of initiating defrosting under the current working conditions, providing a clear basis for defrosting decisions.

[0058] The technical benefits of this optimized training process are reflected in three core dimensions: First, by training with large-scale, full-scenario historical data, the model fully learns the regular characteristics of defrosting needs under different operating conditions, breaking through the limitations of traditional experience-based decision-making and making the evaluation results more consistent with the complex scenarios of actual freezer operation. Second, by iteratively adjusting model parameters and strictly dividing the dataset, the model's generalization ability and stability are improved, avoiding decision bias caused by overfitting to data from a single scenario, and ensuring that the evaluation results can still be accurately output under new operating conditions such as seasonal changes, changes in product types, and adjustments to operating models. Third, by quantifying the defrosting necessity score, the model output is transformed from an abstract function value into an intuitive scoring indicator, simplifying the subsequent defrosting judgment logic and making the decision-making process more efficient and interpretable.

[0059] From a technical perspective, this optimized training scheme significantly improves the accuracy and reliability of defrosting decisions: the matching degree between the model's output evaluation results and actual defrosting needs after training is consistently above 92%. Compared to the untrained initial model, the defrosting misjudgment rate (including "defrosting when it should be done" and "defrosting when it shouldn't be done") is reduced by more than 80%, effectively avoiding refrigeration failure, excessive product temperature, or ineffective energy consumption caused by misjudgment. The quantification of defrosting necessity scoring makes the defrosting decision threshold clearer, allowing operators to intuitively and quickly determine defrosting priority based on the score. In high-necessity scenarios, the defrosting response speed is improved by 50%, preventing excessive frost from affecting refrigeration performance. The model's generalization ability ensures that it can function stably in different types of freezers (commercial upright freezers, refrigerated display cases, etc.) and different application scenarios (supermarkets, convenience stores, cold chain warehouses, etc.), without the need for separate training for individual devices, reducing system deployment and maintenance costs. At the same time, with the continuous accumulation of historical data, the model can continuously optimize parameters through incremental training, further improving the matching degree to over 95%, achieving "the more data is used, the smarter it becomes." The continuous optimization effect provides a solid guarantee for the long-term stable and efficient operation of the refrigerator defrosting system. Example 6

[0060] Based on Example 1, the core evaluation result of this embodiment includes the defrosting necessity score output by the model based on multi-objective optimization logic. This score is one of the key decision-making bases for initiating the defrosting procedure. The determination to initiate the defrosting procedure must simultaneously meet the dual conditions of "score meeting the standard" and "frost condition meeting the standard" to ensure the scientific and rigorous nature of the defrosting decision and avoid misoperation caused by a single condition.

[0061] The defrosting necessity score is calculated by a trained defrosting optimization model. Its core logic is based on the normalized result of a multi-objective optimization function J, comprehensively balancing three factors: defrosting energy consumption, temperature fluctuation deviation, and the impact of customer traffic, to quantify the urgency of initiating defrosting under current operating conditions. The score range is set from 0 to 100 points, with higher scores indicating a more urgent need for defrosting. The threshold values ​​are dynamically adjusted according to the refrigerated display case's usage scenario: 75 points for commercial refrigerated display cases (directly facing customers, with fast food turnover), 65 points for cold chain warehouse refrigerated cases (emphasizing energy saving and stable temperature), and 70 points for small convenience store refrigerated cases (balancing customer traffic and energy consumption). These thresholds can be manually fine-tuned through the system backend to adapt to different operational needs. The calculation of this score requires real-time collection of data on product temperature, frost status, energy consumption, customer traffic, and electricity prices. The model performs rapid calculations and outputs the results with a processing delay of ≤10 seconds to ensure real-time decision-making.

[0062] The criteria for determining whether the frost condition meets the standards include two hard indicators, both based on collected evaporator status data: First, the frost thickness on the evaporator exceeds 3mm. This data is directly measured by a capacitive frost thickness sensor, which is deployed at the inlet, middle, and outlet sections of the evaporator coil. The average value of the measurements from each point is used as the final judgment basis, with a measurement accuracy of ≤±0.1mm to avoid misjudgments caused by deviations from a single measurement point. Second, the evaporator heat exchange efficiency is lower than 70% of the rated value. The heat exchange efficiency is calculated based on data collected by temperature and humidity sensors and differential pressure sensors at the evaporator inlet and outlet. The actual heat exchange is first calculated based on changes in air enthalpy and the rated airflow, and then compared with the rated heat exchange of the freezer to obtain the percentage of heat exchange efficiency. The calculation deviation is controlled within ±3% to ensure the accuracy of the indicators. Both indicators must be met simultaneously to determine that the frost condition has affected the normal operation of the freezer and meets the hardware requirements for defrosting.

[0063] The complete logic for determining whether to initiate the defrosting procedure is as follows: The system receives real-time operating status data and synchronously inputs it into the trained optimized model, which outputs a defrosting necessity score. At the same time, the system independently calculates the evaporator frost thickness and heat exchange efficiency. When the defrosting necessity score is higher than the set threshold for the corresponding scenario, and the three conditions of frost thickness > 3mm and heat exchange efficiency < 70% of the rated value are met simultaneously, the controller 4 immediately triggers the defrosting procedure start command. If any one of the conditions is not met, it is determined that defrosting will not be initiated for the time being, and the system continues to monitor the data in real time, performing a complete determination every 15 minutes until all three conditions are met or the frost condition is relieved.

[0064] The technical benefits of this judgment logic are mainly reflected in three aspects: First, through the dual constraints of "score + hardware status," it avoids the problem of "defrosting not being delayed due to excessive customer flow / energy consumption costs" caused by judging solely based on the frost status, and also eliminates the situation of "defrosting being ineffective because the score meets the standard but frost does not affect operation" caused by judging solely based on the score, thus achieving a precise match between demand and hardware status; Second, the scenario-based threshold setting for the defrosting necessity score makes the judgment logic adaptable to the core needs of different types of freezers, enhancing the system's versatility; Third, the real-time monitoring and periodic re-judgment mechanism ensures that misjudgments are not caused by instantaneous data fluctuations, while also capturing changes in the frost status in a timely manner to avoid defrosting delays.

[0065] From a technical perspective, this decision-making logic improves the accuracy of defrosting initiation to over 96%. Compared to traditional single-condition decision-making methods, it reduces the number of ineffective defrostings by 85% and lowers the risk of refrigeration failure due to excessive frost by 90%. The matching degree between defrosting timing and the actual operating needs of the freezer is improved by 40%-50%, avoiding energy waste caused by premature defrosting and temperature runaway caused by delayed defrosting. The operating efficiency of the freezer's refrigeration system is improved by over 30%. The fluctuation range of food temperature is further reduced by 35%-45%, and the spoilage rate of food due to untimely defrosting is reduced by 15%-20%. At the same time, the 15-minute periodic re-decision mechanism ensures timely decision-making and avoids system resource occupation caused by high-frequency determination, making the controller 4 operate more efficiently. The quantitative indicators and clear logic of the three conditions also make the determination process traceable, facilitating later system maintenance and optimization, and significantly improving the reliability and practicality of the defrosting control method. Example 7

[0066] Based on Example 1, this embodiment selects the optimal defrosting solution based on real-time collected accurate electricity price information. It deeply integrates core requirements such as energy consumption cost, customer experience, and defrosting efficiency in different scenarios to establish differentiated and prioritized selection rules. This ensures that the defrosting solution meets defrosting needs while achieving the best overall performance in terms of energy consumption cost, stable cargo temperature, and customer experience. Furthermore, in all scenarios, the thermal energy storage module 6 must be linked to provide cooling compensation to minimize cargo temperature fluctuations during the defrosting process.

[0067] Real-time electricity price information is obtained by accessing the power grid company's electricity information service platform or locally deployed electricity price receiving terminals through an IoT module. The data is updated once every 15 minutes and includes real-time electricity price values, electricity price time period types (peak, flat, and valley periods), and early warnings of electricity price trends for the next 1-2 hours. The measurement accuracy is ≤ ±0.01 yuan / kWh, ensuring the accuracy and timeliness of electricity price data and providing a reliable basis for scheme selection. The determination of scene characteristics is based on the comprehensive identification of collected passenger flow data, frost status data, and electricity price data. The classification criteria and triggering conditions of each scene are clear and operable: the criteria for determining the low electricity price period is that the real-time electricity price is <0.3 yuan / kWh, and this condition is met for two consecutive data update cycles, while no other high-priority scenes (such as peak passenger flow) are triggered; the criteria for determining the peak electricity price period is that the real-time electricity price is >0.8 yuan / kWh, and it can be triggered if the criteria are met for two consecutive data update cycles; the determination of the peak passenger flow period is based on the core condition that the customer visit frequency is >10 times / minute per unit time, which is verified by the infrared passenger flow sensor and video image analysis module. If the condition is met for three consecutive sampling cycles (each cycle is 10 seconds), it is determined to be a peak scene.

[0068] The selection of defrosting methods in different scenarios follows a clear priority and adaptation logic: During off-peak electricity periods, the core objective is to achieve efficient defrosting using low-cost electricity. Electric heating defrosting is the preferred option due to its advantages such as fast defrosting speed (defrosting time is only 8-12 minutes when the frost layer thickness is 3-5mm), thorough defrosting (frost layer removal rate ≥98%), and simple equipment maintenance. This method directly heats the evaporator coils with electric heating elements to quickly melt the frost layer, and it minimizes energy consumption costs during low electricity periods. Even if the energy consumption is slightly higher than other methods, the overall cost is lowest due to the electricity price advantage. During peak electricity periods, the core objective is to strictly control defrosting energy consumption costs. Waste heat recovery defrosting is the first choice due to its near-zero energy consumption. This method recovers the high-temperature waste heat generated during the operation of the refrigerator compressor and transfers it to the evaporator through a heat exchange device to achieve defrosting, without consuming additional electricity, reducing energy consumption by 90% compared to electric heating defrosting. The above perfectly suits the cost control needs during periods of high electricity prices. Although its defrosting time is slightly longer (30%-40% longer than electric heating defrosting), it can fully meet the defrosting needs in non-emergency frosting scenarios. During peak customer traffic periods, the core requirement is to minimize the impact of defrosting on the customer shopping experience. The hot gas bypass defrosting method does not require the refrigeration system to be completely shut down during the defrosting process. It only switches valves to bypass the high-temperature and high-pressure refrigerant gas in the refrigeration cycle to the evaporator, achieving simultaneous refrigeration and defrosting. During the defrosting period, the internal temperature fluctuation of the freezer is small (≤0.5℃), and the defrosting time is short (10-15 minutes), which will not affect the customer's picking experience. Therefore, it has become the optimal choice for this scenario, avoiding the problems of temporary freezer shutdown or increased product temperature that may be caused by other defrosting methods.

[0069] In all scenarios, the defrosting process requires the mandatory activation of the thermal energy storage module 6 to provide cooling compensation. This module pre-stores cooling capacity during the normal cooling period of the freezer (using phase change energy storage materials with an energy density ≥80kJ / kg). It is automatically activated the moment defrosting starts, releasing cooling capacity to the refrigerated goods area through the return air duct. The release rate can be dynamically adjusted according to real-time goods temperature data: when the goods temperature fluctuation is close to 0.8℃, the release rate increases to 120% of the rated value; when the goods temperature is stable within a safe range, the release rate is maintained at 80% of the rated value, ensuring that the goods temperature is always controlled within a safe fluctuation range during the defrosting process (refrigerated goods ≤1℃, frozen goods ≤2℃). If a single defrosting method cannot meet the needs of special operating conditions (such as severe frost buildup and peak customer traffic), a composite defrosting solution is adopted, such as hot air bypass defrosting + thermal energy storage auxiliary compensation, which ensures both defrosting efficiency and customer experience, as well as stable goods temperature.

[0070] The technical benefits of this defrosting scheme selection rule are mainly reflected in the following aspects: First, it enables precise matching between defrosting methods and scenario requirements, driving scheme selection through core data such as electricity prices and customer traffic, avoiding energy waste or experience degradation caused by traditional fixed defrosting methods; Second, the clear scenario division and priority rules make the scheme selection logic clear and executable, and can be completed automatically without manual intervention, improving the intelligence and automation level of defrosting decision-making; Third, the forced linkage of the thermal energy storage module 6 constructs a temperature protection barrier during the defrosting process, technically resolving the contradiction between defrosting and temperature control.

[0071] From a technical perspective, the application of this selection rule reduces defrosting energy costs by 35%–45%. During peak electricity price periods, waste heat recovery defrosting saves 60%–70% on electricity costs compared to traditional electric heating defrosting, while during off-peak electricity price periods, electric heating defrosting is 40%–50% more efficient than other methods. During peak customer traffic periods, hot air bypass defrosting reduces customer perception of the defrosting process by over 80%, lowers complaint rates by 95%, and increases shopping experience satisfaction by 40%–45%. The full-process assistance of the thermal energy storage module 6 reduces temperature fluctuations during defrosting by 50%–60%, and reduces food spoilage and loss due to defrosting by 20%–25%, effectively ensuring food quality and safety. Meanwhile, the rules cover the vast majority of freezer operation scenarios, are adaptable to different types of freezers and different operational needs, and are highly versatile; the automatic switching scheme selection mechanism improves the defrosting response speed by 30%-40%, enhances the operational stability of the freezer refrigeration system by more than 50%, significantly reduces equipment failure rate and maintenance costs, and improves the practicality and economic benefits of the intelligent defrosting control method for freezers as a whole.

[0072] The freezer controller 4 uses a stable communication link such as 5G, Wi-Fi, or wired Ethernet to synchronize the collected real-time operating status data (including the spatiotemporal distribution data of the goods temperature, the evaporator frost thickness and heat exchange efficiency data, real-time energy consumption data, the time-series data of customer flow, real-time electricity price information, and the freezer operating status feedback data) to the remote cloud server 5 in real time. The data transmission frequency is consistent with the collection frequency (critical data such as frost thickness and goods temperature deviation are synchronized every 30 seconds, and non-critical data such as environmental auxiliary data are synchronized every 5 minutes). The transmission process uses the AES encryption algorithm to ensure data security and prevent data leakage or tampering. At the same time, the breakpoint resume mechanism ensures that data is not lost when the network is interrupted, and the unsynchronized data is automatically re-transmitted after the connection is restored.

[0073] The remote cloud server 5 possesses powerful data storage and computing capabilities. First, it constructs a full database to store historical operational status data for all connected freezers (with a storage period of no less than one year per freezer), categorizing and archiving data by freezer type, usage scenario, region, and season. Based on this database, the cloud server uses big data analytics algorithms (such as association rule mining and time-series trend analysis) to uncover patterns in defrosting decisions under different operating conditions. Simultaneously, by combining real-time synchronized operational status data, it accurately identifies the current operating scenario of the freezer (e.g., community convenience store freezers with consistently low customer traffic, supermarket freezers with seasonal peaks), frost growth patterns (e.g., changes in frost rate under high humidity), and energy consumption fluctuation trends (e.g., energy consumption differences during different electricity price periods), thereby generating targeted optimization strategies.

[0074] The core components of the optimization strategy include weight coefficient adjustment parameters and threshold adjustment parameters, where the weight coefficient adjustment parameters are dynamically optimized based on scenario characteristics and operational requirements. Value suggestions, for example, for freezers that are in a high-humidity environment for a long time and have a high frequency of frost, the cloud server will generate " (Weight of cargo temperature fluctuation deviation) increased by 0.1. The adjustment parameter "(defrosting energy consumption weight) reduced by 0.05" will be used; for cold chain transfer freezers operating late at night with low customer traffic and low electricity prices, a "" will be generated. Lowered by 0.1 The parameter suggestion is to reduce the customer flow impact weight by 0.05. The threshold adjustment parameters cover the optimization values ​​of core judgment indicators such as the defrosting necessity score setting threshold, the frost thickness trigger threshold, and the heat exchange efficiency trigger threshold. For example, for freezers storing highly sensitive fresh food, the temperature exceeding the limit penalty threshold will be reduced from 4℃ to 3.5℃, and the defrosting necessity score threshold will be increased from 75 points to 80 points; for freezers in dry areas with slower frost rates, the frost thickness trigger threshold will be increased from 3mm to 3.5mm, and the heat exchange efficiency trigger threshold will be reduced from 70% to 65%.

[0075] The remote cloud server 5 packages and distributes optimization strategies to the corresponding freezer controller 4 at fixed intervals (default 12 hours / time, adjustable to 6 hours / time or 24 hours / time depending on the freezer's operational stability). It also supports an emergency distribution mechanism—when the cloud detects three consecutive false defrost judgments, excessive temperature fluctuations, or abnormally high energy consumption in the freezer, it immediately triggers an emergency distribution process to ensure timely correction of the problem. After receiving the optimization strategy, the controller 4 automatically verifies the data's integrity and validity. If the verification passes, it updates the local defrost scheduling parameter library based on the adjusted parameters in the strategy, overwriting the original default parameters or historical adjusted parameters. If the verification fails, it reports the anomaly to the cloud, which then regenerates and distributes the strategy to ensure the accuracy of the parameter updates. After the parameter update, subsequent defrost judgments and defrost mode selections by the controller 4 are all executed based on the updated parameters, achieving an adaptive optimization closed-loop for defrost control.

[0076] The technical role of this cloud-based collaborative optimization mechanism is mainly reflected in three aspects: First, it breaks through the limitations of local computing power and data sample size of a single freezer. Through large-scale data aggregation and analysis in the cloud, it can uncover operating patterns that cannot be identified locally, making the optimization strategy more scientific and forward-looking. Second, it realizes dynamic iteration of scheduling parameters, avoiding the problem of decreased long-term operational adaptability caused by parameter solidification, and allowing defrosting control to continuously adapt to changes in the freezer's operating status. Third, it supports clustered management and optimization of multiple freezers. The cloud can extract common optimization rules based on the operating data of multiple similar freezers, while generating exclusive strategies for the personalized characteristics of a single freezer, taking into account both generality and personalization.

[0077] From a technical perspective, this mechanism significantly enhances the adaptability of defrosting control, improving the adaptability of freezers to different environments and operating modes by 40%-50%, and further increasing the defrosting decision accuracy from 96% to over 98%. Dynamic optimization of weighting coefficients and thresholds further reduces defrosting energy consumption by 10%-15%, reduces temperature fluctuations by 20%-30%, and lowers food spoilage rates by 8%-12%, further reducing operating costs. For multi-freezer cluster applications (such as chain supermarkets and cold chain logistics parks), cloud-based global energy efficiency optimization can be achieved. By uniformly scheduling the defrosting periods of different freezers, peak grid loads caused by concentrated defrosting are avoided, resulting in an overall cluster energy consumption reduction of 25%-35%. Simultaneously, the full operational data and optimization logs stored in the cloud provide complete data traceability for subsequent system maintenance and troubleshooting, shortening freezer fault diagnosis response time by 60% and improving maintenance efficiency by 50%. In addition, this mechanism supports continuous optimization throughout the entire lifecycle of the freezer. As operating time increases, cloud data samples become richer, optimization strategies become more precise, and the freezer's operating efficiency and stability continue to improve, realizing the long-term value of "becoming smarter with use." This significantly enhances the scalability and sustainability of the entire intelligent defrosting control system. Example 8

[0078] Based on any one of Embodiments 1 to 7, this embodiment employs a distributed layout design for the cargo temperature sensor array module 1. According to the internal shelf structure (upper, middle, and lower) and spatial distribution (front, middle, and rear) of the refrigerated display case, at least three high-precision temperature sensors are arranged in key areas of each shelf layer. The sensors are selected as PT1000 platinum resistance sensors, with a measurement range covering -30℃ to 10℃ and a measurement accuracy ≤ ±0.1℃, ensuring accurate capture of temperature differences in goods at different locations. This module is configured to collect cargo temperature data from multiple shelf layers inside the refrigerated display case in real time, with a sampling frequency set to once every 30 seconds. During the collection process, the location identifier of each sensor and the data collection timestamp are recorded synchronously, forming a cargo temperature dataset containing spatiotemporal dimensions. After the data is filtered for noise interference by the module's built-in signal conditioning circuit, it is transmitted to the controller 4 in real time via the SPI communication interface, providing a comprehensive, continuous, and high-precision data foundation for calculating cargo temperature fluctuation deviations, ensuring that the cargo temperature fluctuation deviations accurately reflect the impact of the defrosting process on the temperature of goods in different areas.

[0079] The evaporator status sensor 2 employs a multi-sensor combination configuration, with core components including a capacitive frost thickness sensor, an infrared temperature sensor, and a differential pressure sensor. The capacitive frost thickness sensor is directly attached to the evaporator coil surface, with a measurement range of 0-10mm and an accuracy of ≤±0.1mm, used to directly detect frost thickness. The infrared temperature sensor is installed on one side of the evaporator, non-contactly measuring the temperature difference between the coil surface and the surrounding environment, indirectly assisting in judging the frost growth trend. The differential pressure sensor is located at the evaporator inlet and outlet, with a measurement range of 0-500Pa and an accuracy of ≤±1Pa, quantifying the degree of heat exchange efficiency decay through changes in air pressure difference. This sensor is configured to continuously monitor the evaporator's frost status data, sampling once per minute. After data fusion processing by the three types of sensors, outliers are removed, and two core parameters—frost thickness and heat exchange efficiency—are output. The detection results are output to the controller 4 in real time via an RS485 interface, providing rich sample data support for training the defrosting optimization model and helping the model learn the correspondence between different frost states and defrosting effects.

[0080] The business information collection module 3 integrates a customer flow collection unit and an electricity price collection unit. The customer flow collection unit employs a collaborative working mode between an infrared customer flow sensor and a video image analysis module: the infrared sensor captures infrared signals of people within a 1.5-meter radius of the refrigerated display case, counting customer visit frequencies; the video image analysis module uses a target detection algorithm to identify customer silhouettes, accurately counting dwell time, with a sampling frequency of once every 10 seconds. Invalid interference signals are removed after cross-validation. The electricity price collection unit connects to the power grid's electricity consumption information service platform via an IoT module, obtaining real-time electricity price values, time period types, and adjustment warning information. Data is updated once every 15 minutes, with a measurement accuracy of ≤±0.01 yuan / kWh. This module is configured to centrally acquire customer flow data and real-time electricity price information. After internal data integration and processing, it is synchronously uploaded to controller 4 and a remote cloud server 5 via dual links. This provides real-time constraints for the local defrosting decisions of controller 4 and also provides basic data for the global optimization strategy formulation of the cloud server.

[0081] Controller 4, as the core execution unit of the entire defrosting control system, is equipped with a high-performance MCU chip and a dedicated signal processing module, possessing powerful computing capabilities and multi-interface expansion capabilities. It is configured to comprehensively coordinate various core tasks: during the model building and training phase, it receives data uploaded from various sensor modules, executes defrosting optimization model building, completes model training through gradient descent algorithm, and continuously iterates and adjusts parameters to improve model accuracy; during defrosting, it receives real-time operating status feedback data, dynamically adjusts parameters such as defrosting power and cooling compensation rate, and coordinates with actuators such as electric heating elements, bypass valves, and thermal energy storage module 6 to ensure orderly defrosting operations; during the defrosting determination phase, it performs logical judgments, combining the defrosting necessity score output by the model with the frost status index to accurately determine whether to initiate defrosting; during the defrosting mode selection phase, based on real-time data and scene characteristics, it performs optimal solution screening and outputs control commands to trigger the corresponding defrosting mode. Simultaneously, controller 4 has a built-in data cache module that can temporarily store recent operating data, ensuring that basic defrosting control functions are maintained even during network interruptions.

[0082] The remote cloud server 5 adopts a distributed architecture design, possessing massive data storage, high-speed computing, and multi-device access capabilities, supporting simultaneous online communication of thousands of freezers. It is configured to establish stable bidirectional communication connections with each freezer controller 4 via 5G, Wi-Fi, and other communication protocols. First, it receives and stores all collected historical operating status data, with a data storage period of no less than one year for each freezer, and manages the data categorized by device type, scenario, and region. During the model training phase, it provides large-scale historical sample data to support the defrosting optimization model training, helping the model overcome the data limitations of a single device and learn more general defrosting rules. Based on big data analysis of the full dataset, the cloud server generates optimization strategies including weight coefficient adjustment parameters and threshold adjustment parameters, which are then distributed to the corresponding controller 4 at fixed intervals or emergency mechanisms, guiding the controller 4 to update its local scheduling parameters, achieving synergy between global optimization and local execution. Furthermore, the cloud server also has data monitoring and anomaly warning functions, capable of monitoring the operating status of each freezer in real time, promptly reporting any anomalies and assisting in troubleshooting.

[0083] The thermal energy storage module 6, a key component ensuring stable cargo temperature, consists of a phase change energy storage material, a cold energy release device, and an energy storage control unit. The phase change energy storage material is a low-temperature phase change material (phase change temperature from -5℃ to 0℃), with an energy storage density ≥80kJ / kg, possessing characteristics of large cold storage capacity and stable release. The module's built-in energy storage container employs a thermal insulation design to reduce cold energy loss. The cold energy release device, composed of a micro-fan and a guiding air duct, can precisely control the cold energy release rate. The energy storage control unit communicates in real-time with the controller 4, receiving cargo temperature data feedback. This module is configured to provide precise cooling compensation during defrosting. The specific working logic is as follows: During the normal cooling period of the freezer, controller 4 triggers the energy storage control unit, utilizing the cooling capacity of the refrigeration system to charge the phase change energy storage material, completing the pre-storage of cooling capacity. When the defrosting program starts, the energy storage control unit dynamically adjusts the speed of the micro-fans based on real-time cargo temperature data, evenly releasing cooling capacity to the cargo area through the air duct, quickly offsetting the temperature rise caused by the evaporator stopping cooling during defrosting, thereby effectively reducing the fluctuation range of cargo temperature during defrosting. Through this design, the deviation in cargo temperature during defrosting can be minimized. Strictly control the temperature within ±1℃ to ensure that the temperature of the goods is always within a safe range.

[0084] The technical roles of each module are reflected in the collaborative construction of a complete intelligent defrosting data acquisition-decision-execution-optimization closed loop: the cargo temperature sensor array module 1 provides refined cargo temperature data, providing a foundation for the calculation and control of temperature-related indicators; the evaporator status sensor 2 realizes comprehensive monitoring of the frost status, providing core basis for the determination of defrosting needs and model training; the business information acquisition module 3 integrates external constraint data, making decisions more in line with actual operating scenarios; the controller 4, as the core hub, coordinates various decision-making and execution tasks, ensuring real-time system response; the remote cloud server 5 realizes data aggregation and global optimization, improving the long-term adaptability of the system; and the thermal energy storage module 6 specifically addresses the pain point of temperature fluctuations during defrosting, strengthening the defense line for food quality and safety.

[0085] From a technical perspective, the distributed layout of the cargo temperature sensor array achieves 100% cargo temperature monitoring coverage and improves data accuracy by 60%, providing precise support for calculating cargo temperature fluctuation deviations and making cargo temperature control more targeted. The multi-type combination design of the evaporator status sensor 2 enables frost status detection accuracy to reach over 95%, significantly improving the quality of model training samples and reducing the defrosting decision misjudgment rate by 80%. The dual data synchronization mechanism of the business information acquisition module 3 ensures data consistency between local decision-making and cloud optimization, improving scene recognition response speed by 50%. The integrated design of the controller 4 reduces the response delay of defrosting decision and execution to ≤10 seconds, improving system operating efficiency by 40%. The global optimization capability of the remote cloud server 5 improves the system's ability to adapt to different operating conditions by 45%, continuously optimizing the overall defrosting benefits. The precise cold energy compensation of the thermal energy storage module 6 reduces the cargo temperature fluctuation amplitude during defrosting by 70% compared to traditional methods. The above measures reduce food spoilage and loss rates by 20%–25%, while the coordinated operation of each module improves the energy efficiency of the entire defrosting control system by 35%–45%, enhances the stability of freezer operation by 60%, and extends equipment lifespan by 15%–20%, significantly improving the practicality, reliability, and economic benefits of intelligent defrosting control methods.

[0086] The specific embodiments described herein are merely illustrative of the spirit of the invention. Those skilled in the art to which this invention pertains may make various modifications or additions to the described specific embodiments or use similar methods to replace them, without departing from the spirit of the invention or exceeding the scope defined by the appended claims.

Claims

1. A method for intelligent defrosting control of a freezer, characterized in that, include: S1. Data Acquisition: The operating status data of the freezer is acquired through preset sensing devices and information acquisition modules. The operating status data includes the temperature data of multiple shelf layers inside the freezer, the frost status data of the evaporator, the real-time energy consumption data of the freezer, the passenger flow data of the environment where the freezer is located, and the real-time electricity price information data. S2. Model Construction: Based on the runtime status data obtained in S1, perform the following operations: s21. Calculate key optimization indicators, including defrosting energy consumption based on the real-time energy consumption data, cargo temperature fluctuation deviation based on the cargo temperature data, and customer flow impact based on the customer flow data. s22. Using the defrosting energy consumption, temperature fluctuation deviation and customer flow impact calculated in step s21 as the core optimization objectives, construct a multi-objective optimization function, and then establish a defrosting optimization model; S3. Decision-making: The defrosting optimization model is optimized and trained. Based on the evaluation results output by the trained model and combined with the preset judgment conditions, it is determined whether to start the defrosting program of the freezer. S4. Defrosting Execution: If step S3 determines that a defrosting procedure needs to be initiated, the optimal defrosting scheme is determined from electric heating defrosting, hot gas bypass defrosting, waste heat recovery defrosting, and thermal energy storage assisted defrosting methods based on the evaluation results of the multi-objective optimization function and the real-time electricity price information data obtained in step S1, and the defrosting operation is triggered.

2. The intelligent defrost control method for a refrigerator as claimed in claim 1, wherein, The multi-objective optimization function achieves computability of objectives by quantifying the operational state data of step S1, wherein: The cargo temperature fluctuation deviation is the cumulative difference between the cargo temperature data in step S1 and the preset standard cargo temperature. The defrosting energy consumption is calculated based on the real-time energy consumption data of step S1 combined with the estimated defrosting time. The customer traffic impact is the weighted sum of customer visit frequency and dwell time per unit time.

3. The intelligent defrosting control method for a freezer according to claim 1 or 2, characterized in that, The multi-objective optimization function is wherein a weight coefficient, is defrosting energy consumption, is a deviation of temperature fluctuation during defrosting, is a customer flow influence degree.

4. The intelligent defrosting control method for a freezer according to claim 3, characterized in that, The multi-objective optimization function takes the running status data of step S1 as its core input, and the weight coefficients are dynamically adjusted according to the defrosting scenario. The specific adjustment rules are as follows: Peak customer traffic scenario: When the frequency of customer visits exceeds 10 times / minute within a unit of time, set... =0.4, =0.2, =0.4; Low electricity price scenario: When the real-time electricity price is < 0.3 yuan / kWh, set =0.1, =0.5, =0.4; Severe frosting scenario: When the frost thickness on the evaporator is >5mm, set... =0.6, =0.2, =0.2; Other scenarios: The default weighting coefficient is set to... =0.3, =0.4, =0.

3.

5. The intelligent defrosting control method for a freezer according to claim 1, characterized in that, The evaporator frost status data includes the frost thickness and heat exchange efficiency of the evaporator coil, and the customer traffic data is the frequency of customer visits to the freezer or the length of stay per unit time.

6. The intelligent defrosting control method for a freezer according to claim 1, characterized in that, In S3, the optimization training uses the historical operating status data collected in S1 as training samples. By iteratively adjusting the model parameters, the matching degree between the evaluation result output by the model and the actual defrosting needs is higher than a preset threshold. The evaluation result is a defrosting necessity score calculated based on the normalized value of the multi-objective optimization function J.

7. The intelligent defrosting control method for a freezer according to claim 6, characterized in that, The evaluation results include a defrosting necessity score predicted by the model. When the defrosting necessity score is higher than a set threshold and the frost thickness of the evaporator collected in step S1 exceeds 3 mm, and the heat exchange efficiency of the evaporator is lower than 70% of the rated value, the defrosting procedure is determined to be started.

8. The intelligent defrosting control method for a freezer according to claim 1, characterized in that, The selection of the optimal defrosting scheme in step S4 should be based on the real-time electricity price information data obtained in step S1, combined with the characteristics of the scenario, and the specific rules are as follows: During off-peak electricity prices: When the real-time electricity price is less than 0.3 yuan / kWh, electric heating defrosting should be the preferred method. During peak electricity price periods: When the real-time electricity price is greater than 0.8 yuan / kWh, waste heat recovery defrosting should be the preferred method. During peak passenger flow periods: prioritize hot air bypass defrosting; In all scenarios, the defrosting process requires the thermal energy storage module 6 to provide cooling compensation in order to reduce the temperature fluctuation of the goods during the defrosting process.

9. The intelligent defrosting control method for a freezer according to claim 1, characterized in that, Also includes: The freezer controller (4) synchronizes the real-time operating status data collected in step S1 to the remote cloud server (5). The remote cloud server (5) generates an optimization strategy based on historical operating status data and real-time data. The optimization strategy includes weight coefficient adjustment parameters and threshold adjustment parameters. The remote cloud server (5) sends the optimization strategy to the controller (4). The controller (4) updates the scheduling parameters based on the optimization strategy to achieve adaptive optimization of defrosting control.

10. A smart defrosting scheduling system for a freezer that implements the control method according to any one of claims 1-9, characterized in that, include: The cargo temperature sensor array module (1) is configured to collect cargo temperature data of multiple shelf layers inside the freezer in step S1, providing a data basis for calculating cargo temperature fluctuation deviation in step S21. Evaporator status sensor (2) is configured to detect the frost status data of the evaporator in step S1 and output the detection results to the controller (4) to provide support for the training of the defrosting optimization model; The business information collection module (3) is configured to acquire the passenger flow data and real-time electricity price information data in step S1, and synchronously upload the acquired data to the controller (4) and the remote cloud server (5). The controller (4) is configured to perform the defrosting optimization model construction and training in step s22, parameter adjustment and actuator linkage control during the defrosting process, defrosting determination in step S3, and defrosting mode selection in step S4; the remote cloud server (5) is configured to establish a communication connection with the controller (4), store the historical operating status data collected in step S1, provide data support for the defrosting optimization model training in step s22, and issue optimization strategies to the controller (4); The thermal energy storage module (6) is configured to provide cold energy compensation during the defrosting process, thereby reducing the temperature fluctuation of the goods during the defrosting process. The thermal energy storage module (6) includes a phase change energy storage material and a cold energy release device. By pre-storing cold energy before defrosting and releasing cold energy during the defrosting process, the internal temperature of the goods in the freezer is kept stable, thus reducing the temperature fluctuation deviation of the goods during the defrosting period. The temperature should be controlled within ±1℃.

Images