Dynamic optimization system for energy consumption of refrigeration house based on digital twinning

Abstract

A dynamic optimization system for energy consumption of a refrigeration house based on digital twinning is characterized by comprising a data acquisition module used for acquiring basic structure data of the refrigeration house, technical parameters of a refrigeration system, real-time operation data and historical operation data, preprocessing the data and then outputting a standardized multi-dimensional real-time data stream; the model construction module is used for constructing a 3D geometric model, a thermodynamic transfer model and a refrigeration system mathematical model according to the multi-dimensional real-time data flow, performing machine learning calibration on model parameters through historical operation data, and performing fusion to construct a refrigeration house digital twin model; the prediction analysis module is used for predicting future energy consumption demand and load change according to the refrigeration house digital twin model and the real-time operation data, and outputting an energy consumption prediction result and a load analysis report; a strategy generation module; an execution feedback module; and a learning optimization module. Overall energy consumption of the refrigeration house is reduced, energy utilization efficiency is remarkably improved, and goods storage safety is guaranteed.

Description

Technical Field

[0001] This invention relates to the field of cold storage optimization technology, specifically a dynamic energy consumption optimization system for cold storage based on digital twins. Background Technology

[0002] Traditional cold storage energy consumption control solutions mostly adopt a static control mode with fixed parameters and manual adjustment. They rely on preset equipment operating parameters and lack dynamic adaptation to real-time operating scenarios. The data collection dimensions are single and fragmented, resulting in extremely low data value utilization and an inability to provide comprehensive support for subsequent control. In the face of environmental temperature fluctuations and changes in the amount of goods stored, overcooling or undercooling problems often occur, causing unnecessary energy waste.

[0003] Traditional solutions lack systematic model support and the concept of digital twin models at the model construction and application levels. They lack thermodynamic transfer models to simulate cold distribution, and the mathematical models of refrigeration systems are mostly static empirical formulas, failing to quantify the dynamic relationship between equipment operating parameters and energy consumption and cooling effect. Furthermore, traditional solutions do not involve model calibration; as equipment ages and the environment changes, the deviation between the model and actual operating conditions continues to widen, leading to a significant decrease in the accuracy of model-based judgments. This, in turn, causes control decisions to deviate from actual needs, making it difficult to achieve a balance between energy consumption and cooling effect.

[0004] In traditional cold storage, energy consumption prediction and strategy generation rely heavily on manual experience and judgment, lacking scientific algorithmic support and data-driven capabilities. Strategy generation often employs a one-size-fits-all approach with no multi-objective optimization logic, making it difficult to balance multiple objectives such as energy consumption, cost, and equipment lifespan, thus limiting optimization effectiveness.

[0005] In terms of execution feedback and continuous optimization, traditional solutions have weak closed-loop control capabilities and lack self-learning mechanisms. Traditional solutions lack dedicated learning and optimization mechanisms, leaving the system in a rigid, experience-driven state for extended periods. As the cold storage facility ages and the external environment changes, energy consumption control effectiveness gradually declines, making it difficult to achieve long-term stable energy consumption optimization. Summary of the Invention

[0006] To solve the above-mentioned technical problems, the technical solution adopted by the present invention is: a dynamic energy consumption optimization system for cold storage based on digital twins, characterized in that it includes: The data acquisition module is used to collect basic structural data of cold storage, technical parameters of refrigeration system, real-time operation data and historical operation data, and outputs standardized multi-dimensional real-time data streams after data preprocessing; The model building module is used to build 3D geometric models, thermodynamic transfer models, and refrigeration system mathematical models based on multi-dimensional real-time data streams. After machine learning calibration of the model parameters using historical operating data, the models are integrated to build a digital twin model of the cold storage. The predictive analysis module is used to predict future energy consumption demand and load changes based on the cold storage digital twin model and real-time operation data, and output energy consumption prediction results and load analysis reports. The strategy generation module is used to establish a multi-objective optimization mathematical model based on energy consumption prediction results and load analysis, and generate dynamic optimization control strategies. The execution feedback module is used to convert the dynamic optimization control strategy into equipment control commands, monitor the control effect data in real time, dynamically adjust the control parameters, and output the optimized cold storage operation status and energy consumption data. The learning optimization module is used to compare the optimized cold storage operation status and energy consumption data with those before optimization, and to learn and optimize the model.

[0007] The real-time and historical operating data include: real-time operating data includes multi-point temperature data inside the cold storage, ambient temperature data, equipment surface temperature data, relative humidity and absolute humidity data, pressure data including high and low pressure sides of the refrigeration system and static pressure data of the fan, and electrical parameters including current, voltage, power factor and frequency of each piece of equipment; historical operating data includes recent continuous operating data including temperature, humidity, pressure and flow records, energy consumption data under different operating conditions, equipment performance degradation data, fault and maintenance records.

[0008] The 3D geometric model is constructed as follows: a precise geometric model of the cold storage is built based on the basic structural data of the cold storage, including the external wall structure, internal compartments, insulation layer thickness, door and window positions, and shelf layout; at the same time, a coordinate system is established.

[0009] The thermodynamic transfer model is constructed as follows: Based on the 3D geometric model, a mathematical model describing the heat and mass transfer process in the cold storage is established: the energy conservation equation describes the temperature field distribution, the mass conservation equation describes the humidity field change, and the momentum conservation equation describes the air flow; the temperature transfer process is described by the three-dimensional unsteady heat transfer equation; and the coupling effect of phase change heat transfer, convective heat transfer and radiative heat transfer is considered to establish a set of boundary condition equations.

[0010] The mathematical model of the refrigeration system is constructed as follows: dynamic characteristic models are established for the compressor, condenser, evaporator, and expansion valve of the refrigeration system respectively: the compressor model uses a polynomial fitting method to describe the relationship between its cooling capacity and power consumption and the changes in evaporation temperature, condensation temperature, and rotational speed; the condenser and evaporator use the moving boundary method to establish dynamic heat transfer models, considering the two-phase flow and phase change process of the refrigerant; the expansion valve uses an isenthalpic expansion model; the models of each component are coupled and connected through the mass flow rate and enthalpy value of the refrigerant.

[0011] The method for predicting future energy consumption demand is as follows: For short-term demand, based on the thermodynamic transfer characteristics of the digital twin model, the dynamic heat load distribution of the cold storage is calculated in real time by solving the three-dimensional unsteady heat transfer equation; for medium- and long-term energy management needs, a time series prediction model is trained using historical cold load data, environmental temperature and humidity change patterns, cargo storage plans, and electricity price fluctuation cycles as input features, and outputs the peak and valley values ​​of cold load and the corresponding energy consumption prediction values ​​for different future periods.

[0012] The load analysis report is generated as follows: an in-depth analysis of the energy consumption distribution and operating status of the cold storage is conducted: the energy consumption ratio of each subsystem is broken down using a digital twin model to identify abnormal energy consumption patterns; the energy efficiency ratio change trend of each core device is calculated to determine whether the device is in a high-efficiency operating range, and finally a load analysis report is generated that includes energy consumption prediction results, load change curves, anomaly identification reports, and energy efficiency analysis conclusions.

[0013] The method for generating dynamic optimization control strategies is as follows: a multi-objective optimization mathematical model is established, with the minimum total energy consumption of the cold storage as the primary objective, while the temperature fluctuation range and cumulative equipment operating time in the cold storage are used as constraints, and the weight of each objective is quantified; the optimal solution is searched through algorithm iteration using compressor operating frequency, fan speed, valve opening degree, and defrosting time as optimization variables, and finally, a time-segmented equipment operation strategy is generated.

[0014] The method for dynamically adjusting control parameters is as follows: when monitoring data shows a deviation between the actual state and the expected target, the system automatically activates the feedback adjustment mechanism, analyzes the cause of the deviation, simulates the effects of different adjustment schemes based on the digital twin model, and dynamically adjusts the control parameters; at the same time, the feedback data is transmitted to the digital twin model, the equipment performance parameters in the model are updated, and the optimized cold storage operation status and energy consumption data are output.

[0015] The method for learning and optimizing the model is as follows: establish an optimization effect evaluation system, analyze the actual effect and applicable conditions of the optimization strategy by comparing the data before and after optimization and the predicted data; use reinforcement learning to iteratively upgrade the optimization strategy, and at the same time establish an optimization case knowledge base to classify and store optimization strategies with significant effects, so that mature strategies can be quickly called when similar scenarios are encountered; and regularly adjust the core parameters of the optimization algorithm in response to seasonal changes and changes in cargo storage patterns.

[0016] Combining all the above technical solutions, the positive effects of this invention are as follows: 1. This invention comprehensively integrates multi-dimensional data through a data acquisition module and constructs a high-precision digital twin model through a model building module, accurately simulating the cold load transfer and equipment operation patterns. By combining the model with real-time data, changes in cold load can be predicted in advance; based on a multi-objective optimization model, dynamic strategies are formulated in conjunction with peak and off-peak electricity prices and cargo characteristics, such as storing cold during off-peak hours and reducing load during peak hours, effectively reducing ineffective energy consumption. This can reduce the overall energy consumption of cold storage, significantly improve energy utilization efficiency, and reduce operating costs for enterprises.

[0017] 2. This invention transforms optimization strategies into precise control commands, driving equipment in real time through a PLC / DCS system. Simultaneously, it monitors operational deviations and dynamically adjusts parameters. Combined with the real-time synchronization characteristics of a digital twin model, it can control the temperature fluctuation range of cold storage. The predictive analysis module can identify abnormal energy consumption and inefficient equipment operation in advance, such as a drop in compressor COP or valve leakage, providing timely warnings and mitigating risks through strategy optimization. This avoids temperature runaway due to equipment failure and ensures the safety of stored goods.

[0018] 3. This invention significantly improves cold storage management efficiency, reduces reliance on manual labor and operational errors, and achieves full-process automation: the data acquisition module automatically acquires multi-dimensional data and processes it in a standardized manner, eliminating the need for manual data entry; the model building and predictive analysis module replaces manual estimation, outputting accurate load analysis and energy consumption predictions; the strategy generation module automatically generates optimization schemes, and the execution feedback module automatically implements control commands. Managers only need to view results and issue commands through the system, improving management efficiency while avoiding human error and ensuring the scientific nature of decision-making.

[0019] 4. The learning optimization module of this invention analyzes the effect of the strategy by comparing data before and after optimization, and uses reinforcement learning algorithms to iteratively optimize the model and strategy. It can also store successful cases to form a knowledge base. When seasonal changes cause changes in environmental temperature and humidity, adjustments are made to the type of goods stored, or equipment ages and experiences performance degradation, the system can automatically adjust algorithm parameters and control strategies without manual readjustment. With long-term use, the system's optimization accuracy continuously improves, ensuring that the cold storage maintains a highly efficient energy consumption control state throughout its entire life cycle. Attached Figure Description

[0020] The present invention will be further described with reference to the accompanying drawings, but the embodiments in the drawings do not constitute any limitation on the present invention. For those skilled in the art, other drawings can be obtained based on the following drawings without creative effort.

[0021] Figure 1 This is a system framework diagram of the system of the present invention. Detailed Implementation

[0022] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0023] See Figure 1 As shown, this invention proposes a dynamic energy consumption optimization system for cold storage based on digital twins, including a data acquisition module, a model building module, a predictive analysis module, a strategy generation module, an execution feedback module, and a learning optimization module.

[0024] In a more specific application of this invention, the data acquisition module collects basic structural data of the cold storage, technical parameters of the refrigeration system, real-time operating data, and historical operating data. After data preprocessing, it outputs a standardized multi-dimensional real-time data stream. For the basic structural data of the cold storage, laser scanning combined with BIM modeling technology is used to collect the building dimensions of the cold storage, the thickness of the wall insulation layer, the layout of the cold storage room, the location of doors and windows, and the sealing performance parameters. At the same time, the spatial information of pipeline routing, pipe diameter specifications, and equipment installation coordinates is recorded to form a structured basic database, providing accurate physical parameter support for the subsequent construction of 3D geometric models.

[0025] The technical parameters of the refrigeration system are collected by combining equipment nameplate identification with digital factory documentation. The technical parameters of the compressor, condenser, and evaporator are directly read through the Internet of Things interface, including rated power, rated cooling capacity, working pressure range, and efficiency curves. For the control components of expansion valves and solenoid valves, their model specifications, action response time, and flow characteristic curves are recorded. For fans and pumps, the performance parameters of rated speed, air volume, and head are collected. All parameters are manually verified and then entered into the system to form a benchmark performance database for the refrigeration system.

[0026] Historical operational data is collected and analyzed retrospectively through a long-term storage mechanism. The system automatically archives recent continuous operational data, summarizing the change curves of temperature, humidity, pressure, and flow rate. For different operating conditions, such as full load / no load, day / night, and seasonal transitions, corresponding energy consumption data is individually marked and stored. Performance degradation indicators, such as the deviation rate between the actual cooling capacity and the rated value of the compressor, and changes in heat exchanger cleanliness, are recorded through the equipment management module. Simultaneously, the fault alarm system and maintenance records are integrated, including the fault occurrence time, type, handling measures, and post-maintenance performance recovery data. Historical data is stored in a distributed database, providing long-term data support for model calibration and strategy optimization.

[0027] The collected data undergoes preprocessing, employing Kalman filtering to remove sensor noise and interpolating missing data. Data units and formats are standardized, and timestamp alignment enables spatiotemporal synchronization of multi-source data. A data quality assessment mechanism is established to automatically flag outliers and trigger sensor self-checks. The final output is a standardized, multi-dimensional real-time data stream containing updated key operating parameters and an integrated structured historical data index, providing high-quality data input for digital twin model construction and subsequent analysis.

[0028] In the model building module, a 3D geometric model, a thermodynamic transfer model, and a refrigeration system mathematical model are constructed based on multi-dimensional real-time data streams. After machine learning calibration of the model parameters using historical operating data, a digital twin model of the cold storage is built by merging these components. This model can be constructed by using CAD software or 3D modeling tools based on the cold storage's basic structural data to create an accurate geometric model of the cold storage, including the external wall structure, internal compartments, insulation layer thickness, door and window positions, and shelf layout. At the same time, a coordinate system is established to lay the foundation for subsequent mesh generation and numerical calculations.

[0029] Based on the geometric model, a mathematical model describing the heat and mass transfer process within the cold storage is established. The energy conservation equation describes the temperature field distribution, the mass conservation equation describes the humidity field change, and the momentum conservation equation describes the air flow. The temperature transfer process is described by a three-dimensional unsteady-state heat transfer equation. Considering the coupling effects of phase change heat transfer, convective heat transfer, and radiative heat transfer, a set of boundary condition equations is established.

[0030] For the compressor, condenser, evaporator, and expansion valve of the refrigeration system, dynamic characteristic models are established respectively: the compressor model uses a polynomial fitting method to describe the relationship between its cooling capacity and power consumption and the changes in evaporation temperature, condensation temperature, and rotational speed; the condenser and evaporator use the moving boundary method to establish dynamic heat transfer models, considering the two-phase flow and phase change process of the refrigerant; the expansion valve uses an isenthalpic expansion model; the models of each component are coupled and connected through the mass flow rate and enthalpy value of the refrigerant to form a complete dynamic simulation model of the refrigeration cycle.

[0031] The established partial differential equations are numerically discretized using the finite element method or the finite difference method. The cold storage space is divided into regular hexahedral mesh elements, with the mesh density adaptively adjusted according to the temperature gradient, and finer meshing near the heat exchange equipment. The time step is determined using an adaptive algorithm to ensure numerical stability while improving computational efficiency. The solver employs the SIMPLE algorithm to handle pressure-velocity coupling problems and the PISO algorithm to handle unsteady-state terms, ensuring the convergence and accuracy of the numerical solution.

[0032] The digital twin model is calibrated using historical operating data collected in Step 1. Empirical parameters in the model, such as heat transfer coefficient, drag coefficient, and equipment efficiency, are optimized using the least squares method or genetic algorithm. The calibration process is divided into two stages: first, 70% of the historical data is used for parameter identification, and then the remaining 30% of the data is used for model verification. The model accuracy is evaluated by comparing key indicators such as temperature field distribution and energy consumption trends between simulation results and measured data.

[0033] Establish a real-time data synchronization mechanism between the digital twin model and the physical cold storage. Obtain real-time operating data from the cold storage control system via OPCUA or MQTT protocols, including temperature and humidity at various points, equipment operating parameters, and power consumption. At the same time, establish a model state update algorithm to dynamically adjust the boundary conditions and initial conditions of the model based on real-time data, maintaining a high degree of consistency between the digital twin model and the physical entity.

[0034] The predictive analysis module predicts future energy consumption demand and load changes based on the cold storage digital twin model and real-time operational data, outputting energy consumption prediction results and load analysis reports. The method for predicting future energy consumption demand is as follows: For short-term demand, based on the thermodynamic transfer characteristics of the digital twin model, the dynamic heat load distribution of the cold storage is calculated in real-time by solving the three-dimensional unsteady heat transfer equation. Based on real-time environmental temperature and humidity, cargo heat load characteristics, and personnel activity data, the module simulates cold energy loss paths, accurately quantifying the cold load required by the cold storage in the current and short-term future, providing a basis for equipment operation and adjustment.

[0035] To address medium- and long-term energy management needs, a deep learning algorithm is used to train a time series prediction model with historical cooling load data, environmental temperature and humidity variation patterns, cargo storage plans, and electricity price fluctuation cycles as input features. The model outputs the peak and valley values ​​of cooling load and the corresponding energy consumption prediction values ​​for different future periods.

[0036] A deep analysis of the energy consumption distribution and operating status of the cold storage is conducted: the energy consumption ratio of each subsystem is broken down using a digital twin model to identify abnormal energy consumption patterns; the energy efficiency ratio change trend of each core device is calculated to determine whether the device is in a high-efficiency operating range, and finally a load analysis report is generated that includes energy consumption prediction results, load change curves, anomaly identification reports, and energy efficiency analysis conclusions, clarifying the key directions for subsequent optimization.

[0037] In the strategy generation module, a multi-objective optimization mathematical model is established based on energy consumption prediction results and load analysis to generate a dynamic optimization control strategy. This can be achieved by: establishing a multi-objective optimization mathematical model with minimizing the total energy consumption of the cold storage as the primary objective, while using the temperature fluctuation range within the cold storage and the cumulative operating time of the equipment as constraints, and quantifying the weight of each objective; clearly defining specific constraint boundaries in the model, setting temperature range thresholds based on cargo storage requirements, setting operating limits based on equipment technical parameters, and setting hard constraints in conjunction with safety regulations.

[0038] Intelligent algorithms suitable for multi-objective optimization, such as genetic algorithms and particle swarm optimization, are employed to solve the model. Using compressor operating frequency, fan speed, valve opening degree, and defrosting time as optimization variables, the algorithm iteratively searches for the optimal solution, ultimately generating time-segmented equipment operation strategies. Simultaneously, alternative strategies are generated for different scenarios, such as peak cargo inbound periods and extreme weather, forming a dynamic optimization control strategy set to ensure the flexibility and adaptability of the strategies.

[0039] In the execution feedback module, the dynamic optimization control strategy is transformed into equipment control commands. Real-time monitoring of control effect data dynamically adjusts control parameters, outputting optimized cold storage operation status and energy consumption data. The module automatically triggers a feedback adjustment mechanism when monitoring data shows a deviation between the actual state and the expected target. The system uses the real-time status of the digital twin model to pinpoint the cause of the deviation: if the temperature is too high and the monitored cold load value is higher than the predicted value, it may be because the initial cold load forecast did not consider the additional heat brought by the goods entering the warehouse; if the compressor's actual frequency meets the standard but the cooling effect is insufficient, it may be due to condenser blockage leading to decreased heat exchange efficiency, or compressor aging causing a decrease in COP; if the ambient temperature suddenly rises, it may be due to increased external heat penetration leading to increased cooling loss. After identifying the cause, the module relies on the pre-simulation capability of the digital twin model to simulate the execution effect of different adjustment schemes: for example, for the case of "cold load exceeding prediction," it simulates two schemes: "increasing the compressor frequency to 55Hz" and "opening the expansion valve to 40%." The model calculates the temperature drop rate and additional energy consumption increment under both schemes, and selects the optimal adjustment scheme by comprehensively considering the objectives of "temperature stability" and "energy consumption control."

[0040] After determining the optimal adjustment scheme, the module generates new control parameters and sends them to the PLC / DCS system through the existing command conversion and transmission link to drive the equipment to adjust its operating status. Simultaneously, the deviation data, adjustment scheme, and execution effect data are transmitted to the digital twin model in real time to update key parameters in the model: for example, if the actual COP of the compressor is lower than the initial value in the model, the compressor efficiency coefficient in the refrigeration system mathematical model is corrected; if the external heat penetration exceeds expectations, the wall heat transfer coefficient in the thermodynamic transfer model is updated. Through parameter updates, the operating status of the digital twin model and the physical entity of the cold storage is ensured to remain highly consistent, avoiding errors in subsequent predictive analysis and strategy generation due to model deviations, thus laying the foundation for continuous and accurate system optimization.

[0041] In the learning and optimization module, the optimized cold storage operation status and energy consumption data are compared with those before optimization. The model is then learned and optimized through: establishing an optimization effect evaluation system by comparing before and after optimization data and predicted data to analyze the actual effect and applicable conditions of the optimization strategy; using reinforcement learning to iteratively upgrade the optimization strategy, while establishing an optimization case knowledge base to classify and store optimization strategies with significant effects, so that mature strategies can be quickly invoked when similar scenarios are encountered; and periodically adjusting the core parameters of the optimization algorithm in response to seasonal changes and changes in cargo storage patterns.

[0042] Secondly: The accompanying drawings of the embodiments disclosed in this invention only involve the structures involved in the embodiments disclosed in this invention. Other structures can refer to the general design. In the absence of conflict, the same embodiment and different embodiments of this invention can be combined with each other. In conclusion, the above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A dynamic energy consumption optimization system for cold storage based on digital twins, characterized in that, include: The data acquisition module is used to collect basic structural data of cold storage, technical parameters of refrigeration system, real-time operation data and historical operation data, and outputs standardized multi-dimensional real-time data streams after data preprocessing; The model building module is used to build 3D geometric models, thermodynamic transfer models, and refrigeration system mathematical models based on multi-dimensional real-time data streams. After machine learning calibration of the model parameters using historical operating data, the models are integrated to build a digital twin model of the cold storage. The predictive analysis module is used to predict future energy consumption demand and load changes based on the cold storage digital twin model and real-time operation data, and output energy consumption prediction results and load analysis reports. The strategy generation module is used to establish a multi-objective optimization mathematical model based on energy consumption prediction results and load analysis, and generate dynamic optimization control strategies. The execution feedback module is used to convert the dynamic optimization control strategy into equipment control commands, monitor the control effect data in real time, dynamically adjust the control parameters, and output the optimized cold storage operation status and energy consumption data. The learning optimization module is used to compare the optimized cold storage operation status and energy consumption data with those before optimization, and to learn and optimize the model.

2. The cold storage energy consumption dynamic optimization system based on digital twin as described in claim 1, characterized in that: The real-time and historical operating data include: real-time operating data includes multi-point temperature data inside the cold storage, ambient temperature data, equipment surface temperature data, relative humidity and absolute humidity data, pressure data including high and low pressure sides of the refrigeration system and static pressure data of the fan, and electrical parameters including current, voltage, power factor and frequency of each piece of equipment; historical operating data includes recent continuous operating data including temperature, humidity, pressure and flow records, energy consumption data under different operating conditions, equipment performance degradation data, fault and maintenance records.

3. The cold storage energy consumption dynamic optimization system based on digital twin as described in claim 1, characterized in that: The 3D geometric model is constructed as follows: a precise geometric model of the cold storage is built based on the basic structural data of the cold storage, including the external wall structure, internal compartments, insulation layer thickness, door and window positions, and shelf layout; at the same time, a coordinate system is established.

4. The cold storage energy consumption dynamic optimization system based on digital twin as described in claim 1, characterized in that: The thermodynamic transfer model is constructed as follows: Based on the 3D geometric model, a mathematical model describing the heat and mass transfer process in the cold storage is established: the energy conservation equation describes the temperature field distribution, the mass conservation equation describes the humidity field change, and the momentum conservation equation describes the air flow; the temperature transfer process is described by the three-dimensional unsteady heat transfer equation; and the coupling effect of phase change heat transfer, convective heat transfer and radiative heat transfer is considered to establish a set of boundary condition equations.

5. The cold storage energy consumption dynamic optimization system based on digital twin as described in claim 1, characterized in that: The mathematical model of the refrigeration system is constructed as follows: dynamic characteristic models are established for the compressor, condenser, evaporator, and expansion valve of the refrigeration system respectively: the compressor model uses a polynomial fitting method to describe the relationship between its cooling capacity and power consumption and the changes in evaporation temperature, condensation temperature, and rotational speed; the condenser and evaporator use the moving boundary method to establish dynamic heat transfer models, considering the two-phase flow and phase change process of the refrigerant; the expansion valve uses an isenthalpic expansion model; the models of each component are coupled and connected through the mass flow rate and enthalpy value of the refrigerant.

6. The cold storage energy consumption dynamic optimization system based on digital twin as described in claim 1, characterized in that: The method for predicting future energy consumption demand is as follows: For short-term demand, based on the thermodynamic transfer characteristics of the digital twin model, the dynamic heat load distribution of the cold storage is calculated in real time by solving the three-dimensional unsteady heat transfer equation; for medium- and long-term energy management needs, a time series prediction model is trained using historical cold load data, environmental temperature and humidity change patterns, cargo storage plans, and electricity price fluctuation cycles as input features, and outputs the peak and valley values ​​of cold load and the corresponding energy consumption prediction values ​​for different future periods.

7. The cold storage energy consumption dynamic optimization system based on digital twin as described in claim 1, characterized in that: The load analysis report is generated by: conducting in-depth analysis of the energy consumption distribution and operating status of the cold storage; breaking down the energy consumption ratio of each subsystem through a digital twin model and identifying abnormal energy consumption patterns. Calculate the energy efficiency ratio change trend of each core device, determine whether the device is in the high-efficiency operating range, and finally generate a load analysis report that includes energy consumption prediction results, load change curves, anomaly identification reports, and energy efficiency analysis conclusions.

8. The cold storage energy consumption dynamic optimization system based on digital twin as described in claim 1, characterized in that: The method for generating dynamic optimization control strategy is as follows: establish a multi-objective optimization mathematical model, with the minimum total energy consumption of the cold storage as the primary objective, while taking the temperature fluctuation range and cumulative operating time of the equipment in the cold storage as constraints, and quantify the weight of each objective. Using compressor operating frequency, fan speed, valve opening degree, and defrosting time as optimization variables, the algorithm iteratively searches for the optimal solution and finally generates a time-segmented equipment operation strategy.

9. The cold storage energy consumption dynamic optimization system based on digital twin as described in claim 1, characterized in that: The method for dynamically adjusting control parameters is as follows: when monitoring data shows a deviation between the actual state and the expected target, the system automatically activates the feedback adjustment mechanism, analyzes the cause of the deviation, simulates the effects of different adjustment schemes based on the digital twin model, and dynamically adjusts the control parameters; at the same time, the feedback data is transmitted to the digital twin model, the equipment performance parameters in the model are updated, and the optimized cold storage operation status and energy consumption data are output.

10. The cold storage energy consumption dynamic optimization system based on digital twin as described in claim 1, characterized in that: The method for learning and optimizing the model is as follows: establish an optimization effect evaluation system, analyze the actual effect and applicable conditions of the optimization strategy by comparing the data before and after optimization and the predicted data; use reinforcement learning to iteratively upgrade the optimization strategy, and at the same time establish an optimization case knowledge base to classify and store optimization strategies with significant effects, so that mature strategies can be quickly called when similar scenarios are encountered; and regularly adjust the core parameters of the optimization algorithm in response to seasonal changes and changes in cargo storage patterns.

Images