Phase change cold storage method applied to refrigerating system, computer equipment and storage medium

Abstract

The invention relates to the technical field of industrial energy management, and discloses a phase change cold storage method applied to a refrigerating system, computer equipment and a storage medium. According to the phase change cold storage method, historical operation data of a refrigeration system is obtained, a prediction cold load model of the refrigeration system is constructed by adopting an LSTM algorithm, the minimum total energy consumption cost of the refrigeration system is used as a target function based on the prediction cold load model, a reference strategy of the refrigeration system is generated by adopting a reverse recursive algorithm, the reference strategy is executed, and the phase change cold storage is achieved. Dynamic correction is conducted on line, an instruction in the finally-determined reference strategy is issued to the refrigerating system, and regulation and control operation of the refrigerating system is achieved; the benchmark strategy generated by combining the precise prediction cold load model and the reverse recursive algorithm can improve the energy efficiency of the refrigerating system, the production cost is further reduced, meanwhile, the application scene is wide, and the control requirement of the refrigerating system on precise load regulation and control is met; the refrigerating system is high in adaptability and has a good industrial prospect.

Description

Technical Field

[0001] This invention relates to the field of industrial energy management technology, and more specifically to a phase change cold storage method, computer equipment, and storage medium applied to a refrigeration system. Background Technology

[0002] Refrigeration systems are a critical component of modern industrial and commercial infrastructure, playing a vital role in maintaining product quality, ensuring production process stability, and improving energy efficiency. In cigarette factories, refrigeration systems control and regulate temperature and humidity in production workshops, storage areas, and other critical zones to ensure cigarettes and finished products are stored and processed under optimal conditions, guaranteeing consistent quality. Currently, most refrigeration systems are energy inefficient and consume a lot of energy, increasing operating costs. Furthermore, most refrigeration systems lack flexible operation control strategies and cannot adjust to environmental conditions and user needs, resulting in slow response to load changes, poor system adaptability, and an inability to meet the current demands of cigarette factories. Summary of the Invention

[0003] To achieve the above objectives, this invention provides a phase change cold storage method for refrigeration systems. This method acquires historical operating data of the refrigeration system, constructs a predictive cooling load model using the LSTM algorithm, and generates a baseline strategy for the refrigeration system based on the predicted cooling load model, with the goal of minimizing the total energy consumption cost of the refrigeration system. The baseline strategy is then executed, and dynamic corrections are performed online. Finally, the instructions from the final determined baseline strategy are issued to the refrigeration system to achieve regulated operation. The accurate predicted cooling load model combined with the baseline strategy generated by the inverse recursive algorithm improves the energy efficiency of the refrigeration system, further reducing production costs. It also has wide applicability, meeting the requirements for precise load control in refrigeration systems. Furthermore, it exhibits high adaptability to refrigeration systems and has promising industrial prospects.

[0004] To achieve the above objectives, a first aspect of the present invention provides a phase change cold storage method applied to a refrigeration system, the phase change cold storage method comprising: Acquire and preprocess historical operating data of the refrigeration system; The LSTM algorithm is used to construct a predictive cooling load model for the refrigeration system; Based on the predicted cooling load model, with the objective function of minimizing the total energy consumption cost of the refrigeration system, a baseline strategy for the refrigeration system is generated using a reverse recursive algorithm. The baseline strategy is executed, and dynamic correction is performed online. The instructions in the final baseline strategy are then sent to the refrigeration system to achieve operation control of the refrigeration system.

[0005] Preferably, the LSTM algorithm is used to construct a predictive cooling load model for the refrigeration system, including: Based on the historical operating data, the LSTM algorithm is used to construct a feature sequence dataset for predicting the cooling load model according to formula (1). (1) in, for The feature sequence dataset at time step, for The input temperature of the refrigeration system at any given time. for The output temperature of the cooling system at any given time. for The cooling load of the refrigeration system at any given time. for The ambient temperature at any given moment; The feature sequence dataset is input into the LSTM network using the LSTM algorithm, and the final output is obtained according to formula (2). (2) in, for The hidden energy state at all times for The calculation result of the unit output gate at time 1. for The unit energy state at any given time; The final output of the LSTM network is input into a fully connected network, and the predicted cooling load output of the refrigeration system is obtained according to formula (3). (3) in, for Predicted cooling load output at any given time. These are the weighting coefficients. Bias coefficient The weighting and bias coefficients of the predicted cooling load model are optimized based on the historical operating data to obtain the final predicted cooling load model.

[0006] Preferably, based on the predicted cooling load model, with the objective function of minimizing the total energy consumption cost of the refrigeration system, a baseline strategy for generating the refrigeration system is generated using a reverse recursive algorithm, including: The optimized predictive cooling load model is used as input to the latest feature sequence dataset to obtain the predicted cooling load curve for future periods. Obtain the required phase change material, and based on the heat exchange efficiency, obtain the actual cooling capacity of the phase change material according to formula (4). (4) in, This represents the actual cooling capacity of the refrigeration system. For heat exchange efficiency, The latent heat of phase change in phase change materials. For the mass of the phase change material; Based on the predicted cooling load curve output for the future period, the future period is divided into... In a series of consecutive time periods, the refrigeration system performs one of the refrigeration operations in formula (5) during each time period. To achieve the transfer of energy states, (5) in, Representing the The refrigeration operations that the refrigeration system needs to perform during a given time period; A reverse recursive algorithm is used to perform reverse recursion to obtain the energy state of the cooling system at the beginning of each time period, the feasible cooling operations, and the energy state of the corresponding next time period. Calculate the energy state of the cooling system at the beginning of each time period and the instantaneous cost under feasible cooling operations, and calculate the energy state of the cooling system at the beginning of each time period and the final cost under optimal cooling operations. The objective function for the total energy consumption cost of the refrigeration system is constructed using formula (6). (6) in, Let the objective function be the total energy cost. For the first Real-time cost for a given time period For the first Energy state over a period of time For the first Cooling operation for a specific time period For from the first The sum of the minimum future costs starting from the energy state over a given time period; After minimizing the objective function, obtain the energy state and optimal cooling operation strategy table for all time periods in the future cycle, and generate the baseline strategy.

[0007] Preferably, a reverse recursive algorithm is used to perform reverse recursion to obtain the energy state of the cooling system at the beginning of each time period, the feasible cooling operations, and the energy state of the corresponding next time period, including: Obtain the energy state of the refrigeration system at the beginning of each time period, and determine whether the cold storage operation under this energy state can meet the constraints of formulas (7)-(8). (7) (8) in, For the first The planned cold storage capacity for a given period of time. This represents the maximum cooling capacity of the refrigeration system. The duration of each time period, For safety reasons, This represents the minimum load requirement during nighttime cold storage. Under the condition that the constraints are met, the cold storage operation is feasible, and the energy state for the next time period is calculated according to formula (9); (9) in, For the first Energy status of the refrigeration system over a +1 time period. As the baseline cold storage capacity; Obtain the energy state of the refrigeration system at the beginning of each time period, and determine whether the heat release operation under that energy state can satisfy the constraints of formulas (10)-(11). (10) (11) in, This refers to the power required by the refrigeration system with the aid of heat release. For the first The planned release of cooling capacity over a given time period. For the first Forecast cooling load for a given time period; Under the condition that the constraints are met, the refrigeration operation is feasible, and the energy state for the next time period can be calculated according to formula (12). (12) When the system is idle, the energy state of the refrigeration system does not change.

[0008] Preferably, the energy state of the refrigeration system at the beginning of each time period and the immediate cost under feasible refrigeration operation are calculated, as well as the energy state of the refrigeration system at the beginning of each time period and the final cost under optimal refrigeration operation, including: Based on the energy state of the refrigeration system and feasible refrigeration operations at the beginning of each time period, the instantaneous cost incurred during cold storage operations is obtained through formula (13). (13) in, For the first Real-time cost for a given time period For the first Electricity price for a specific time period, As the baseline cold storage capacity; Based on the energy state of the refrigeration system and feasible refrigeration operations at the beginning of each time period, the instantaneous cost incurred when performing either a cooling release operation or an idle operation is obtained through formula (14). (14) in, This refers to the power required by the refrigeration system with cooling assistance. Based on the recursive calculation using the reverse recursive algorithm, the energy state of the refrigeration system at the beginning of each time period and the final cost under optimal refrigeration operation are determined according to formula (15). (15) in, For from the first The sum of the minimum future costs starting from the energy state over a given time period.

[0009] Preferably, a baseline strategy is executed, dynamic correction is performed online, and the instructions in the final baseline strategy are issued to the refrigeration system to achieve operation control of the refrigeration system, including: Execute the baseline strategy; A high-frequency real-time compensation method is used to perform real-time online correction of the refrigeration system, while a low-frequency rolling optimization method is used to perform local planning optimization of the refrigeration system. The instructions in the finalized baseline strategy are issued to the refrigeration system to achieve operation and control of the refrigeration system.

[0010] Preferably, a high-frequency real-time compensation method is used to perform real-time online correction of the refrigeration system, including: Determine whether the current refrigeration operation is a cold release operation; When the current refrigeration operation is a cold release operation, the target cold release amount is corrected in real time using formula (16). (16) in, for The target amount of cooling released at any given time. for The baseline cooling rate in the baseline strategy at that time. for Constant load disturbances for Dynamic weighting coefficients at any given time; Determine whether the current refrigeration operation is a cold storage operation; When the current refrigeration operation is a cold storage operation, the target cold storage power is corrected in real time using formula (17). (17) in, for Target cold storage capacity at any time for The baseline cold storage power in the baseline strategy at any given time. This is the proportional control gain coefficient. for The deviation in the amount of cold storage at any given moment.

[0011] Preferably, a low-frequency rolling optimization method is used simultaneously to perform local planning optimization of the refrigeration system, including: Determine whether the current moment triggers the condition for using a low-frequency rolling optimization method to perform local planning optimization of the refrigeration system; When the conditions for local planning optimization of the refrigeration system using the low-frequency rolling optimization method are triggered at the current moment, the current moment is taken as the new starting point. Based on the latest data up to the current moment, the LSTM algorithm is used to regenerate the latest predicted cooling load output curve for the remaining time period within the cycle. With the minimization of the total energy consumption cost of the remaining time period as the objective function, the inverse recursive algorithm is used to generate a new baseline strategy for the remaining time period, which replaces the remaining time period part in the original baseline strategy, thereby achieving local planning optimization.

[0012] A second aspect of the present invention provides a computer device, including a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to implement the phase change cooling method as described in any of the preceding claims.

[0013] A third aspect of the present invention provides a computer-readable storage medium storing a computer program, which, when executed by a processor, implements the phase change cooling method as described in any of the preceding claims.

[0014] Through the above technical solution, this phase change cold storage method uses historical operating data of the refrigeration system to construct a predictive cooling load model of the refrigeration system using the LSTM algorithm, generating predictive cooling load curves for future periods. Based on the operating temperature range of the refrigeration system, suitable phase change materials are selected for cold storage. Using the predicted cooling load model and minimizing the total energy consumption cost of the refrigeration system as the objective function, suitable phase change materials are selected as cold storage materials. Considering electricity price fluctuations and system operating constraints, a reverse recursive algorithm is used to generate a baseline strategy for the refrigeration system. This baseline strategy is executed, and dynamic correction is performed online through a dual-layer control structure of high-frequency real-time compensation and low-frequency rolling optimization. Finally, the instructions in the determined baseline strategy are issued to the refrigeration system, realizing the regulation and operation of the refrigeration system. The system is capable of improving energy efficiency through a precise predictive cooling load model and a reverse recursive algorithm-generated baseline strategy. A dual-layer control structure combining high-frequency real-time compensation and low-frequency rolling optimization enables rapid response to load fluctuations and prediction errors, enhancing the system's adaptability and stability under complex operating conditions. By storing cold during low-electricity-price periods at night and releasing it during high-electricity-price periods during the day, electricity costs are effectively reduced. Simultaneously, pre-cooling lowers the chiller's inlet water temperature, reducing the main unit's power requirements and further saving energy. The introduction of minimum load constraints and real-time compensation mechanisms during the cold storage phase ensures that the refrigeration system still meets the rigid requirements of the production process during cold storage / release, improving system reliability and production quality. Furthermore, the refrigeration system boasts high adaptability and possesses promising industrial application prospects and promotional value. Attached Figure Description

[0015] Figure 1 This is a connection block diagram of a phase change cold storage method applied to a refrigeration system according to an embodiment of the present invention; Figure 2 This is a connection block diagram for constructing a predictive cooling load model of a refrigeration system according to an embodiment of the present invention, which is applied to a phase change cold storage method for a refrigeration system. Figure 3 This is a connection block diagram generated from the baseline strategy of a refrigeration system for a phase change cold storage method applied to a refrigeration system according to an embodiment of the present invention. Figure 4 This is a connection block diagram of a phase change cold storage method applied to a refrigeration system according to an embodiment of the present invention, which realizes high-frequency real-time compensation and online correction of the refrigeration system. Detailed Implementation

[0016] The specific embodiments of the present invention will be described in detail below with reference to the accompanying drawings. It should be understood that the specific embodiments described herein are for illustration and explanation only and are not intended to limit the scope of the present invention.

[0017] like Figure 1The diagram shown is a connection block diagram of a phase change cold storage method applied to a refrigeration system according to an embodiment of the present invention; in Figure 1 The phase change cold storage method includes the following steps: In step S10, historical operating data of the refrigeration system is acquired and preprocessed; In step S11, the LSTM algorithm is used to construct a predictive cooling load model for the refrigeration system; In step S12, a baseline strategy for the refrigeration system is generated based on the predicted cooling load model and with the goal of minimizing the total energy consumption cost of the refrigeration system. A reverse recursive algorithm is then used.

[0018] In step S13, the baseline strategy is executed, dynamic correction is performed online, and the instructions in the final baseline strategy are issued to the refrigeration system to realize the operation control of the refrigeration system.

[0019] Step S10 can collect historical operating data of the refrigeration system in real time through sensors, obtain energy consumption data, electricity price data and operating temperature data of the refrigeration system's chiller, and obtain data of the phase change material of the refrigeration system; and perform standardized cleaning on the energy consumption data, electricity price data, operating temperature data of the chiller and the data of the phase change material to obtain the final historical operating data of the refrigeration system.

[0020] Step S11 uses the LSTM algorithm to learn from the historical operating data of the refrigeration system and construct a predictive cooling load model for the refrigeration system, which facilitates the subsequent prediction of the predicted cooling load curve for a certain period of time in the future. Step S12 generates a predicted cooling load curve for a certain period of time in the future by using a predicted cooling load model, obtains the required phase change material as a cold storage material, and uses the minimum total energy consumption cost of the refrigeration system in the future as the objective function, and uses a reverse recursive algorithm to generate the baseline strategy of the refrigeration system.

[0021] Step S12 involves first allowing the refrigeration system to operate according to the instructions in the baseline strategy. Since errors and disturbances are prone to occur during actual operation, online dynamic correction is required. The generated new baseline strategy replaces the original baseline strategy, and the refrigeration system operates according to the instructions in the finally determined baseline strategy.

[0022] Through the above technical solution, this phase change cold storage method can improve the energy efficiency of the refrigeration system and further reduce production costs by combining a precise predictive cooling load model with a reverse recursive algorithm to generate a baseline strategy. It also has wide applicability, meeting the control requirements of precise load regulation in refrigeration systems; the refrigeration system has high adaptability and good industrial prospects. like Figure 2The diagram shown is a connection block diagram for constructing a predictive cooling load model of a refrigeration system according to an embodiment of the present invention, which applies a phase change cold storage method to a refrigeration system; in Figure 2 In order to establish a predictive cooling load model for the refrigeration system, in one embodiment of the present invention, the LSTM algorithm is used to construct the predictive cooling load model of the refrigeration system, including the following steps: In step S20, based on the historical operating data, the LSTM algorithm is used to construct a feature sequence dataset for predicting the cooling load model according to formula (1). (1) in, for The feature sequence dataset at time step, for The input temperature of the refrigeration system at any given time. for The output temperature of the cooling system at any given time. for The cooling load of the refrigeration system at any given time. for The ambient temperature at any given moment; In step S21, the feature sequence dataset is input into the LSTM network using the LSTM algorithm, and the final output is obtained according to formula (2). (2) in, for The hidden state at all times for The calculation result of the unit output gate at time 1. for The cell state at any given time; In step S22, the final output of the LSTM network is input into the fully connected network, and the predicted cooling load output of the refrigeration system is obtained according to formula (3). (3) in, for Predicted cooling load output at any given time. These are the weighting coefficients. Bias coefficient In step S23, the weight coefficients and bias coefficients of the predicted cooling load model are optimized based on the historical operating data to obtain the final predicted cooling load model. A feature sequence dataset is constructed based on the historical operating data, input into an LSTM network to obtain the final output, which is then passed through a fully connected layer to output the predicted cooling load. The response parameters are optimized to complete the model construction.

[0023] like Figure 3 The diagram shown is a connection block diagram of a refrigeration system based on a phase change cold storage method applied to a refrigeration system according to an embodiment of the present invention; in Figure 3 In order to generate a baseline strategy and implement subsequent regulation of the refrigeration system, in one embodiment of the present invention, ensuring that the refrigeration system, based on a predicted cooling load model, uses the minimization of the total energy consumption cost of the refrigeration system as the objective function, and generating the baseline strategy of the refrigeration system using a reverse recursive algorithm may include the following steps: In step S30, the latest feature sequence dataset is input using the optimized predicted cooling load model to obtain the predicted cooling load curve output for future periods; In step S31, the required phase change material is obtained, and the actual cooling capacity of the phase change material is obtained according to formula (4) based on the heat exchange efficiency. (4) in, This represents the actual cooling capacity of the refrigeration system. For heat exchange efficiency, The latent heat of phase change in phase change materials. For the mass of the phase change material; In step S32, based on the output of the predicted cooling load curve for the future period, the future period is divided into... In a series of consecutive time periods, the refrigeration system performs one of the refrigeration operations in formula (5) during each time period. To achieve the transfer of energy states, (5) in, Representing the The refrigeration operations that the refrigeration system needs to perform during a given time period; In step S33, a reverse recursive algorithm is used to perform reverse recursion to obtain the energy state of the refrigeration system at the beginning of each time period, the feasible refrigeration operation, and the energy state of the next time period. In step S34, the energy state of the refrigeration system at the beginning of each time period and the instantaneous cost under feasible refrigeration operation are calculated, and the energy state of the refrigeration system at the beginning of each time period and the final cost under optimal refrigeration operation are calculated. In step S35, the objective function for the total energy consumption cost of the refrigeration system is constructed using formula (6). (6) in, Let the objective function be the total energy cost. For the first Real-time cost for a given time period For the first Energy state over a period of time For the first Cooling operation for a specific time period For from the first The sum of the minimum future costs starting from the energy state over a given time period; In step S36, after minimizing the objective function, the energy state and optimal cooling operation strategy table for all time periods in the future cycle are obtained, and a baseline strategy is generated. The trained predictive cooling load model is used to predict the cooling load curve for the future cycle, combined with the actual cold storage capacity of the phase change material. The entire cycle is divided into multiple consecutive time periods, and the cold storage, release, or idle operations that the cooling system can perform in each time period are defined. The core is the use of a reverse recursive dynamic programming method: starting from the end of the cycle, the calculation proceeds backwards, evaluating all feasible operations for each possible cold storage state of the cooling system at the beginning of each time period, calculating its immediate operating cost and the sum of the minimum future cost from the next state to the end of the cycle after executing the operation. Through repeated comparisons, the minimum total cost and its corresponding optimal operation for each state are recorded. Starting from the beginning state of the cycle, the optimal operation sequence for all time periods is extracted forwards based on the record table, thus generating the baseline control strategy with the lowest global total energy consumption cost.

[0024] To ensure that feasible cooling operations can be performed within each time period and meet the corresponding constraints, in one embodiment of the present invention, a reverse recursive algorithm is used to perform reverse recursion to obtain the energy state of the cooling system at the beginning of each time period, feasible cooling operations, and the energy state of the corresponding next time period. This may include the following steps: In step S40, the energy state of the refrigeration system at the beginning of each time period is obtained, and it is determined whether the cold storage operation under this energy state can meet the constraints of formulas (7)-(8). (7) (8) in, For the first The planned cold storage capacity for a given period of time. This represents the maximum cooling capacity of the refrigeration system. The duration of each time period, For safety reasons, This represents the minimum load requirement during nighttime cold storage. In step S41, the cold storage operation is feasible if the constraints are met, and the energy state for the next time period is calculated according to formula (9). (9) in, For the first Energy status of the refrigeration system over a +1 time period. As the baseline cold storage capacity; In step S42, the energy state of the refrigeration system at the beginning of each time period is obtained, and it is determined whether the cooling operation under this energy state can satisfy the constraints of formulas (10)-(11). (10) (11) in, This refers to the power required by the refrigeration system with the aid of heat release. For the first The planned release of cooling capacity over a given time period. For the first Forecast cooling load for a given time period; In step S43, the refrigeration operation is feasible if the constraints are met, and the energy state for the next time period is calculated according to formula (12). (12) In step S44, during the idle operation, the energy state of the refrigeration system does not change.

[0025] During the generation of the baseline strategy, the feasibility of each planned cooling operation must be assessed for the initial system energy level in each time period. For planned cold storage operations, two constraints must be met simultaneously: first, the total energy of the cooling system after cold storage must not exceed the maximum capacity of the cold storage device; second, the remaining cooling capacity of the cooling system during cold storage must be sufficient to meet the minimum load requirements for nighttime production. If both are satisfied, the cold storage operation is feasible, and the system energy level in the next time period after execution will be equal to the current energy plus the stored cold energy. For planned cold release operations, two constraints must also be met: first, the energy currently stored in the cooling system must be greater than or equal to the planned cold energy release, i.e., it cannot be overdrawn; second, when the cooling system releases cold energy, the sum of the cooling output of the cooling system and the cold energy released by the cold storage device must at least meet the predicted load demand for that time period. If both are satisfied, the cold release operation is feasible, and the system energy level in the next time period after execution will be the current energy minus the released cold energy. If an idle operation is selected, the system energy level remains unchanged, and the process proceeds directly to the next time period.

[0026] To find the cooling operation path with the minimum total loss cost within a cycle, in one embodiment of the present invention, the energy state of the cooling system at the beginning of each time period and the immediate cost under feasible cooling operations are calculated, and the energy state of the cooling system at the beginning of each time period and the final cost under optimal cooling operations are calculated, including: In step S50, based on the energy state of the refrigeration system and feasible refrigeration operations at the beginning of each time period, the instantaneous cost incurred during cold storage operations is obtained using formula (13). (13) in, For the first Real-time cost for a given time period For the first Electricity price for a specific time period, As the baseline cold storage capacity; In step S51, based on the energy state of the refrigeration system and feasible refrigeration operations at the beginning of each time period, the instantaneous cost incurred when performing either a cooling release operation or an idle operation is obtained through formula (14). (14) in, This refers to the power required by the refrigeration system with cooling assistance. In step S52, based on the recursive calculation of the reverse recursive algorithm, the energy state of the refrigeration system at the beginning of each time period and the final cost under optimal refrigeration operation are determined according to formula (15). (15) in, For from the first The sum of minimum future costs starting from the energy state of each time period. In the reverse recursive calculation, it is necessary to determine the cost of each time period under various feasible operations.

[0027] The immediate cost of performing a cold storage operation is the product of the electricity price for that time period, the cold storage capacity, and the time. The immediate cost of performing a release or idle operation is the product of the electricity price for that time period, the refrigeration system's operating power, and the time. Through recursive calculation, starting from the last time period and working backwards, for each energy state at the beginning of each time period, the operation that minimizes the sum of its immediate cost and the minimum total cost from the start of the next energy state to the end of the next time period is selected from all feasible operations. This minimum value is the minimum total cost from the current state to the end of the cycle, and the corresponding optimal operation is recorded.

[0028] Considering that prediction deviations and disturbances may occur during the actual execution of the baseline strategy, requiring online correction, in one embodiment of the present invention, executing the baseline strategy, performing dynamic correction online, and issuing the instructions in the finally determined baseline strategy to the refrigeration system to achieve the operation control of the refrigeration system may include the following steps: In step S60, the baseline strategy is executed; In step S61, a high-frequency real-time compensation method is used to perform real-time online correction of the refrigeration system, while a low-frequency rolling optimization method is used to perform local planning optimization of the refrigeration system. In step S62, the instructions from the finalized baseline strategy are issued to the refrigeration system to control its operation. The refrigeration system operates according to the instructions in the baseline strategy. Since errors and disturbances are prone to occur during actual operation, online dynamic correction is necessary. High-frequency real-time compensation is used for real-time correction, while a low-frequency rolling optimization method is used to generate a new baseline strategy based on triggering conditions to replace the original remaining baseline strategy. The refrigeration system then operates according to the instructions in the finalized baseline strategy.

[0029] like Figure 4 The diagram shown is a connection block diagram of a phase change cold storage method applied to a refrigeration system according to an embodiment of the present invention, which implements high-frequency real-time compensation and online correction. Figure 4 In order to achieve accurate real-time compensation, in one embodiment of the present invention, the real-time online correction of the refrigeration system using a high-frequency real-time compensation method may include the following steps: In step S70, it is determined whether the current refrigeration operation is a cold release operation; In step S71, when the current refrigeration operation is a cold release operation, the target cold release amount is corrected in real time using formula (16). (16) in, for The target amount of cooling released at any given time. for The baseline cooling rate in the baseline strategy at that time. for Constant load disturbances for Dynamic weighting coefficients at any given time; In step S72, it is determined whether the current refrigeration operation is a cold storage operation; In step S73, when the current refrigeration operation is a cold storage operation, the target cold storage power is corrected in real time using formula (17). (17) in, for Target cold storage capacity at any time for The baseline cold storage power in the baseline strategy at any given time. This is the proportional control gain coefficient. for The system dynamically fine-tunes its cooling capacity based on real-time compensation control, considering the deviation in cooling capacity at any given moment. If a cooling release operation is currently in progress, the load deviation is determined by comparing the instantaneous actual load with the predicted value. The target cooling release capacity is then dynamically adjusted based on this deviation to quickly smooth out demand fluctuations. If a cooling storage operation is currently in progress, the deviation between the current actual cooling storage progress and the planned value is calculated. The target cooling storage power is then adjusted through proportional feedback to ensure that the cooling storage progress closely follows the baseline plan. This mechanism effectively addresses instantaneous disturbances and guarantees the refrigeration system's accurate tracking of actual demand.

[0030] To achieve local optimization of the refrigeration system by applying a baseline strategy, in one embodiment of the present invention, a low-frequency rolling optimization method is simultaneously employed to perform local planning optimization of the refrigeration system, including: In step S80, it is determined whether the condition for using the low-frequency rolling optimization method to perform local planning optimization of the refrigeration system is triggered at the current moment; In step S81, when the condition for using the low-frequency rolling optimization method to perform local planning optimization of the refrigeration system is triggered at the current moment, the current moment is taken as the new starting point. Based on the latest data up to the current moment, the LSTM algorithm is used to regenerate the latest predicted cooling load output curve for the remaining time period within the cycle. With the minimization of the total energy consumption cost of the remaining time period as the objective function, the inverse recursive algorithm is used to generate a new baseline strategy for the remaining time period, which replaces the part of the remaining time period in the original baseline strategy, thereby achieving local planning optimization.

[0031] The refrigeration system triggers low-frequency rolling optimization when any of the following conditions are met: reaching a fixed cycle, prediction error continuously exceeding a threshold, or a significant change in the production plan. Upon triggering, starting from the current moment, the system re-predicts the cooling load for the remaining period using the latest operating data. With the objective of minimizing the total energy cost for the remaining period, a new locally optimal baseline strategy is generated using a reverse recursive algorithm to replace a portion of the original plan for the remaining period. This dynamically corrects the long-term control trajectory and improves the system's adaptability to changes in operating conditions.

[0032] A second aspect of the present invention provides a computer device, including a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to implement the phase change cooling method as described in any of the preceding claims.

[0033] A third aspect of the present invention provides a computer-readable storage medium storing a computer program, which, when executed by a processor, implements the phase change cooling method as described in any of the preceding claims.

[0034] Through the above technical solution, this phase change cold storage method uses historical operating data of the refrigeration system to construct a predictive cooling load model of the refrigeration system using the LSTM algorithm, generating predictive cooling load curves for future periods. Based on the operating temperature range of the refrigeration system, suitable phase change materials are selected for cold storage. Using the predicted cooling load model and minimizing the total energy consumption cost of the refrigeration system as the objective function, suitable phase change materials are selected as cold storage materials. Considering electricity price fluctuations and system operating constraints, a reverse recursive algorithm is used to generate a baseline strategy for the refrigeration system. This baseline strategy is executed, and dynamic correction is performed online through a dual-layer control structure of high-frequency real-time compensation and low-frequency rolling optimization. Finally, the instructions in the determined baseline strategy are issued to the refrigeration system, realizing the regulation and operation of the refrigeration system. The system is capable of improving energy efficiency through a precise predictive cooling load model and a reverse recursive algorithm-generated baseline strategy. A dual-layer control structure combining high-frequency real-time compensation and low-frequency rolling optimization enables rapid response to load fluctuations and prediction errors, enhancing the system's adaptability and stability under complex operating conditions. By storing cold during low-electricity-price periods at night and releasing it during high-electricity-price periods during the day, electricity costs are effectively reduced. Simultaneously, pre-cooling lowers the chiller's inlet water temperature, reducing the main unit's power requirements and further saving energy. The introduction of minimum load constraints and real-time compensation mechanisms during the cold storage phase ensures that the refrigeration system still meets the rigid requirements of the production process during cold storage / release, improving system reliability and production quality. Furthermore, the refrigeration system boasts high adaptability and possesses promising industrial application prospects and promotional value.

[0035] The preferred embodiments of the present invention have been described in detail above with reference to the accompanying drawings. However, the present invention is not limited to the specific details of the above embodiments. Within the scope of the technical concept of the present invention, various simple modifications can be made to the technical solutions of the present invention, and these simple modifications all fall within the protection scope of the present invention. Furthermore, it should be noted that the various specific technical features described in the above embodiments can be combined in any suitable manner without contradiction. To avoid unnecessary repetition, the present invention will not describe the various possible combinations separately.

Claims

1. A phase change cold storage method applied to a refrigeration system, characterized in that, The phase change cold storage method includes: Acquire and preprocess historical operating data of the refrigeration system; The LSTM algorithm is used to construct a predictive cooling load model for the refrigeration system; Based on the predicted cooling load model, with the objective function of minimizing the total energy consumption cost of the refrigeration system, a baseline strategy for the refrigeration system is generated using a reverse recursive algorithm. The baseline strategy is executed, and dynamic correction is performed online. The instructions in the final baseline strategy are then sent to the refrigeration system to achieve operation control of the refrigeration system.

2. The phase change cold storage method according to claim 1, characterized in that, A predictive cooling load model for the refrigeration system is constructed using the LSTM algorithm, including: Based on the historical operating data, the LSTM algorithm is used to construct a feature sequence dataset for predicting the cooling load model according to formula (1). ，（1） in, for The feature sequence dataset at time step, for The input temperature of the refrigeration system at any given time. for The output temperature of the cooling system at any given time. for The cooling load of the refrigeration system at any given time. for The ambient temperature at any given moment; The feature sequence dataset is input into the LSTM network using the LSTM algorithm, and the final output is obtained according to formula (2). ，（2） in, for The hidden energy state at all times for The calculation result of the unit output gate at time 1. for The unit energy state at any given time; The final output of the LSTM network is input into a fully connected network, and the predicted cooling load output of the refrigeration system is obtained according to formula (3). ，（3） in, for Predicted cooling load output at any given time. These are the weighting coefficients. Bias coefficient The weighting and bias coefficients of the predicted cooling load model are optimized based on the historical operating data to obtain the final predicted cooling load model.

3. The phase change cold storage method according to claim 2, characterized in that, Based on a predicted cooling load model, with the objective function of minimizing the total energy consumption cost of the refrigeration system, a baseline strategy for the refrigeration system is generated using a reverse recursive algorithm, including: The optimized predictive cooling load model is used as input to the latest feature sequence dataset to obtain the predicted cooling load curve for future periods. Obtain the required phase change material, and based on the heat exchange efficiency, obtain the actual cooling capacity of the phase change material according to formula (4). ，（4） in, This represents the actual cooling capacity of the refrigeration system. For heat exchange efficiency, The latent heat of phase change in phase change materials. For the mass of the phase change material; Based on the predicted cooling load curve output for the future period, the future period is divided into... In a series of consecutive time periods, the refrigeration system performs one of the refrigeration operations in formula (5) during each time period. To achieve the transfer of energy states, ，（5） in, Representing the The refrigeration operations that the refrigeration system needs to perform during a given time period; A reverse recursive algorithm is used to perform reverse recursion to obtain the energy state of the cooling system at the beginning of each time period, the feasible cooling operations, and the energy state of the corresponding next time period. Calculate the energy state of the cooling system at the beginning of each time period and the instantaneous cost under feasible cooling operations, and calculate the energy state of the cooling system at the beginning of each time period and the final cost under optimal cooling operations. The objective function for the total energy consumption cost of the refrigeration system is constructed using formula (6). ，（6） in, Let the objective function be the total energy cost. For the first Real-time cost for a given time period For the first Energy state over a period of time For the first Cooling operation for a specific time period For from the first The sum of the minimum future costs starting from the energy state over a given time period; After minimizing the objective function, obtain the energy state and optimal cooling operation strategy table for all time periods in the future cycle, and generate the baseline strategy.

4. The phase change cold storage method according to claim 3, characterized in that, A reverse recursive algorithm is used to perform backward recursion to obtain the energy state of the cooling system at the beginning of each time period, the feasible cooling operations, and the energy state of the corresponding next time period, including: Obtain the energy state of the refrigeration system at the beginning of each time period, and determine whether the cold storage operation under this energy state can meet the constraints of formulas (7)-(8). ，（7） ，（8） in, For the first The planned cold storage capacity for a given period of time. This represents the maximum cooling capacity of the refrigeration system. The duration of each time period, For safety reasons, This represents the minimum load requirement during nighttime cold storage. Under the condition that the constraints are met, the cold storage operation is feasible, and the energy state for the next time period is calculated according to formula (9); ，（9） in, For the first Energy status of the refrigeration system over a +1 time period. As the baseline cold storage capacity; Obtain the energy state of the refrigeration system at the beginning of each time period, and determine whether the heat release operation under that energy state can satisfy the constraints of formulas (10)-(11). ，（10） ，（11） in, This refers to the power required by the refrigeration system with the aid of heat release. For the first The planned release of cooling capacity over a given time period. For the first Forecast cooling load for a given time period; Under the condition that the constraints are met, the refrigeration operation is feasible, and the energy state for the next time period can be calculated according to formula (12). ，（12） When the system is idle, the energy state of the refrigeration system does not change.

5. The phase change cold storage method according to claim 3, characterized in that, Calculate the energy state of the cooling system at the beginning of each time period and the immediate cost under feasible cooling operations, and calculate the energy state of the cooling system at the beginning of each time period and the final cost under optimal cooling operations, including: Based on the energy state of the refrigeration system and feasible refrigeration operations at the beginning of each time period, the instantaneous cost incurred during cold storage operations is obtained through formula (13). ，（13） in, For the first Real-time cost for a given time period For the first Electricity price for a specific time period, As the baseline cold storage capacity; Based on the energy state of the refrigeration system and feasible refrigeration operations at the beginning of each time period, the instantaneous cost incurred when performing either a cooling release operation or an idle operation is obtained through formula (14). ，（14） in, This refers to the power required by the refrigeration system with cooling assistance. Based on the recursive calculation using the reverse recursive algorithm, the energy state of the refrigeration system at the beginning of each time period and the final cost under optimal refrigeration operation are determined according to formula (15). ，（15） in, For from the first The sum of the minimum future costs starting from the energy state over a given time period.

6. The phase change cold storage method according to claim 1, characterized in that, The baseline strategy is executed, and dynamic corrections are performed online. The instructions from the finalized baseline strategy are then sent to the refrigeration system to achieve operational control of the refrigeration system, including: Execute the baseline strategy; A high-frequency real-time compensation method is used to perform real-time online correction of the refrigeration system, while a low-frequency rolling optimization method is used to perform local planning optimization of the refrigeration system. The instructions in the finalized baseline strategy are issued to the refrigeration system to achieve operation and control of the refrigeration system.

7. The phase change cold storage method according to claim 6, characterized in that, A high-frequency real-time compensation method is used to perform real-time online correction of the refrigeration system, including: Determine whether the current refrigeration operation is a cold release operation; When the current refrigeration operation is a cold release operation, the target cold release amount is corrected in real time using formula (16). ，（16） in, for The target amount of cooling released at any given time. for The baseline cooling rate in the baseline strategy at that time. for Constant load disturbances for Dynamic weighting coefficients at any given time; Determine whether the current refrigeration operation is a cold storage operation; When the current refrigeration operation is a cold storage operation, the target cold storage power is corrected in real time using formula (17). ，（17） in, for Target cold storage capacity at any time for The baseline cold storage power in the baseline strategy at any given time. This is the proportional control gain coefficient. for The deviation in the amount of cold storage at any given moment.

8. The phase change cold storage method according to claim 7, characterized in that, Simultaneously, a low-frequency rolling optimization method is used to perform local planning optimization of the refrigeration system, including: Determine whether the current moment triggers the condition for using a low-frequency rolling optimization method to perform local planning optimization of the refrigeration system; When the conditions for local planning optimization of the refrigeration system using the low-frequency rolling optimization method are triggered at the current moment, the current moment is taken as the new starting point. Based on the latest data up to the current moment, the LSTM algorithm is used to regenerate the latest predicted cooling load output curve for the remaining time period within the cycle. With the minimization of the total energy consumption cost of the remaining time period as the objective function, the inverse recursive algorithm is used to generate a new baseline strategy for the remaining time period, which replaces the remaining time period part in the original baseline strategy, thereby achieving local planning optimization.

9. A computer device comprising a memory and a processor, wherein the memory stores a computer program, characterized in that, When the processor executes the computer program, it implements the phase change cold storage method according to any one of claims 1 to 8.

10. A computer-readable storage medium storing a computer program, characterized in that, When the computer program is executed by the processor, it implements the phase change cold storage method according to any one of claims 1 to 8.

Images