A method and system for intelligent optimization and control of cold storage equipment for storage objects

Abstract

The present invention relates to the field of information technology, and discloses a method and system for intelligent optimization and control of cold storage equipment for storage objects. The method obtains the temperature sensitivity coefficients of different types of cold storage storage objects, and determines the respective temperature fluctuation acceptance ranges accordingly. Meanwhile, a temperature prediction model is constructed using historical data to predict the temperature change trend of the cold storage equipment under different control strategies. Based on the prediction results and characteristic parameters of the cold storage storage objects, the cold storage equipment control strategies that meet the temperature requirements are screened, and their energy consumption is evaluated to construct the Arrhenius equation to characterize the mass loss index of the cold storage storage objects in the candidate solutions. The candidate strategies are optimized in combination with a genetic algorithm, and the fitness function of the genetic algorithm is constructed through the mass loss index. The optimal control strategy is determined through the genetic algorithm and the execution of the cold storage equipment is controlled, thereby realizing accurate control of the cold storage temperature and energy efficiency optimization, and improving the intelligent level of cold storage management.

Description

Technical Field

[0001] The present invention relates to the field of information technology, and in particular to an intelligent optimization and control method and system for cold storage equipment for storage objects. Background Art

[0002] At present, in order to ensure the safety of objects stored in cold storage, several compressors, condensers, air coolers and other equipment are usually installed in the cold storage, and these devices are controlled to control the temperature of the area where the storage objects are located within the tolerance range of the corresponding storage objects.

[0003] However, different types of storage objects have different temperature resistance, and their corresponding cold storage equipment control strategies must also be different. In addition, temperature fluctuations in the cold storage will affect the operating efficiency of its refrigeration system. The greater the temperature fluctuation, the higher the energy consumption of the refrigeration system. Therefore, how to intelligently optimize and regulate the control strategy of cold storage equipment while taking into account the needs of energy saving and efficiency improvement has become a technical problem that technical personnel in this field need to urgently solve. Summary of the invention

[0004] The present invention provides a method and system for intelligent optimization and regulation of cold storage equipment for storage objects, which solves the problem of how to intelligently optimize and regulate the control strategy of the cold storage equipment while taking into account the needs of energy saving and efficiency improvement.

[0005] In order to solve the above technical problems, the first aspect of the present invention provides a method for intelligent optimization and control of cold storage equipment for storage objects, comprising:

[0006] Acquire temperature sensitivity coefficients of different types of cold storage storage objects, and determine temperature fluctuation acceptance ranges corresponding to the different types of cold storage storage objects based on the temperature sensitivity coefficients;

[0007] Sequentially obtain historical temperature data of each type of cold storage storage object stored in the cold storage and corresponding cold storage equipment operation data, and obtain a trained cold storage temperature prediction model based at least on the temperature sensitivity coefficient and the cold storage equipment operation data;

[0008] Based on each of the temperature sensitivity coefficients, different initial control strategies are simulated and generated and input into the cold storage temperature prediction model for processing to obtain temperature change data under each initial control strategy;

[0009] Compare and analyze each of the temperature change data with the corresponding temperature fluctuation acceptance range, and calculate the energy consumption data corresponding to the target strategy determined from the initial control strategy according to the analysis result;

[0010] Based on the energy consumption data, a candidate solution is selected from the target strategy, and a mass loss index of the candidate solution is determined based on an Arrhenius equation constructed by historical temperature data corresponding to different types of cold storage objects;

[0011] The candidate solutions are screened according to a genetic algorithm, and the cold storage equipment is controlled to execute the determined final control strategy, wherein the screening process is configured to use the fitness function constructed by the quality loss index corresponding to each candidate solution as the objective function for screening.

[0012] As one of the preferred solutions, the step of obtaining temperature sensitivity coefficients of different types of cold storage objects and determining temperature fluctuation acceptance ranges corresponding to the different types of cold storage objects based on the temperature sensitivity coefficients includes:

[0013] Acquire historical temperature data of different types of cold storage objects stored in the cold storage, and construct a feature vector based on the historical temperature data, the types of the cold storage objects and their shelf life, wherein the historical temperature data includes a historical temperature fluctuation amplitude and a historical temperature tolerance range;

[0014] Performing cluster analysis on the feature vectors by using a K-means algorithm to obtain a number of cold storage object groups, and calculating the average temperature tolerance range and average temperature fluctuation amplitude of each of the cold storage object groups;

[0015] quantifying the temperature sensitivity coefficient of each cold storage storage object group based on the average temperature fluctuation amplitude, the average temperature tolerance range and the corresponding shelf life of each cold storage storage object group;

[0016] The temperature fluctuation acceptance range of the corresponding cold storage object group is set by the temperature sensitivity coefficient of each cold storage object group, wherein the temperature fluctuation acceptance range of each cold storage object group is the temperature fluctuation acceptance range of the cold storage storage objects of the type contained in the corresponding group.

[0017] As one of the preferred solutions, the historical temperature data of each type of cold storage storage object stored in the cold storage and the corresponding cold storage equipment operation data are obtained in sequence, and a trained cold storage temperature prediction model is obtained based at least on the temperature sensitivity coefficient and the cold storage equipment operation data, including:

[0018] Sequentially obtain the historical temperature data of each type of cold storage storage object stored in the cold storage and the corresponding cold storage equipment operation data, and combine the temperature sensitivity coefficient corresponding to each type of cold storage storage object to construct a time series data set;

[0019] After preprocessing, the time series data set is divided into a training set and a test set in proportion, and the long short-term memory neural network is trained as an input feature matrix and an output feature matrix respectively. The network structure and hyperparameters are selected by a grid search method, and the trained long short-term memory neural network is used as the cold storage temperature prediction model. The long short-term memory neural network includes an input layer, two hidden layers and an output layer.

[0020] As one of the preferred solutions, based on each of the temperature sensitivity coefficients, different initial control strategies are simulated and generated and input into the cold storage temperature prediction model for processing to obtain temperature change data under each initial control strategy, including:

[0021] Based on each of the temperature sensitivity coefficients, the cold storage equipment operation data of the corresponding type of the cold storage storage object is retrieved to simulate and generate different initial control strategies;

[0022] Inputting a plurality of different initial control strategies into the cold storage temperature prediction model for processing, and outputting temperature change data under each initial control strategy;

[0023] Based on each of the temperature sensitivity coefficients, the temperature change data under the corresponding initial control strategy is weighted averaged to obtain a comprehensive temperature prediction curve.

[0024] As one of the preferred solutions, comparing and analyzing each of the temperature change data with the corresponding temperature fluctuation acceptance range, and calculating the energy consumption data corresponding to the target strategy determined from the initial control strategy according to the analysis result, includes:

[0025] Comparing the comprehensive temperature prediction curve with the temperature fluctuation acceptance range of the corresponding type of the cold storage storage object, and selecting a target strategy that meets the temperature requirements of the cold storage storage object from the multiple initial control strategies according to the comparison result;

[0026] According to the cold storage equipment operation data corresponding to each target strategy, the energy consumption data of each target strategy is calculated.

[0027] As one of the preferred solutions, based on the energy consumption data, a candidate solution is selected from the target strategy, and based on the Arrhenius equation constructed by the historical temperature data corresponding to different types of cold storage objects, the mass loss index of the candidate solution is determined, including:

[0028] Sort the target strategies according to the energy consumption data, and use the target strategies whose energy consumption data does not exceed the energy consumption threshold as the candidate solutions;

[0029] An Arrhenius equation is constructed through characteristic parameters of different types of cold storage objects and their corresponding historical temperature fluctuation amplitudes to characterize the degree of influence on the mass loss of various types of cold storage objects;

[0030] The historical temperature fluctuation amplitude and characteristic parameters of the cold storage storage object corresponding to the candidate solution are input into the constructed Arrhenius equation for processing, and the mass loss index of each candidate solution is output.

[0031] As one of the preferred solutions, the candidate solutions are screened according to the genetic algorithm, and the cold storage equipment is controlled to execute the final control strategy determined, wherein the screening process is configured to use the fitness function constructed by the quality loss index corresponding to each candidate solution as the objective function for screening, including:

[0032] Performing weighted summation on the quality loss indexes of the candidate solutions to obtain a comprehensive score, and using the comprehensive score as a fitness function;

[0033] Encoding the start and stop time series of the candidate solutions into binary chromosomes, selecting individuals using a roulette wheel selection method, and generating new candidate solutions using single-point crossover and bit reversal mutation to optimize the candidate solutions;

[0034] With the minimum fitness function as the goal, the optimization process of the candidate solution is iterated, and the optimized candidate solution obtained when the fitness function reaches the optimum is used as the final control strategy to control the execution of the cold storage equipment.

[0035] As one of the preferred solutions, the candidate solutions are screened according to the genetic algorithm, and the cold storage equipment is controlled to execute the final control strategy determined, wherein the screening process is configured to use the fitness function constructed by the quality loss index corresponding to each candidate solution as the objective function for screening, and then includes:

[0036] Collecting real-time operation data and real-time energy consumption data of the cold storage equipment when executing the final control strategy, as well as real-time temperature data in the cold storage;

[0037] The real-time operation data, the real-time energy consumption data and the real-time temperature data are used as the state space, and the equipment start and stop actions in the final control strategy are used as the action space. A reward function is designed according to the temperature accuracy corresponding to the real-time energy consumption data and the real-time temperature data and the equipment life of the cold storage equipment to construct a deep reinforcement learning environment;

[0038] Training an agent in a deep reinforcement learning environment through a deep Q-network algorithm, taking the historical operation data and current operation data of the cold storage equipment as input, and outputting the latest control strategy;

[0039] The latest control strategy and the final control strategy are integrated to obtain a fusion strategy, and the exponential moving average method is used to adjust the weight of the fusion strategy to control the execution of the cold storage equipment.

[0040] As one of the preferred solutions, after fusing the latest control strategy and the final control strategy to obtain a fusion strategy, and adjusting the weight of the fusion strategy by using the exponential moving average method to control the execution of the cold storage equipment, the method further includes:

[0041] Collecting temperature change data of the cold storage equipment when executing the fusion strategy, and performing real-time analysis on the temperature change data through a multi-scale sliding time window algorithm to obtain a comprehensive temperature in each window, wherein the comprehensive temperature includes a maximum temperature and an average temperature;

[0042] When the comprehensive temperature exceeds the temperature fluctuation acceptance range corresponding to each cold storage storage object, the temperature deviation and the temperature change rate are determined according to the comprehensive temperature, and based on the temperature deviation and the temperature change rate, the start and stop actions in the fusion strategy are updated through a fuzzy adaptive PID control algorithm;

[0043] The updated fusion strategy is executed, and the comprehensive temperature analysis step is repeated until the maximum temperature and the average temperature do not exceed the temperature fluctuation acceptance range corresponding to each cold storage storage object.

[0044] The second aspect of the present invention provides an intelligent optimization and control system for cold storage equipment for storage objects, comprising:

[0045] A data determination module, used to obtain temperature sensitivity coefficients of different types of cold storage storage objects, and determine each temperature fluctuation acceptance range corresponding to each type of cold storage storage object based on each of the temperature sensitivity coefficients;

[0046] A model building module, used to sequentially obtain the historical temperature data of each type of cold storage storage object stored in the cold storage and the corresponding cold storage equipment operation data, and obtain a trained cold storage temperature prediction model based at least on the temperature sensitivity coefficient and the cold storage equipment operation data;

[0047] A data prediction module, used to simulate and generate different initial control strategies based on each of the temperature sensitivity coefficients and input them into the cold storage temperature prediction model for processing to obtain temperature change data under each initial control strategy;

[0048] A strategy selection module, used for comparing and analyzing each of the temperature change data with the corresponding temperature fluctuation acceptance range, and calculating the energy consumption data corresponding to the target strategy determined from the initial control strategy according to the analysis result;

[0049] An index quantification module, configured to select a candidate solution from the target strategy based on the energy consumption data, and determine a mass loss index of the candidate solution based on an Arrhenius equation constructed by using historical temperature data corresponding to different types of cold storage objects;

[0050] The strategy optimization module is used to screen the candidate solutions according to the genetic algorithm and control the cold storage equipment to execute the determined final control strategy, wherein the screening process is configured to use the fitness function constructed by the quality loss index corresponding to each candidate solution as the objective function for screening.

[0051] Compared with the prior art, the embodiments of the present invention have the following advantages:

[0052] (1) By obtaining the temperature sensitivity coefficients of different types of cold storage objects and determining their respective acceptable temperature fluctuation ranges, refined management of the temperatures of various storage objects is achieved, which helps to better protect temperature-sensitive items and reduce cargo losses caused by temperature fluctuations. The temperature prediction model is trained using historical temperature data and cold storage equipment operation data, and the temperature changes are predicted by simulating different control strategies for the cold storage equipment. The temperature changes are then compared and analyzed with the acceptable temperature fluctuation range, and the cold storage equipment control strategy that can meet the temperature requirements and reduce energy consumption can be identified, thereby significantly improving the energy efficiency of the cold storage and achieving energy conservation and emission reduction.

[0053] (2) Based on the temperature sensitivity coefficient of each storage object in the cold storage, a variety of control strategies are generated. Combined with the temperature fluctuation acceptance range of various cold storage storage objects, strategies that meet the temperature requirements of all cold storage storage objects are screened out. The energy consumption data of these strategies are calculated to provide a quantitative basis for the selection of strategies. Based on the characteristic parameters of the storage object itself, such as shelf life and temperature sensitivity, the Arrhenius equation is constructed to evaluate the mass loss index of the storage object under different control strategies of the cold storage equipment, so that the optimization process is closer to actual needs, aiming to find equipment control strategies that can both meet temperature requirements and minimize mass loss.

[0054] (3) The candidate solutions are optimized through genetic algorithms, and the fitness function is constructed through the quality loss index. With the help of the global search ability and adaptability of genetic algorithms, the optimal solution can be effectively found among multiple candidate solutions, that is, the equipment control strategy with the minimum quality loss, providing scientific and efficient decision support for cold storage management. BRIEF DESCRIPTION OF THE DRAWINGS

[0055] In order to more clearly illustrate the technical solution of the present invention, the drawings required for use in the implementation mode will be briefly introduced below. Obviously, the drawings described below are only some implementation modes of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0056] Figure 1 It is a flow chart of a method for intelligent optimization and control of cold storage equipment for storage objects provided by a certain embodiment of the present invention;

[0057] Figure 2 is a flowchart of step S1 provided in one embodiment of the present invention;

[0058] Figure 3 is a flowchart of step S2 provided by an embodiment of the present invention;

[0059] Figure 4 is a flowchart of step S3 provided by an embodiment of the present invention;

[0060] Figure 5 is a flowchart of step S4 provided by an embodiment of the present invention;

[0061] Figure 6 is a flowchart of step S5 provided by an embodiment of the present invention;

[0062] Figure 7 is a flowchart of step S6 provided in one embodiment of the present invention;

[0063] Figure 8 is a final control strategy optimization process diagram provided by a certain embodiment of the present invention;

[0064] Fig. 9 is a final control strategy optimization process diagram provided by another embodiment of the present invention;

[0065] Fig.10 It is a structural diagram of an intelligent optimization and control system for cold storage equipment for storage objects provided by a certain embodiment of the present invention. DETAILED DESCRIPTION

[0066] The following is a clear and complete description of the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings and embodiments. Obviously, the described embodiments are only part of the embodiments of the present invention, rather than all of the embodiments. The purpose of providing these embodiments is to make the disclosure of the present invention more thorough and comprehensive. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0067] In the description of this application, the terms "first", "second", "third", etc. are used for descriptive purposes only and are not to be understood as indicating or implying relative importance or implicitly indicating the number of the indicated technical features. Thus, a feature defined as "first", "second", "third", etc. may explicitly or implicitly include one or more of the feature. In the description of this application, unless otherwise specified, "plurality" means two or more.

[0068] In the description of the present application, it should be noted that, unless otherwise clearly specified and limited, the terms "installed", "connected", and "connected" should be understood in a broad sense. For example, it can be a fixed connection, a detachable connection, or an integral connection; it can be a mechanical connection or an electrical connection; it can be a direct connection, or it can be indirectly connected through an intermediate medium, or it can be a connection between the two components. The terms "vertical", "horizontal", "left", "right", "upper", "lower" and similar expressions used herein are for illustrative purposes only, and do not indicate or imply that the system or component referred to must have a specific orientation, be constructed and operated in a specific orientation, and therefore cannot be understood as a limitation on the present invention. The term "and / or" used herein includes any and all combinations of one or more related listed items. For those of ordinary skill in the art, the specific meanings of the above terms in this application can be understood according to specific circumstances.

[0069] In the description of this application, it should be noted that, unless otherwise defined, all technical and scientific terms used in the present invention have the same meaning as those commonly understood by those skilled in the art. The terms used in the specification of the present invention are only for the purpose of describing specific embodiments and are not intended to limit the present invention. For those of ordinary skill in the art, the specific meanings of the above terms in this application can be understood by specific circumstances.

[0070] In one embodiment, if Figure 1 As shown, the first aspect of the present invention provides a method for intelligent optimization and control of cold storage equipment for storage objects, comprising:

[0071] S1. Obtaining temperature sensitivity coefficients of different types of cold storage storage objects, and determining temperature fluctuation acceptance ranges and historical temperature tolerance ranges corresponding to different types of cold storage storage objects based on the temperature sensitivity coefficients;

[0072] In one embodiment, step S1 is as follows: Figure 2 As shown, including:

[0073] S11, obtaining historical temperature data of different types of cold storage objects stored in the cold storage, and constructing a feature vector based on the historical temperature data, the types of the cold storage objects and their shelf life, wherein the historical temperature data includes a historical temperature fluctuation amplitude and a historical temperature tolerance range;

[0074] S12, performing cluster analysis on the feature vectors by using a K-means algorithm to obtain a number of cold storage object groups, and calculating the average temperature tolerance range and the average temperature fluctuation amplitude of each of the cold storage object groups;

[0075] S13, quantifying the temperature sensitivity coefficient of each cold storage storage object group based on the average temperature fluctuation amplitude, average temperature tolerance range and corresponding shelf life of each cold storage storage object group;

[0076] S14. Setting a temperature fluctuation acceptance range of the corresponding cold storage object group by using the temperature sensitivity coefficient of each cold storage object group, wherein the temperature fluctuation acceptance range of each cold storage object group is the temperature fluctuation acceptance range of the cold storage objects of the type contained in the corresponding group.

[0077] Specifically, the objects stored in the cold storage include food, medicine, chemical preparations, some special materials of the power grid department, etc., that is, items that are sensitive to temperature and have temperature requirements for the storage environment. This application extracts the temperature fluctuation data of the objects stored in the past preset time from several cold storages. Taking food as an example, the temperature fluctuation data of various types of food can be collected within 7 days to calculate the temperature fluctuation amplitude of each food area, such as ±1.5℃ for meat area, ±2.0℃ for vegetable area, ±2.5℃ for fruit area, etc.; at the same time, the upper and lower limits of the temperature tolerance range and the shelf life of various storage objects are obtained from the storage object attribute database. Taking food as an example, the temperature tolerance range of meat is 0-4℃, and the shelf life is 7 days; vegetables are 4-8℃, and the shelf life is 14 days; fruits are 7-12℃, and the shelf life is 21 days. The storage object attribute database contains fields such as the type of storage object, the upper and lower limits of the temperature tolerance range, and the shelf life. Subsequently, a feature vector including the type of storage object, historical temperature fluctuation amplitude, historical temperature tolerance range, and shelf life is constructed, and the K-means clustering algorithm is used to perform cluster analysis on the feature vector. The number of clusters is set to 5, and the convergence condition is set to a moving distance from the cluster center that is less than 0.01. After iterative calculation, 5 cold storage object groups can be obtained: extremely sensitive, highly sensitive, moderately sensitive, lowly sensitive, and insensitive. In addition, the temperature sensitivity coefficients of different types of cold storage objects can also be obtained by table lookup; among them, the cold storage object group can contain only one type of storage object, such as a single cold storage object cold storage, or it can contain multiple types of storage objects, such as a composite cold storage containing medicines and chemical preparations, etc., which is determined according to the facility conditions and needs of the cold storage and is not specifically limited here.

[0078] Then calculate the average temperature tolerance range and average temperature fluctuation range in each cold storage storage object group, divide the average temperature fluctuation range of each group of cold storage storage objects by the average temperature tolerance range, and then multiply it by the inverse of its shelf life to obtain the temperature sensitivity coefficient of the group. For example, the average temperature fluctuation range of the extremely sensitive group is ±1.2℃, the average temperature tolerance range is 3℃, and the average shelf life is 5 days, so the temperature sensitivity coefficient is 0.08. The higher the temperature sensitivity coefficient, the more sensitive the group is to temperature fluctuations. Then set the temperature fluctuation acceptance range according to the temperature sensitivity coefficient, such as the temperature fluctuation acceptance range of the extremely sensitive group is ±0.5℃, the highly sensitive group is ±1.0℃, the moderately sensitive group is ±1.5℃, the low sensitive group is ±2.0℃, and the insensitive group is ±2.5℃. These range values ​​can also be converted into specific control parameters for the start and stop of cold storage equipment.

[0079] The present application constructs a feature vector through multi-dimensional information such as the type of storage object, shelf life, historical temperature fluctuation amplitude, and temperature tolerance range, so as to realize refined management and temperature control of each cold storage storage object in the cold storage, which is helpful to improve the quality and safety of the cold storage storage objects; the K-means algorithm is used to perform cluster analysis on the feature vector, and the cold storage storage objects with similar temperature sensitivity characteristics are divided into the same group, which can more accurately set the temperature conditions of different areas in the cold storage, optimize the storage layout, reduce energy consumption and improve the preservation effect of the cold storage storage objects; the quantification process of the temperature sensitivity coefficient of each cold storage storage object group can enable the management personnel to intuitively understand the sensitivity of different cold storage storage objects to temperature changes, so as to formulate a more scientific and reasonable temperature control strategy, and avoid the adverse effects on the quality of the cold storage storage objects due to excessive or too small temperature fluctuations; the temperature fluctuation acceptance range is dynamically adjusted through the temperature sensitivity coefficient of each cold storage storage object group, which can not only ensure that the cold storage storage objects are always in the most suitable temperature environment during the storage process, but also can be flexibly adjusted according to the characteristics of different cold storage storage objects, so as to maximize the shelf life of the cold storage storage objects and maintain their nutritional content and taste.

[0080] S2, sequentially obtaining historical temperature data of each type of the cold storage storage object stored in the cold storage and corresponding cold storage equipment operation data, and obtaining a trained cold storage temperature prediction model based at least on the temperature sensitivity coefficient and the cold storage equipment operation data;

[0081] In one embodiment, step S2 is as follows: Figure 3 As shown, including:

[0082] S21, sequentially obtaining historical temperature data of each type of cold storage storage object stored in the cold storage and corresponding cold storage equipment operation data, and combining the temperature sensitivity coefficient corresponding to each type of cold storage storage object to construct a time series data set;

[0083] S22. After preprocessing the time series data set, divide it into a training set and a test set in proportion, and use them as an input feature matrix and an output feature matrix to train the long short-term memory neural network, select the network structure and hyperparameters through a grid search method, and use the trained long short-term memory neural network as the cold storage temperature prediction model, wherein the long short-term memory neural network includes an input layer, two hidden layers and an output layer.

[0084] Specifically, the present application performs data sampling at intervals of 10 minutes to extract the temperature data of each type of cold storage storage object stored in the cold storage in the past year and the corresponding cold storage equipment operation data, wherein the cold storage equipment includes compressors, condensers, air coolers, etc., and the cold storage equipment operation data at least includes the working time and working frequency of the cold storage equipment. Taking the compressor as an example, the cold storage equipment operation data includes the compressor start time, stop time, operation frequency, operation time, etc.; then the temperature sensitivity coefficient corresponding to each type of cold storage storage object, the historical temperature data and the cold storage equipment operation data are combined into a time series data set, and the time series data set is preprocessed. The preprocessing process includes data standardization and filling in the data using linear interpolation. Missing values ​​and outliers are handled using the 3 times standard deviation method. Then the preprocessed data set is divided into training set and test set in a ratio of 8:2, and used as input feature matrix and output feature matrix to train the long short-term memory neural network. The grid search method is used to select the best network structure and hyperparameters, including the number of hidden layer nodes, learning rate and batch size. The Adam optimizer and mean square error loss function are used to train the model on the training set, and the early stopping mechanism is set. When the validation set loss does not decrease for 10 consecutive epochs, the training is stopped. The long short-term memory neural network after training is obtained and used as the cold storage temperature prediction model to predict the temperature change trend of various cold storage storage objects under different cold storage equipment operation data.

[0085] This application constructs a time series data set based on the temperature sensitivity coefficients of various cold storage objects, which improves the accuracy and reliability of the model in predicting the temperature changes of different types of cold storage objects. The long short-term memory neural network (LSTM) is used to process time series data, which can effectively capture the time characteristics and dynamic trends of cold storage temperature changes, provide a more scientific basis for the management of cold storage equipment, and achieve the purpose of energy saving and consumption reduction, reducing energy consumption fluctuations, while ensuring the safe storage of cold storage objects.

[0086] S3. Based on each of the temperature sensitivity coefficients, different initial control strategies are simulated and generated and input into the cold storage temperature prediction model for processing to obtain temperature change data under each initial control strategy;

[0087] In one embodiment, step S3 is as follows Figure 4 As shown, including:

[0088] S31, based on each of the temperature sensitivity coefficients, retrieve the cold storage equipment operation data of the corresponding type of the cold storage storage object to simulate and generate different initial control strategies;

[0089] S32, inputting a plurality of different initial control strategies into the cold storage temperature prediction model for processing, and outputting the temperature change data under each initial control strategy;

[0090] S33. Based on each of the temperature sensitivity coefficients, weighted averaging is performed on the temperature change data under the corresponding initial control strategy to obtain a comprehensive temperature prediction curve.

[0091] Specifically, the present application retrieves the cold storage equipment operation data of the corresponding type of cold storage storage object according to each calculated temperature sensitivity coefficient, summarizes it, summarizes the equipment start and stop rules of various cold storage storage objects, forms a variety of equipment control strategies and numbers them, so that each strategy has its corresponding strategy ID. Taking food and compressors as examples, the meat area adopts a strategy of 30 minutes on and 15 minutes off, and the vegetable area adopts a strategy of 40 minutes on and 20 minutes off, etc., wherein the specific form of the strategy ID is not specifically limited here; then the trained cold storage temperature prediction model is used to take these strategy sequences as input to predict the temperature change trend of the cold storage in the next 24 hours; and a corresponding temperature prediction curve is generated for each cold storage storage object, the curve includes a timestamp and a corresponding temperature value, and the root mean square error and the mean absolute percentage error of the prediction results can also be calculated to evaluate the prediction accuracy, and finally the prediction results are weighted averaged to obtain a comprehensive temperature prediction curve, and the comprehensive temperature prediction curve and the evaluation index are stored in the prediction result database for subsequent optimization of the equipment control strategy.

[0092] This application generates multiple control strategies based on the temperature sensitivity coefficients of various cold storage storage objects and their corresponding cold storage equipment operation data, ensuring that different types of cold storage storage objects can obtain temperature control that best suits their characteristics during the storage process; by inputting multiple control strategies into the trained cold storage temperature prediction model, the temperature changes under different strategies can be simulated. This process not only takes into account the working status of the cold storage equipment, but also combines the temperature sensitivity of the cold storage storage objects, and can more accurately predict the temperature changes in actual operation; through the comprehensive temperature prediction curve, managers can intuitively see the impact of different control strategies adopted by the equipment on the cold storage temperature, avoiding the limitations of traditional empirical judgments and improving the scientificity and accuracy of decision-making; by simulating the temperature changes of the cold storage equipment under different control strategies, and comprehensively considering the temperature sensitivity of the cold storage storage objects, a strategy that can meet the preservation requirements of the cold storage storage objects and reduce energy consumption can be selected, which helps to reduce unnecessary equipment operation time, reduce energy consumption costs, and achieve the goal of green energy saving.

[0093] S4, comparing and analyzing each of the temperature change data with the corresponding temperature fluctuation acceptance range, and calculating the energy consumption data corresponding to the target strategy determined from the initial control strategy according to the analysis result;

[0094] In one embodiment, step S4 is as follows: Figure 5 As shown, including:

[0095] S41, comparing the comprehensive temperature prediction curve with the temperature fluctuation acceptance range of the cold storage storage object of the corresponding type, and selecting a target strategy that meets the temperature requirements of the cold storage storage object from the multiple initial control strategies according to the comparison result;

[0096] S42. Calculate the energy consumption data of each target strategy according to the cold storage equipment operation data corresponding to each target strategy.

[0097] Specifically, based on the comprehensive temperature prediction curve, the temperature change trends of 100 different control strategies of cold storage equipment and their corresponding temperature fluctuation acceptance ranges are selected. Each strategy contains temperature data at 144 time points within 24 hours, and the temperature requirements of various cold storage storage objects, still taking food as an example, such as meat needs to be stored in an environment of 2-4℃, vegetables need to be stored in an environment of 4-8℃, etc., are compared and analyzed with the selected data, and a set of target control strategies that meet the temperature requirements of cold storage storage objects are screened out based on the comparison results; the start and stop frequency in the target control strategy is decomposed into start and stop cycles, and the energy data of each start and stop cycle is calculated in combination with the cold storage equipment operation data. The energy consumption data of each target control strategy can be obtained by accumulating the energy consumption values ​​of all cycles.

[0098] This application calculates the energy consumption data of each strategy through the cold storage equipment operation data corresponding to each target control strategy, so that managers can clearly understand the performance of different strategies in terms of energy consumption, and thus select the strategy with the lowest energy consumption and the highest efficiency for implementation, which helps to reduce the operating costs of the cold storage and improve economic benefits; the solution has high flexibility and adaptability. With the changes in the types of cold storage storage objects and the adjustment of storage needs, managers can adjust the temperature sensitivity coefficients and temperature fluctuation acceptance ranges of various types of cold storage storage objects according to actual conditions, and reselect the target control strategy to ensure that the cold storage can always maintain the best operating state and meet various storage needs.

[0099] S5. Based on the energy consumption data, select a candidate solution from the target strategy, and determine the mass loss index of the candidate solution based on the Arrhenius equation constructed by the historical temperature data corresponding to different types of cold storage objects;

[0100] In one embodiment, step S5 is as follows: Figure 6 As shown, including:

[0101] S51, sorting the target strategies according to the energy consumption data, and taking the target strategies whose energy consumption data does not exceed the energy consumption threshold as the candidate solutions;

[0102] S52, constructing an Arrhenius equation based on characteristic parameters of different types of cold storage objects and their corresponding historical temperature fluctuation amplitudes to characterize the degree of influence on the mass loss of the cold storage objects of various types;

[0103] S53, inputting the historical temperature fluctuation amplitude and characteristic parameters of the cold storage storage object corresponding to the candidate solution into the constructed Arrhenius equation for processing, and outputting the mass loss index of each candidate solution.

[0104] Specifically, based on the calculated energy consumption data, the target control strategies are ranked, and the target strategies whose energy consumption value does not exceed 120% of the minimum energy consumption are selected as candidate solutions; then the Arrhenius equation is used to establish the relationship between temperature and the quality degradation rate of cold storage objects, and a cold storage object quality impact model is constructed to characterize the impact degree on the quality loss of various types of cold storage objects, that is, Q=Q 0 exp[−Aexp(−Ea / RT)t], where Q is the mass of the cold storage object, Q 0 is the initial mass of the cold storage object, A is the prefactor (also called frequency factor), which is a constant independent of temperature but depends on the specific degradation process, Ea is the activation energy, which indicates the minimum energy required for the reactant molecules to react, and its unit is joule per mole (J / mol), R is the ideal gas constant, and its value is about 8.314J / (mol*K), T is the temperature, and t is the time. The change of the mass of the cold storage object at different temperatures over time can be measured experimentally, and then the values ​​of A and Ea can be estimated using methods such as nonlinear regression; the model inputs temperature fluctuation data and characteristic parameters of the cold storage object, such as activation energy, etc., and can output the mass loss index under each candidate scheme.

[0105] This application selects strategies whose energy consumption data does not exceed the energy consumption threshold as candidate solutions, which helps to reduce overall energy consumption and reduce operating costs. Setting the energy consumption threshold makes the solution flexible and helps to adapt to different operating scenarios and changes in demand. The Arrhenius equation is used to express the influence of the characteristic parameters of cold storage objects and the acceptable range of temperature fluctuations on the quality of cold storage objects, which can more comprehensively evaluate the changes in the quality of cold storage objects under different conditions. By evaluating the quality loss index, potential quality risks can be discovered in a timely manner, and corresponding measures can be taken to prevent and control them, which helps to ensure the safety and quality stability of cold storage objects.

[0106] S6. Screening the candidate solutions according to the genetic algorithm, and controlling the cold storage equipment to execute the determined final control strategy, wherein the screening process is configured to use the fitness function constructed by the quality loss index corresponding to each candidate solution as the objective function for screening;

[0107] In one embodiment, step S6 is as follows: Figure 7 As shown, including:

[0108] S61, performing weighted summation on the quality loss indexes of the candidate solutions to obtain a comprehensive score, and using the comprehensive score as a fitness function;

[0109] S62, encoding the start and stop time series of the candidate solutions into binary chromosomes, selecting individuals using a roulette wheel selection method, and generating new candidate solutions using single-point crossover and bit reversal mutation to optimize the candidate solutions;

[0110] S63, with the goal of minimizing the fitness function, iterating the optimization process of the candidate solutions, and taking the optimized candidate solutions obtained when the fitness function reaches the optimum as the final control strategy to control the execution of the cold storage equipment.

[0111] Specifically, this application obtains a comprehensive score by weighting and summing the quality loss index of the candidate solutions, where the weights can be determined according to the value and energy cost of the cold storage storage object; then a genetic algorithm is used to optimize the candidate solutions, with the comprehensive score as the fitness function, that is, with the goal of minimizing quality loss, the start and stop states of the candidate solutions at 144 time points within 24 hours are encoded into 144 binary chromosomes, the population size is set to 100, and individuals are selected using the roulette wheel selection method. New candidate solutions are generated using single-point crossover and bit reversal mutation, where the crossover probability is 0.8 and the mutation probability is 0.01. The convergence condition is set to a continuous 50 generations where the optimal fitness changes by less than 0.1% or reaches 1000 generations, and the start and stop time series corresponding to the optimal fitness is selected as the final control strategy for execution. In addition, a cold storage virtual environment can be constructed through digital twin technology, and the optimal start and stop time series is simulated for 24 hours, energy consumption, temperature fluctuations, and changes in the quality of cold storage storage objects are recorded, and its actual effect is evaluated and fine-tuned.

[0112] The present application uses the comprehensive score obtained by weighted summation of the quality loss index and the energy consumption value as the fitness function, which not only considers the economy of energy consumption, but also takes into account the importance of the quality of the cold storage storage objects, so that the final selected scheme is more comprehensive and reasonable, and the weights in the weighted summation can be adjusted according to actual needs to balance the priority between the quality of the cold storage storage objects and the energy consumption; encoding the start and stop time series of the candidate schemes into binary chromosomes is simple and easy to process, and the roulette wheel selection method is used to select individuals, so that individuals with higher fitness have a greater chance of being selected for genetic operations, thereby accelerating the optimization process, and using single-point crossover and bit reversal mutation to generate new candidate schemes can increase the diversity of the population, avoid premature convergence, and improve the global search capability of the optimization algorithm; through the iterative optimization process, the candidate schemes are continuously improved until the fitness function reaches the optimal level, so as to improve the accuracy and reliability of the optimization results.

[0113] In one embodiment, the final control strategy optimization process after step S6 is as follows: Figure 8 As shown, including:

[0114] S71, collecting real-time operation data and real-time energy consumption data of the cold storage equipment when executing the final control strategy, as well as real-time temperature data in the cold storage;

[0115] S72, taking the real-time operation data, the real-time energy consumption data and the real-time temperature data as the state space, and taking the equipment start and stop action in the final control strategy as the action space, designing a reward function according to the temperature accuracy corresponding to the real-time energy consumption data and the real-time temperature data and the equipment life of the cold storage equipment, so as to construct a deep reinforcement learning environment;

[0116] S73, training an intelligent agent in a deep reinforcement learning environment by a deep Q network algorithm, taking the historical operation data and current operation data of the cold storage equipment as input, and outputting the latest control strategy;

[0117] S74, fusing the latest control strategy and the final control strategy to obtain a fusion strategy, and using the exponential moving average method to adjust the weight of the fusion strategy to control the execution of the cold storage equipment.

[0118] This application uses the data after the final control strategy is executed to re-optimize, update and execute it. Specifically, it collects the operating data of the cold storage equipment in the process of executing the final control strategy in real time, including the start and stop time, operating current, temperature and refrigerant flow of the cold storage equipment, and records its energy consumption data and temperature changes in the cold storage; then constructs a deep reinforcement learning environment, takes the equipment operating status, energy consumption data, cold storage temperature and external environmental parameters as the state space, and takes the equipment start and stop action on or off as the action space, and designs a reward function, in which energy consumption accounts for 40%, temperature control accuracy accounts for 40%, and equipment life accounts for 20%. Through normalization, each factor is unified into the range of 0-1; then a deep Q network is used to train the intelligent agent. The network contains 3 hidden layers with 64, 32, and 16 nodes respectively, with historical operating data and current operating data. The data is input, the latest control strategy is output, and the experience pool of size 10,000 is used to store state transition samples. 128 samples are randomly sampled for batch learning each time training. The target network is updated every 100 trainings, and the policy network parameters are updated every 4 hours. Finally, the latest control strategy output by the deep reinforcement learning algorithm is integrated with the original optimal time series. The exponential moving average method is used to give the original strategy a weight of 0.7 and the latest strategy a weight of 0.3. The dynamic time window method is used to adjust the window size adaptively according to the temperature change rate, ranging from 2 to 8 hours. When the change rate is greater than 0.5℃ / h, the window is reduced to 2 hours; when it is less than 0.1℃ / h, it is expanded to 8 hours. The model is retrained every 30 days, and the learning rate and discount factor are adjusted according to the average reward value of the last 7 days. In addition, an abnormality detection mechanism based on statistical process control can be added. When the monitoring index exceeds the range of 3 times the standard deviation, the emergency control logic is triggered and temporarily switched to the preset conservative strategy.

[0119] This application collects real-time data after the cold storage equipment executes the final control strategy to update the plan, ensuring the rapid response and accuracy of the strategy adjustment; constructs the real-time data as a state space, and uses the equipment start and stop actions as the action space, combined with the reward function design, to construct a deep reinforcement learning environment, which can automatically learn and optimize the cold storage equipment control strategy without the need for frequent manual adjustment of parameters, thereby improving the system's adaptability and intelligence level; when designing the reward function, multiple dimensions such as real-time energy consumption, temperature accuracy, and cold storage equipment life are taken into consideration, which helps to reduce operating costs while ensuring the quality of cold storage operation; the strategy update mechanism based on historical data and current status can continuously learn and adapt to changes in the cold storage operating environment, thereby improving the effectiveness and accuracy of the strategy.

[0120] In one embodiment, the final control strategy optimization process is as follows: Fig. 9 As shown, it also includes:

[0121] S75, collecting temperature change data when the cold storage equipment executes the fusion strategy, and performing real-time analysis on the temperature change data through a multi-scale sliding time window algorithm to obtain a comprehensive temperature in each window, wherein the comprehensive temperature includes a maximum temperature and an average temperature;

[0122] S76, when the comprehensive temperature exceeds the temperature fluctuation acceptance range corresponding to each of the cold storage storage objects, determine the temperature deviation and the temperature change rate according to the comprehensive temperature, and based on the temperature deviation and the temperature change rate, update the start and stop actions in the fusion strategy through a fuzzy adaptive PID control algorithm;

[0123] S77, executing the updated fusion strategy, and repeating the comprehensive temperature analysis step until the maximum temperature and the average temperature do not exceed the temperature fluctuation acceptance range corresponding to each of the cold storage storage objects.

[0124] Specifically, the present application optimizes and updates the data after the fusion strategy is executed, and executes it. Specifically, the temperature change data of the cold storage equipment after the fusion strategy is executed is collected, that is, the temperature change data, and a multi-scale sliding time window algorithm is used to perform real-time analysis on the collected temperature data using windows of 5 minutes, 10 minutes and 30 minutes respectively, sliding once every 1 minute, calculating the average temperature, maximum temperature and minimum temperature in each window, and comprehensively judging whether the temperature change trend exceeds the acceptable range of temperature fluctuations of the cold storage storage object. For example, the suitable temperature for beef is 0-4°C, with a maximum fluctuation of ±1°C. The 5-minute window detects that the temperature rises from 2°C to 5°C, the 10-minute window shows an average temperature of 3.5°C, and the 30-minute window trend is stable, which is comprehensively judged to be a short-term fluctuation.

[0125] If it is detected that the temperature change data exceeds the acceptable range of temperature fluctuation of the cold storage storage object, the early warning mechanism is immediately triggered, and an emergency adjustment instruction is sent to the refrigeration control unit. The adaptive PID control combined with the fuzzy control algorithm is used to dynamically adjust the equipment control strategy. According to the size and change rate of the temperature deviation, the PID parameters are dynamically adjusted using fuzzy rules. The fuzzy rules in the fuzzy rule base are established according to the relevant historical temperature data and their corresponding strategies, which are: when the temperature deviation is greater than the deviation threshold and the change rate is greater than the change threshold, Kp is adaptively increased according to the preset ratio, and the new start and stop time points are determined accordingly. Taking the aforementioned beef as an example, when the temperature deviation is greater than 1°C and the change rate exceeds 0.2°C / minute, the PID parameter Kp is adjusted from 0.5 to 0.8, Ki is adjusted from 0.1 to 0.15, and Kd is adjusted from 0.05 to 0.08, and the operation time of the cold storage equipment is increased from 30 minutes / hour to 40 minutes / hour. Repeat the start and stop time update process until there is no overtemperature phenomenon in multiple consecutive iterations. It should be noted that the thresholds, presets and other values ​​involved in this scheme can be determined based on historical data and actual needs, and no specific numerical limits are made here.

[0126] The detailed information of the over-temperature event, including the time of occurrence, duration, maximum deviation value, adjusted equipment control strategy and PID parameters, can then be recorded in the abnormal event database, and the historical data in the abnormal event database can be analyzed using seasonal and trend decomposition algorithms to identify potential periodic problems or systemic defects. At the same time, the effect of the adjusted strategy is evaluated, its temperature stability index and energy consumption index are calculated, and the evaluation results are fed back to the optimization algorithm to dynamically adjust the control parameters and abnormal detection thresholds to form a closed-loop control.

[0127] This application processes temperature change data through a multi-scale sliding time window algorithm, which can capture the trend and characteristics of temperature changes; when the comprehensive temperature exceeds the acceptable range of the cold storage storage object grouping, based on the temperature deviation and temperature change rate, the fuzzy adaptive PID control algorithm is used to update the start and stop time points in the fusion strategy, and the control parameters are dynamically adjusted according to the actual temperature change, so that the strategy adjustment is more accurate and faster, which helps to better maintain the temperature stability in the cold storage; the entire solution forms a closed-loop feedback system, which realizes the continuous optimization of the cold storage equipment control strategy by continuously collecting data, analyzing data, adjusting strategies and executing strategies, so as to continuously adapt to changes in the environment in the cold storage and improve the overall performance and efficiency of the system.

[0128] In one embodiment, the equipment energy consumption data and cold storage temperature change data before and after the fusion strategy optimization are obtained, and the seasonal decomposition method is used to process the periodic impact of these data to calculate the average daily energy consumption and temperature standard deviation; the paired t-test method is used to evaluate the significance of the optimization strategy, and the p-value and effect size are obtained. According to the difference in energy consumption before and after optimization and the degree of improvement in temperature stability, the weights of the deep neural network model are dynamically adjusted using a non-dominated sorting genetic algorithm, and the population size is set to 100, and iterated 500 times. With energy consumption reduction and temperature stability improvement as the objective function, a new solution set is generated through crossover and mutation operations to obtain energy consumption and temperature stability. The Pareto frontier selects a solution that reduces energy consumption by 13% and temperature fluctuations by 25% as the optimal solution. Then, external environmental data, including outdoor temperature, humidity, and grid load, is collected, combined with the operating data after executing the optimal solution, and the online stochastic gradient descent algorithm is used to update the deep neural network model to dynamically adjust the feature weights of the input layer and the number of neurons in the hidden layer. The pre-trained ResNet50 model is used for transfer learning, the first 40 layers are frozen, and only the parameters of the last 10 layers are fine-tuned. 50 epochs are trained on 100 cold storage samples to recalculate the optimal equipment control strategy. At the same time, the virtual energy storage capacity is adjusted according to the grid load.

[0129] In another embodiment, the optimized device control strategy and its key parameters are uploaded to the cloud server through a secure encrypted channel, and the ApacheCassandra distributed database storage structure is used to achieve high availability and fast reading and writing of data, and a data backup and disaster recovery mechanism is set up at the same time; at the same time, the React.js framework and the Echarts chart library are used to build a Web visualization interface, and a responsive design is used to adapt to different terminal devices. Real-time data push is realized through WebSocket to display the real-time operation status data and real-time energy consumption data of cold storage equipment. In addition, based on the ARIMA model, time series analysis and mining of historical temperature, operation status and other data are performed to generate an optimization effect report including energy saving ratio, temperature stability improvement degree and economic benefit analysis, and it is intuitively presented through data visualization charts. At the same time, a remote control and early warning module is developed, and the Drools rule engine is integrated for anomaly detection. Specific rules such as "triggering an early warning when the temperature exceeds the threshold and lasts for 5 minutes" are written and set. When the monitoring index exceeds the preset threshold, the early warning is automatically triggered, and the early warning threshold is dynamically adjusted according to the optimization effect report.

[0130] In the embodiment of the present application, based on the problem of how to intelligently optimize and regulate the control strategy of cold storage equipment while taking into account the needs of energy saving and efficiency improvement, a cold storage equipment intelligent optimization and regulation method for storage objects is designed. By obtaining the temperature sensitivity coefficients of different types of cold storage storage objects and determining their respective temperature fluctuation acceptance ranges based on this, the method realizes refined management of the temperatures of various types of storage objects, helps to better protect temperature-sensitive items and reduce cargo losses caused by temperature fluctuations; the temperature prediction model is trained using historical temperature data and cold storage equipment operation data, and the temperature changes are predicted by simulating different control strategies for the cold storage equipment, and then compared and analyzed with the temperature fluctuation acceptance range, the equipment control strategy that can meet the temperature requirements and reduce energy consumption can be identified, thereby significantly improving the energy efficiency of the cold storage and achieving energy saving and emission reduction; based on the temperature sensitivity coefficients of each storage object in the cold storage coefficient, generate a variety of control strategies, and combine the temperature fluctuation acceptance range of various cold storage objects to screen out strategies that meet the temperature requirements of all cold storage objects, and provide a quantitative basis for strategy selection by calculating the energy consumption data of these strategies; construct the Arrhenius equation based on the characteristic parameters of the storage object itself, such as shelf life, temperature sensitivity, etc., to evaluate the mass loss index of the storage object under different control strategies of the cold storage equipment, so that the optimization process is closer to actual needs, aiming to find equipment control strategies that can meet temperature requirements and minimize quality loss; optimize the candidate solutions through genetic algorithms, and construct a fitness function through the quality loss index, with the help of the global search ability and adaptability of genetic algorithms, to effectively find the optimal solution among multiple candidate solutions, that is, the equipment control strategy with the smallest quality loss, providing scientific and efficient decision support for cold storage management.

[0131] It should be noted that although the steps in the above flowchart are shown in sequence according to the arrows, these steps are not necessarily executed in the order indicated by the arrows. Unless otherwise specified in this document, there is no strict order restriction for the execution of these steps, and these steps can be executed in other orders.

[0132] In another embodiment, if Fig.10 As shown, the second aspect of the present invention provides an intelligent optimization and control system for cold storage equipment for storage objects, comprising:

[0133] The data determination module 10 is used to obtain temperature sensitivity coefficients of different types of cold storage storage objects, and determine the temperature fluctuation acceptance ranges corresponding to the different types of cold storage storage objects based on the temperature sensitivity coefficients;

[0134] The model building module 20 is used to sequentially obtain the historical temperature data of each type of the cold storage storage object stored in the cold storage and the corresponding cold storage equipment operation data, and obtain a trained cold storage temperature prediction model based at least on the temperature sensitivity coefficient and the cold storage equipment operation data;

[0135] The data prediction module 30 is used to simulate and generate different initial control strategies based on each of the temperature sensitivity coefficients and input them into the cold storage temperature prediction model for processing to obtain temperature change data under each initial control strategy;

[0136] A strategy selection module 40 is used to compare and analyze each of the temperature change data with the corresponding temperature fluctuation acceptance range, and calculate the energy consumption data corresponding to the target strategy determined from the initial control strategy according to the analysis result;

[0137] An index quantification module 50 is used to select a candidate solution from the target strategy based on the energy consumption data, and determine the mass loss index of the candidate solution based on the Arrhenius equation constructed by the historical temperature data corresponding to different types of cold storage objects;

[0138] The strategy optimization module 60 is used to screen the candidate solutions according to the genetic algorithm and control the cold storage equipment to execute the determined final control strategy, wherein the screening process is configured to use the fitness function constructed by the quality loss index corresponding to each candidate solution as the objective function for screening.

[0139] It should be noted that each module in the above-mentioned intelligent optimization and control system for cold storage equipment based on storage objects can be fully or partially implemented by software, hardware and their combination. The above-mentioned modules can be embedded in or independent of the processor in the computer device in hardware form, or can be stored in the memory of the computer device in software form, so that the processor can call and execute the operations corresponding to the above modules. For the specific definition of an intelligent optimization and control system for cold storage equipment for storage objects, please refer to the definition of an intelligent optimization and control method for cold storage equipment for storage objects in the above text. The two have the same functions and effects, which will not be repeated here.

[0140] In summary, the present invention relates to the field of information technology, and discloses a method and system for intelligent optimization and control of cold storage equipment for storage objects. The method obtains the temperature sensitivity coefficients of different types of cold storage storage objects, and determines the respective temperature fluctuation acceptance ranges accordingly, and uses historical data to construct a temperature prediction model to predict the temperature change trend of the cold storage equipment under different control strategies. Based on the prediction results and the characteristic parameters of the cold storage storage objects, the cold storage equipment control strategies that meet the temperature requirements are screened, and their energy consumption is evaluated to construct the Arrhenius equation to characterize the mass loss index of the cold storage storage objects in the candidate schemes, and the candidate strategies are optimized in combination with the genetic algorithm, and the fitness function of the genetic algorithm is constructed through the mass loss index, so as to determine the optimal control strategy through the genetic algorithm and control the execution of the cold storage equipment, thereby realizing precise control of the cold storage temperature and energy efficiency optimization, and improving the intelligent level of cold storage management.

[0141] Each embodiment in this specification is described in a progressive manner, and the same or similar parts of each embodiment can be directly referred to each other, and each embodiment focuses on the differences from other embodiments. In particular, for the system embodiment, since it is basically similar to the method embodiment, the description is relatively simple, and the relevant parts can be referred to the partial description of the method embodiment. It should be noted that the technical features of the above-mentioned embodiments can be combined arbitrarily. In order to make the description concise, all possible combinations of the technical features in the above-mentioned embodiments are not described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0142] The above-mentioned embodiments only express several preferred implementation modes of the present application, and the descriptions thereof are relatively specific and detailed, but they cannot be understood as limiting the scope of the invention patent. It should be pointed out that, for ordinary technicians in the technical field, several improvements and substitutions can be made without departing from the technical principles of the present invention, and these improvements and substitutions should also be regarded as the protection scope of the present application. Therefore, the protection scope of the patent of the present application shall be based on the protection scope of the claims.

Claims

1. A method for intelligent optimization and control of cold storage equipment for storage objects, characterized in that: include: Acquire temperature sensitivity coefficients of different types of cold storage storage objects, and determine temperature fluctuation acceptance ranges corresponding to the different types of cold storage storage objects based on the temperature sensitivity coefficients; Sequentially obtain historical temperature data of each type of cold storage storage object stored in the cold storage and corresponding cold storage equipment operation data, and obtain a trained cold storage temperature prediction model based at least on the temperature sensitivity coefficient and the cold storage equipment operation data; Based on each of the temperature sensitivity coefficients, different initial control strategies are simulated and generated and input into the cold storage temperature prediction model for processing to obtain temperature change data under each initial control strategy; Compare and analyze each of the temperature change data with the corresponding temperature fluctuation acceptance range, and calculate the energy consumption data corresponding to the target strategy determined from the initial control strategy according to the analysis result; Based on the energy consumption data, a candidate solution is selected from the target strategy, and a mass loss index of the candidate solution is determined based on an Arrhenius equation constructed by historical temperature data corresponding to different types of cold storage objects; The candidate solutions are screened according to a genetic algorithm, and the cold storage equipment is controlled to execute the final control strategy determined, wherein the screening process is configured to use the fitness function constructed by the quality loss index corresponding to each candidate solution as the objective function for screening; The step of obtaining temperature sensitivity coefficients of different types of cold storage objects, and determining temperature fluctuation acceptance ranges corresponding to the different types of cold storage objects based on the temperature sensitivity coefficients, includes: Acquire historical temperature data of different types of cold storage objects stored in the cold storage, and construct a feature vector based on the historical temperature data, the types of the cold storage objects and their shelf life, wherein the historical temperature data includes a historical temperature fluctuation amplitude and a historical temperature tolerance range; Performing cluster analysis on the feature vectors by using a K-means algorithm to obtain a number of cold storage object groups, and calculating the average temperature tolerance range and average temperature fluctuation amplitude of each of the cold storage object groups; quantifying the temperature sensitivity coefficient of each cold storage storage object group based on the average temperature fluctuation amplitude, the average temperature tolerance range and the corresponding shelf life of each cold storage storage object group; The temperature fluctuation acceptance range of the corresponding cold storage object group is set by the temperature sensitivity coefficient of each cold storage object group, wherein the temperature fluctuation acceptance range of each cold storage object group is the temperature fluctuation acceptance range of the cold storage objects of the type contained in the corresponding group; The selecting a candidate solution from the target strategy based on the energy consumption data, and determining the mass loss index of the candidate solution based on the Arrhenius equation constructed by the historical temperature data corresponding to different types of cold storage objects, includes: Sort the target strategies according to the energy consumption data, and use the target strategies whose energy consumption data does not exceed the energy consumption threshold as the candidate solutions; An Arrhenius equation is constructed through characteristic parameters of different types of cold storage objects and their corresponding historical temperature fluctuation amplitudes to characterize the degree of influence on the mass loss of various types of cold storage objects; Input the historical temperature fluctuation amplitude and characteristic parameters of the cold storage storage object corresponding to the candidate solution into the constructed Arrhenius equation for processing, and output the mass loss index of each candidate solution; The candidate solutions are screened according to the genetic algorithm, and the cold storage equipment is controlled to execute the final control strategy determined, wherein the screening process is configured to use the fitness function constructed by the quality loss index corresponding to each candidate solution as the objective function for screening, including: Performing weighted summation on the quality loss index and energy consumption data of the candidate solutions to obtain a comprehensive score, and using the comprehensive score as a fitness function; Encoding the start and stop time series of the candidate solutions into binary chromosomes, selecting individuals using a roulette wheel selection method, and generating new candidate solutions using single-point crossover and bit reversal mutation to optimize the candidate solutions; With the minimum fitness function as the goal, the optimization process of the candidate solution is iterated, and the optimized candidate solution obtained when the fitness function reaches the optimum is used as the final control strategy to control the execution of the cold storage equipment.

2. According to claim 1, a method for intelligent optimization and control of cold storage equipment for storage objects is characterized in that: The step of sequentially acquiring historical temperature data of each type of cold storage storage object stored in the cold storage and corresponding cold storage equipment operation data, and obtaining a trained cold storage temperature prediction model based at least on the temperature sensitivity coefficient and the cold storage equipment operation data, includes: Sequentially obtain the historical temperature data of each type of cold storage storage object stored in the cold storage and the corresponding cold storage equipment operation data, and combine the temperature sensitivity coefficient corresponding to each type of cold storage storage object to construct a time series data set; After preprocessing, the time series data set is divided into a training set and a test set in proportion, and the long short-term memory neural network is trained as an input feature matrix and an output feature matrix respectively. The network structure and hyperparameters are selected by a grid search method, and the trained long short-term memory neural network is used as the cold storage temperature prediction model. The long short-term memory neural network includes an input layer, two hidden layers and an output layer.

3. According to claim 1, a method for intelligent optimization and control of cold storage equipment for storage objects is characterized in that: The candidate solutions are screened according to the genetic algorithm, and the cold storage equipment is controlled to execute the final control strategy determined, wherein the screening process is configured to use the fitness function constructed by the quality loss index corresponding to each candidate solution as the objective function for screening, and then includes: Collecting real-time operation data and real-time energy consumption data of the cold storage equipment when executing the final control strategy, as well as real-time temperature data in the cold storage; The real-time operation data, the real-time energy consumption data and the real-time temperature data are used as the state space, and the equipment start and stop actions in the final control strategy are used as the action space. A reward function is designed according to the temperature accuracy corresponding to the real-time energy consumption data and the real-time temperature data and the equipment life of the cold storage equipment to construct a deep reinforcement learning environment; Training an agent in a deep reinforcement learning environment through a deep Q-network algorithm, taking the historical operation data and current operation data of the cold storage equipment as input, and outputting the latest control strategy; The latest control strategy and the final control strategy are integrated to obtain a fusion strategy, and the exponential moving average method is used to adjust the weight of the fusion strategy to control the execution of the cold storage equipment.

4. According to claim 3, a method for intelligent optimization and control of cold storage equipment for storage objects is characterized in that: After fusing the latest control strategy and the final control strategy to obtain a fusion strategy, and adjusting the weight of the fusion strategy by using an exponential moving average method to control the execution of the cold storage equipment, the method further includes: Collecting temperature change data of the cold storage equipment when executing the fusion strategy, and performing real-time analysis on the temperature change data through a multi-scale sliding time window algorithm to obtain a comprehensive temperature in each window, wherein the comprehensive temperature includes a maximum temperature and an average temperature; When the comprehensive temperature exceeds the temperature fluctuation acceptance range corresponding to each cold storage storage object, the temperature deviation and the temperature change rate are determined according to the comprehensive temperature, and based on the temperature deviation and the temperature change rate, the start and stop actions in the fusion strategy are updated through a fuzzy adaptive PID control algorithm; The updated fusion strategy is executed, and the comprehensive temperature analysis step is repeated until the maximum temperature and the average temperature do not exceed the temperature fluctuation acceptance range corresponding to each cold storage storage object.

5. An intelligent optimization and control system for cold storage equipment for storage objects, characterized in that: include: A data determination module, used to obtain temperature sensitivity coefficients of different types of cold storage storage objects, and determine each temperature fluctuation acceptance range corresponding to the different types of cold storage storage objects based on each of the temperature sensitivity coefficients; A model building module, used to sequentially obtain the historical temperature data of each type of cold storage storage object stored in the cold storage and the corresponding cold storage equipment operation data, and obtain a trained cold storage temperature prediction model based at least on the temperature sensitivity coefficient and the cold storage equipment operation data; A data prediction module, used to simulate and generate different initial control strategies based on each of the temperature sensitivity coefficients and input them into the cold storage temperature prediction model for processing to obtain temperature change data under each initial control strategy; A strategy selection module, used for comparing and analyzing each of the temperature change data with the corresponding temperature fluctuation acceptance range, and calculating the energy consumption data corresponding to the target strategy determined from the initial control strategy according to the analysis result; An index quantification module, configured to select a candidate solution from the target strategy based on the energy consumption data, and determine a mass loss index of the candidate solution based on an Arrhenius equation constructed by using historical temperature data corresponding to different types of cold storage objects; A strategy optimization module, for screening the candidate solutions according to a genetic algorithm, and controlling the cold storage equipment to execute the determined final control strategy, wherein the screening process is configured to use the fitness function constructed by the quality loss index corresponding to each candidate solution as the objective function for screening; The step of obtaining temperature sensitivity coefficients of different types of cold storage objects and determining temperature fluctuation acceptance ranges corresponding to the different types of cold storage objects based on the temperature sensitivity coefficients includes: Acquire historical temperature data of different types of cold storage objects stored in the cold storage, and construct a feature vector based on the historical temperature data, the types of the cold storage objects and their shelf life, wherein the historical temperature data includes a historical temperature fluctuation amplitude and a historical temperature tolerance range; Performing cluster analysis on the feature vectors by using a K-means algorithm to obtain a number of cold storage object groups, and calculating the average temperature tolerance range and average temperature fluctuation amplitude of each of the cold storage object groups; quantifying the temperature sensitivity coefficient of each cold storage storage object group based on the average temperature fluctuation amplitude, the average temperature tolerance range and the corresponding shelf life of each cold storage storage object group; The temperature fluctuation acceptance range of the corresponding cold storage object group is set by the temperature sensitivity coefficient of each cold storage object group, wherein the temperature fluctuation acceptance range of each cold storage object group is the temperature fluctuation acceptance range of the cold storage objects of the type contained in the corresponding group; The selecting a candidate solution from the target strategy based on the energy consumption data, and determining the mass loss index of the candidate solution based on the Arrhenius equation constructed by the historical temperature data corresponding to different types of cold storage objects, includes: Sort the target strategies according to the energy consumption data, and use the target strategies whose energy consumption data does not exceed the energy consumption threshold as the candidate solutions; An Arrhenius equation is constructed through characteristic parameters of different types of cold storage objects and their corresponding historical temperature fluctuation amplitudes to characterize the degree of influence on the mass loss of various types of cold storage objects; Input the historical temperature fluctuation amplitude and characteristic parameters of the cold storage storage object corresponding to the candidate solution into the constructed Arrhenius equation for processing, and output the mass loss index of each candidate solution; The candidate solutions are screened according to the genetic algorithm, and the cold storage equipment is controlled to execute the final control strategy determined, wherein the screening process is configured to use the fitness function constructed by the quality loss index corresponding to each candidate solution as the objective function for screening, including: Performing weighted summation on the quality loss index and energy consumption data of the candidate solutions to obtain a comprehensive score, and using the comprehensive score as a fitness function; Encoding the start and stop time series of the candidate solutions into binary chromosomes, selecting individuals using a roulette wheel selection method, and generating new candidate solutions using single-point crossover and bit reversal mutation to optimize the candidate solutions; With the minimum fitness function as the goal, the optimization process of the candidate solution is iterated, and the optimized candidate solution obtained when the fitness function reaches the optimum is used as the final control strategy to control the execution of the cold storage equipment.

Images