Data center liquid cooling uniform cooling system

Abstract

The invention discloses a liquid cooling uniform cooling system for a data center. The method comprises the steps of obtaining real-time monitoring data of a data center, performing data preprocessing on the real-time monitoring data, establishing an initial Bi LSTM liquid cooling system temperature prediction model by utilizing a Bi LSTM bidirectional long-short-term memory network, performing optimization on hyper-parameters of the initial Bi LSTM liquid cooling system temperature prediction model by utilizing an ISCSO (International Standard Convergence Optimization) improved Sodat swarm algorithm, and performing optimization on the optimal hyper-parameters of the initial Bi LSTM liquid cooling system temperature prediction model. And inputting the training monitoring data into the I SCSO-B i LSTM liquid cooling system temperature prediction model for training, calculating the temperature drop rate of the data center according to the data center temperature, and if the temperature drop rate is greater than a preset rate, generating a target liquid cooling system working temperature to perform target temperature control on the liquid cooling uniform cooling system. The working temperature of the liquid cooling system is monitored and controlled in real time, the cooling strategy is rapidly adjusted according to the actual situation, and the cooling efficiency is effectively improved.

Description

Technical Field

[0001] The invention relates to the field of data centers, and in particular to a liquid cooling uniform cooling system for a data center. Background Art

[0002] With the development of information technology, data centers have higher and higher requirements for cooling systems. Traditional cooling methods often fail to meet the heat dissipation needs of high-density server clusters, and there are problems such as high energy consumption and uneven temperature control. If the local temperature rate drops too fast, a large temperature gradient will be generated inside the liquid cooling system. This temperature difference may cause local overcooling, which in turn affects the normal operation of electronic equipment and even causes hardware damage. If the local temperature rate drops too slowly, the temperature of the data center will not drop enough, and the data center will be overheated. Therefore, how to uniformly control the temperature of the data center is a technical problem that needs to be solved at this stage. Summary of the invention

[0003] The purpose of the present invention is to solve the above problems and designs a data center liquid cooling uniform cooling system.

[0004] To achieve the above object, the technical solution of the present invention is that, further, in the above data center liquid cooling uniform cooling system, the data center liquid cooling uniform cooling system comprises the following modules:

[0005] A detection data acquisition module is used to acquire real-time monitoring data of the data center, and perform data preprocessing on the real-time monitoring data to obtain training monitoring data;

[0006] A prediction model establishment module is used to establish an initial Bi LSTM liquid cooling system temperature prediction model by using a Bi LSTM bidirectional long short-term memory network, and to optimize the hyperparameters of the initial Bi LSTM liquid cooling system temperature prediction model by using an ISCSSO-improved sand cat swarm algorithm to obtain an ISCSSO-B i LSTM liquid cooling system temperature prediction model;

[0007] An operating temperature prediction module, used for inputting the training monitoring data into the ISCS0-B i LSTM liquid cooling system temperature prediction model for training to obtain an initial liquid cooling system operating temperature;

[0008] A temperature rate calculation module, used to perform real-time temperature control on the liquid cooling uniform cooling system based on the initial liquid cooling system operating temperature, and obtain the controlled data center temperature;

[0009] The target temperature control module is used to calculate the temperature drop rate of the data center according to the temperature of the data center. If the temperature drop rate is greater than a preset rate, a target liquid cooling system operating temperature is generated to perform target temperature control on the liquid cooling uniform cooling system.

[0010] Furthermore, in the above data center liquid cooling uniform cooling system, the detection data acquisition module includes the following submodules:

[0011] The acquisition submodule is used to acquire real-time monitoring data of the data center, wherein the real-time monitoring data includes at least the server CPU usage rate, memory usage ratio, coolant flow rate, ambient temperature and ambient humidity of the data center;

[0012] A processing submodule, used to delete outliers and fill missing values ​​in the real-time monitoring data to obtain complete monitoring data;

[0013] A submodule is obtained, which is used to perform standardization processing on the complete monitoring data to obtain training monitoring data.

[0014] Furthermore, in the above data center liquid cooling uniform cooling system, it is characterized in that the prediction model building module includes the following submodules:

[0015] Establish a submodule for establishing an initial Bi LSTM liquid cooling system temperature prediction model using a Bi LSTM bidirectional long short-term memory network;

[0016] The setting submodule is used to set the parameters of the SCSO sand cat swarm algorithm and use the K-means clustering algorithm to initialize the population of the parameters of the sand cat swarm algorithm;

[0017] Substitute the submodule into the objective function to obtain the fitness value of each sand cat, and determine the mutation probability of the current iteration of the SCSO sand cat swarm algorithm and the parameters of the evolution process according to the fitness value;

[0018] A judgment submodule is used to judge the mutation probability. If it is less than the mutation probability, random walk is performed, and further evolution attack and boundary correction are performed to obtain the ISCSO improved sand cat swarm algorithm;

[0019] A submodule is obtained, which is used to optimize the hyperparameters of the initial Bi LSTM liquid cooling system temperature prediction model by using the sand cat swarm algorithm improved by I SCSO, and obtain the I SCSO-B i LSTM liquid cooling system temperature prediction model.

[0020] Further, in the above data center liquid cooling uniform cooling system, the working temperature prediction module includes the following submodules:

[0021] A partitioning submodule, used for randomly partitioning the training monitoring data to obtain a training monitoring data set and a test monitoring data set;

[0022] An optimization submodule, used for performing hyperparameter optimization on the learning rate, hidden nodes and regularization coefficient by using the sand cat swarm algorithm improved by ISCS in the temperature prediction model of the ISCSSO-B i LSTM liquid cooling system;

[0023] An input submodule, used for inputting the training monitoring data set into the forward LSTM in the I SCSO-B i LSTM liquid cooling system temperature prediction model for calculation to obtain an output result;

[0024] The superposition submodule is used to store the calculated output data of the reverse LSTM layer and perform weighted superposition with the forward LSTM to obtain the initial liquid cooling system operating temperature.

[0025] Further, in the above data center liquid cooling uniform cooling system, the temperature rate calculation module includes the following submodules:

[0026] An instruction submodule, for generating a control instruction of a liquid cooling uniform cooling system based on the initial liquid cooling system operating temperature, wherein the control instruction at least includes liquid temperature, liquid flow rate, and liquid flow rate;

[0027] The control submodule is used to perform real-time temperature control on the liquid-cooled uniform cooling system according to control instructions, and to make real-time adjustments using the control logic of the liquid cooling system to obtain the controlled data center temperature.

[0028] Further, in the above data center liquid cooling uniform cooling system, the target temperature control module includes the following submodules:

[0029] A calculation submodule, used to calculate the temperature of the data center by using a differential method to obtain a temperature drop rate;

[0030] A judgment submodule is used to reduce the liquid flow rate, increase the liquid temperature and reduce the liquid flow if the temperature drop rate is greater than a preset rate; if the temperature drop rate is less than a preset rate, increase the liquid flow rate, reduce the liquid temperature and increase the liquid flow; and obtain a target liquid cooling system operating temperature;

[0031] The control submodule is used to use the target liquid cooling system operating temperature to perform target temperature control on the liquid cooling uniform cooling system.

[0032] Furthermore, in the above data center liquid cooling uniform cooling system, a method for implementing the data center liquid cooling uniform cooling system is provided, and the method comprises the following steps:

[0033] Acquire real-time monitoring data of the data center, and perform data preprocessing on the real-time monitoring data to obtain training monitoring data;

[0034] An initial BiLSTM liquid cooling system temperature prediction model is established by using a BiLSTM bidirectional long short-term memory network, and the hyperparameters of the initial BiLSTM liquid cooling system temperature prediction model are optimized by using the sand cat swarm algorithm improved by ISCS, so as to obtain an ISCSSO-BiLSTM liquid cooling system temperature prediction model;

[0035] Inputting the training monitoring data into the I SCSO-B i LSTM liquid cooling system temperature prediction model for training to obtain the initial liquid cooling system operating temperature;

[0036] Based on the initial liquid cooling system operating temperature, the liquid cooling uniform cooling system is temperature controlled in real time, and the controlled data center temperature is obtained;

[0037] The temperature drop rate of the data center is calculated according to the data center temperature. If the temperature drop rate is greater than a preset rate, a target liquid cooling system operating temperature is generated to perform target temperature control on the liquid cooling uniform cooling system.

[0038] Furthermore, in the above data center liquid cooling uniform cooling system, a method for implementing the data center liquid cooling uniform cooling system is provided, and the method comprises the following steps:

[0039] The initial Bi LSTM liquid cooling system temperature prediction model was established using the Bi LSTM bidirectional long short-term memory network;

[0040] Set the parameters of the SCSO sand cat swarm algorithm and use the K-means clustering algorithm to initialize the population of the parameters of the sand cat swarm algorithm;

[0041] Substitute the objective function to obtain the fitness value of each sand cat, and determine the mutation probability of the current iteration of the SCSO sand cat swarm algorithm and the parameters of the evolution process according to the fitness value;

[0042] The mutation probability is judged, and if it is less than the mutation probability, random walk is performed, and further evolution attack and boundary correction are performed to obtain the I SCSO improved sand cat swarm algorithm;

[0043] The ISCSSO-B i LSTM liquid cooling system temperature prediction model is obtained by optimizing the hyperparameters of the initial B i LSTM liquid cooling system temperature prediction model using the sand cat swarm algorithm improved by ISCSSO.

[0044] Furthermore, in the above data center liquid cooling uniform cooling system, a method for implementing the data center liquid cooling uniform cooling system is provided, and the method comprises the following steps:

[0045] Randomly divide the training monitoring data to obtain a training monitoring data set and a test monitoring data set;

[0046] In the I SCSO-B i LSTM liquid cooling system temperature prediction model, the I SCSO improved sand cat swarm algorithm is used to perform hyperparameter optimization on the learning rate, hidden nodes and regularization coefficient;

[0047] Input the training monitoring data set into the forward LSTM in the I SCSO-B i LSTM liquid cooling system temperature prediction model for calculation to obtain an output result;

[0048] The reverse LSTM layer saves the calculated output data and performs weighted superposition with the forward LSTM to obtain the initial liquid cooling system operating temperature.

[0049] Its beneficial effects are as follows: a detection data acquisition module is used to acquire real-time monitoring data of a data center, and data preprocessing is performed on the real-time monitoring data to obtain training monitoring data; a prediction model establishment module is used to establish an initial BiLSTM liquid cooling system temperature prediction model using a BiLSTM bidirectional long short-term memory network, and optimize the hyperparameters of the initial BiLSTM liquid cooling system temperature prediction model using an ISCSO-improved sand cat swarm algorithm to obtain an ISCSO-BiLSTM liquid cooling system temperature prediction model; an operating temperature prediction module is used to input the training monitoring data into the ISCSO-BiLSTM liquid cooling system temperature prediction model for training to obtain an initial liquid cooling system operating temperature; a temperature rate calculation module is used to perform real-time temperature control of a liquid cooling uniform cooling system based on the initial liquid cooling system operating temperature, and obtain the controlled data center temperature; a target temperature control module is used to calculate the temperature drop rate of the data center according to the data center temperature, and if the temperature drop rate is greater than a preset rate, a target liquid cooling system operating temperature is generated to perform target temperature control on the liquid cooling uniform cooling system. 1. Improve cooling efficiency: It realizes real-time monitoring and control of the working temperature of the liquid cooling system, and can quickly adjust the cooling strategy according to the actual situation, effectively improving the cooling efficiency. It avoids the phenomenon of insufficient heat dissipation or overcooling caused by inaccurate temperature control in traditional cooling methods. 2. Reduce energy consumption: By accurately controlling the temperature of the liquid cooling system, unnecessary energy consumption is reduced to achieve the purpose of energy conservation and emission reduction. While ensuring the normal operation of the server, the overall operating cost of the data center is reduced. 3. Uniform temperature distribution: By uniformly controlling the temperature of the data center, the occurrence of local temperatures that are too high or too low is avoided, and the risk of equipment failure caused by excessive temperature gradients is reduced. It helps to extend the service life of servers and related electronic equipment and reduce hardware damage caused by overheating or overcooling. BRIEF DESCRIPTION OF THE DRAWINGS

[0050] Various other advantages and benefits will become apparent to those of ordinary skill in the art by reading the following detailed description of the preferred embodiment.The drawings are only for the purpose of illustrating the preferred embodiments and are not to be construed as limiting the invention.

[0051] Figure 1 This is a schematic diagram of a first embodiment of a liquid cooling uniform cooling system for a data center according to an embodiment of the present invention;

[0052] Figure 2 This is a schematic diagram of a second embodiment of a liquid cooling uniform cooling system for a data center according to an embodiment of the present invention;

[0053] Figure 3 This is a schematic diagram of a third embodiment of a liquid cooling uniform cooling system for a data center according to an embodiment of the present invention. DETAILED DESCRIPTION

[0054] In order to make the purpose, technical solution and advantages of the present invention more clearly understood, the present invention is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not intended to limit the present invention.

[0055] Those skilled in the art will appreciate that, unless otherwise stated, the singular forms "a", "an", "the" used herein may also include plural forms. It should be further understood that the term "comprising" used in the specification of the present invention refers to the presence of the features, integers, steps, operations, elements and / or components, but does not exclude the presence or addition of one or more other features, integers, steps, operations, elements, components and / or groups thereof.

[0056] The present invention will be described in detail below in conjunction with the accompanying drawings. Figure 1 As shown, a data center liquid cooling uniform cooling system, the data center liquid cooling uniform cooling system includes the following modules:

[0057] The detection data acquisition module is used to obtain the real-time monitoring data of the data center, perform data preprocessing on the real-time monitoring data, and obtain training monitoring data;

[0058] Specifically, the present embodiment also includes an acquisition submodule for acquiring real-time monitoring data of the data center, the real-time monitoring data including at least the server CPU usage rate, memory usage percentage, coolant flow rate, ambient temperature and ambient humidity of the data center; a processing submodule for deleting outliers and filling missing values ​​in the real-time monitoring data to obtain complete monitoring data; and an acquisition submodule for standardizing the complete monitoring data to obtain training monitoring data.

[0059] The prediction model establishment module is used to establish the initial Bi LSTM liquid cooling system temperature prediction model by using the Bi LSTM bidirectional long short-term memory network, and optimize the hyperparameters of the initial Bi LSTM liquid cooling system temperature prediction model by using the sand cat swarm algorithm improved by ISCS, so as to obtain the ISCSSO-B i LSTM liquid cooling system temperature prediction model;

[0060] Specifically, the present embodiment also includes an establishment submodule for establishing an initial BiLSTM liquid cooling system temperature prediction model using a BiLSTM bidirectional long short-term memory network; a setting submodule for setting the parameters of the SCSO sand cat swarm algorithm, and using the K-means clustering algorithm to initialize the population of the parameters of the sand cat swarm algorithm; a substitution submodule for substituting the objective function to obtain the fitness value of each sand cat, and determining the mutation probability of the current iteration of the SCSO sand cat swarm algorithm and the parameters of the evolution process according to the fitness value; a judgment submodule for judging the mutation probability, and if it is less than the mutation probability, performing a random walk, and further performing an evolution attack and boundary correction to obtain an ISCSO improved sand cat swarm algorithm; an obtaining submodule for optimizing the hyperparameters of the initial BiLSTM liquid cooling system temperature prediction model using the ISCSO improved sand cat swarm algorithm to obtain an ISCSO-B iLSTM liquid cooling system temperature prediction model.

[0061] The working temperature prediction module is used to input the training monitoring data into the ISCSSO-B i LSTM liquid cooling system temperature prediction model for training to obtain the initial liquid cooling system working temperature;

[0062] Specifically, the present embodiment also includes a partitioning submodule, which is used to randomly divide the training monitoring data to obtain a training monitoring data set and a test monitoring data set; an optimization submodule, which is used to use the ISCSO improved sand cat swarm algorithm in the ISCSO-B i LSTM liquid cooling system temperature prediction model to perform hyperparameter optimization on the learning rate, hidden nodes and regularization coefficient; an input submodule, which is used to input the training monitoring data set into the forward LSTM in the ISCSO-B i LSTM liquid cooling system temperature prediction model for calculation to obtain an output result; and a superposition submodule, which is used to perform weighted superposition of the calculated output data saved in the reverse LSTM layer and the forward LSTM to obtain the initial liquid cooling system operating temperature.

[0063] Sand Cat Swarm Optimization (SCSO) is a newer metaheuristic optimization algorithm that simulates the survival and hunting behavior of sand cats in their natural environment. This algorithm was proposed by Amir Seyyedabbasi and Farzad Kiani in 2022.

[0064] Basic principle: The sand cat (Felis margarita) is a small cat living in the desert. They have excellent adaptability and hunting skills. The SCSO algorithm mainly imitates the following behaviors of the sand cat:

[0065] Foraging: The process of the sand cat searching for food is used to explore the problem space. Attacking prey: The action of capturing prey is used in the local search in the algorithm to improve the solution.

[0066] Following are the basic steps of SCSO algorithm:

[0067] Initialization: Generate an initial population of sand cats, each representing a potential solution. Evaluation: Calculate the fitness value of each sand cat. Update Position: Update the position of each sand cat according to a specific rule. This usually involves using the position of the current best sand cat as well as randomness to guide the search. Local Search: In order to improve the quality of the solution, some local search strategies may be applied. Iteration: Repeat the steps until a termination condition is reached (such as a maximum number of iterations or a certain accuracy criterion is met)

[0068] A temperature rate calculation module is used to perform real-time temperature control on the liquid cooling uniform cooling system based on the initial liquid cooling system operating temperature and obtain the controlled data center temperature;

[0069] Specifically, the present embodiment also includes an instruction submodule, which is used to generate control instructions for the liquid-cooled uniform cooling system based on the initial liquid cooling system operating temperature, and the control instructions include at least liquid temperature, liquid flow rate, and liquid flow rate; a control submodule, which is used to perform real-time temperature control of the liquid-cooled uniform cooling system according to the control instructions, and use the control logic of the liquid cooling system to make real-time adjustments to obtain the controlled data center temperature.

[0070] The target temperature control module is used to calculate the temperature drop rate of the data center according to the temperature of the data center. If the temperature drop rate is greater than the preset rate, the target liquid cooling system operating temperature is generated to perform target temperature control on the liquid cooling uniform cooling system.

[0071] Specifically, the present embodiment also includes a calculation submodule, which is used to calculate the temperature of the data center using the differential method to obtain the temperature drop rate; a judgment submodule, which is used to reduce the liquid flow rate, increase the liquid temperature and reduce the liquid flow rate if the temperature drop rate is greater than the preset rate; if the temperature drop rate is less than the preset rate, increase the liquid flow rate, reduce the liquid temperature and increase the liquid flow rate; obtain the target liquid cooling system operating temperature; and a control submodule, which is used to use the target liquid cooling system operating temperature to perform target temperature control on the liquid cooling uniform cooling system.

[0072] The beneficial effect lies in that the detection data acquisition module is used to obtain the real-time monitoring data of the data center, and the real-time monitoring data is preprocessed to obtain the training monitoring data; the prediction model establishment module is used to use the BiLSTM bidirectional long short-term memory network to establish the initial BiLSTM liquid cooling system temperature prediction model, and the hyperparameters of the initial BiLSTM liquid cooling system temperature prediction model are optimized by using the sand cat swarm algorithm improved by ISCSO to obtain the ISCSO-B iLSTM liquid cooling system temperature prediction model; the working temperature prediction module is used to input the training monitoring data into the ISCSO-B iLSTM liquid cooling system temperature prediction model for training to obtain the initial liquid cooling system working temperature; the temperature rate calculation module is used to perform real-time temperature control of the liquid cooling uniform cooling system based on the initial liquid cooling system working temperature, and obtain the controlled data center temperature; the target temperature control module is used to calculate the temperature drop rate of the data center according to the data center temperature. If the temperature drop rate is greater than the preset rate, the target liquid cooling system working temperature is generated to perform target temperature control on the liquid cooling uniform cooling system. 1. Improve cooling efficiency: It realizes real-time monitoring and control of the working temperature of the liquid cooling system, and can quickly adjust the cooling strategy according to the actual situation, effectively improving the cooling efficiency. It avoids the phenomenon of insufficient heat dissipation or overcooling caused by inaccurate temperature control in traditional cooling methods. 2. Reduce energy consumption: By accurately controlling the temperature of the liquid cooling system, unnecessary energy consumption is reduced to achieve the purpose of energy conservation and emission reduction. While ensuring the normal operation of the server, the overall operating cost of the data center is reduced. 3. Uniform temperature distribution: By uniformly controlling the temperature of the data center, the occurrence of local temperatures that are too high or too low is avoided, and the risk of equipment failure caused by excessive temperature gradients is reduced. It helps to extend the service life of servers and related electronic equipment and reduce hardware damage caused by overheating or overcooling.

[0073] In this embodiment, please refer to Figure 2 In a second embodiment of a liquid cooling uniform cooling system for a data center according to an embodiment of the present invention, the prediction model building module includes the following submodules:

[0074] Establish a submodule for establishing an initial Bi LSTM liquid cooling system temperature prediction model using a Bi LSTM bidirectional long short-term memory network;

[0075] The setting submodule is used to set the parameters of the SCSO sand cat swarm algorithm and use the K-means clustering algorithm to initialize the population of the parameters of the sand cat swarm algorithm;

[0076] Substitute the submodule into the objective function to obtain the fitness value of each sand cat, and determine the mutation probability of the current iteration of the SCSO sand cat swarm algorithm and the parameters of the evolution process according to the fitness value;

[0077] The judgment submodule is used to judge the mutation probability. If it is less than the mutation probability, random walk is performed, and further evolution attack and boundary correction are performed to obtain the I SCSO improved sand cat swarm algorithm;

[0078] A submodule is obtained, which is used to optimize the hyperparameters of the initial BiLSTM liquid cooling system temperature prediction model by using the sand cat swarm algorithm improved by ISCSO, and the ISCSO-B iLSTM liquid cooling system temperature prediction model is obtained.

[0079] In this embodiment, please refer to Figure 3 In a third embodiment of a liquid cooling uniform cooling system for a data center according to an embodiment of the present invention, the working temperature prediction module includes the following submodules:

[0080] A partitioning submodule is used to randomly partition the training monitoring data to obtain a training monitoring data set and a test monitoring data set;

[0081] The optimization submodule is used to optimize the learning rate, hidden nodes and regularization coefficient using the sand cat swarm algorithm improved by ISCS in the temperature prediction model of the ISCS-B i LSTM liquid cooling system;

[0082] The input submodule is used to input the training monitoring data set into the forward LSTM in the I SCSO-B i LSTM liquid cooling system temperature prediction model to calculate and obtain the output result;

[0083] The superposition submodule is used to store the calculated output data of the reverse LSTM layer and perform weighted superposition with the forward LSTM to obtain the initial liquid cooling system operating temperature.

[0084] The above shows and describes the basic principles, main features and advantages of the present invention. Those skilled in the art should understand that the present invention is not limited by the above embodiments. The above embodiments and descriptions are only preferred examples of the present invention and are not intended to limit the present invention. Without departing from the spirit and scope of the present invention, the present invention may have various changes and improvements, which fall within the scope of the present invention. The scope of protection of the present invention is defined by the attached claims and their equivalents.

Claims

1. A data center liquid cooling uniform cooling system, It is characterized in that The data center liquid cooling uniform cooling system includes the following modules: A detection data acquisition module is used to acquire real-time monitoring data of the data center, and perform data preprocessing on the real-time monitoring data to obtain training monitoring data; A prediction model establishment module is used to establish an initial Bi LSTM liquid cooling system temperature prediction model by using a Bi LSTM bidirectional long short-term memory network, and to optimize the hyperparameters of the initial Bi LSTM liquid cooling system temperature prediction model by using an ISCSSO-improved sand cat swarm algorithm to obtain an ISCSSO-Bi LSTM liquid cooling system temperature prediction model; An operating temperature prediction module, used for inputting the training monitoring data into the ISCS0-Bi LSTM liquid cooling system temperature prediction model for training to obtain an initial liquid cooling system operating temperature; A temperature rate calculation module, used to perform real-time temperature control on the liquid cooling uniform cooling system based on the initial liquid cooling system operating temperature, and obtain the controlled data center temperature; The target temperature control module is used to calculate the temperature drop rate of the data center according to the temperature of the data center. If the temperature drop rate is greater than a preset rate, a target liquid cooling system operating temperature is generated to perform target temperature control on the liquid cooling uniform cooling system.

2. A data center liquid cooling uniform cooling system as claimed in claim 1, It is characterized in that The detection data acquisition module includes the following submodules: The acquisition submodule is used to acquire real-time monitoring data of the data center, wherein the real-time monitoring data includes at least the server CPU usage rate, memory usage ratio, coolant flow rate, ambient temperature and ambient humidity of the data center; A processing submodule, used to delete outliers and fill missing values ​​in the real-time monitoring data to obtain complete monitoring data; A submodule is obtained, which is used to perform standardization processing on the complete monitoring data to obtain training monitoring data.

3. A data center liquid cooling uniform cooling system as claimed in claim 1, It is characterized in that The prediction model building module includes the following submodules: Establish a submodule for establishing an initial Bi LSTM liquid cooling system temperature prediction model using a Bi LSTM bidirectional long short-term memory network; The setting submodule is used to set the parameters of the SCSO sand cat swarm algorithm and use the K-means clustering algorithm to initialize the population of the parameters of the sand cat swarm algorithm; Substitute the submodule into the objective function to obtain the fitness value of each sand cat, and determine the mutation probability of the current iteration of the SCSO sand cat swarm algorithm and the parameters of the evolution process according to the fitness value; A judgment submodule is used to judge the mutation probability. If it is less than the mutation probability, random walk is performed, and further evolution attack and boundary correction are performed to obtain the ISCSO improved sand cat swarm algorithm; A submodule is obtained, which is used to optimize the hyperparameters of the initial Bi LSTM liquid cooling system temperature prediction model by using the sand cat swarm algorithm improved by I SCSO, and obtain the I SCSO-Bi LSTM liquid cooling system temperature prediction model.

4. A data center liquid cooling uniform cooling system as claimed in claim 1, It is characterized in that The working temperature prediction module includes the following submodules: A partitioning submodule, used for randomly partitioning the training monitoring data to obtain a training monitoring data set and a test monitoring data set; An optimization submodule, used for performing hyperparameter optimization on the learning rate, hidden nodes and regularization coefficient by using the sand cat swarm algorithm improved by I SCSO in the temperature prediction model of the I SCSO-Bi LSTM liquid cooling system; An input submodule, used for inputting the training monitoring data set into the forward LSTM in the I SCSO-Bi LSTM liquid cooling system temperature prediction model for calculation to obtain an output result; The superposition submodule is used to store the calculated output data of the reverse LSTM layer and perform weighted superposition with the forward LSTM to obtain the initial liquid cooling system operating temperature.

5. A data center liquid cooling uniform cooling system as claimed in claim 1, It is characterized in that The temperature rate calculation module includes the following submodules: An instruction submodule, for generating a control instruction of a liquid cooling uniform cooling system based on the initial liquid cooling system operating temperature, wherein the control instruction at least includes liquid temperature, liquid flow rate, and liquid flow rate; The control submodule is used to perform real-time temperature control on the liquid-cooled uniform cooling system according to control instructions, and to make real-time adjustments using the control logic of the liquid cooling system to obtain the controlled data center temperature.

6. A data center liquid cooling uniform cooling system as claimed in claim 1, It is characterized in that The target temperature control module includes the following submodules: A calculation submodule, used to calculate the temperature of the data center by using a differential method to obtain a temperature drop rate; A judgment submodule is used to reduce the liquid flow rate, increase the liquid temperature and reduce the liquid flow if the temperature drop rate is greater than a preset rate; if the temperature drop rate is less than a preset rate, increase the liquid flow rate, reduce the liquid temperature and increase the liquid flow; and obtain a target liquid cooling system operating temperature; The control submodule is used to use the target liquid cooling system operating temperature to perform target temperature control on the liquid cooling uniform cooling system.

7. A method for implementing a data center liquid cooling uniform cooling system as claimed in claim 1, It is characterized in that The method comprises the following steps: Acquire real-time monitoring data of the data center, and perform data preprocessing on the real-time monitoring data to obtain training monitoring data; An initial Bi LSTM liquid cooling system temperature prediction model is established by using a Bi LSTM bidirectional long short-term memory network, and the hyperparameters of the initial Bi LSTM liquid cooling system temperature prediction model are optimized by using an ISCSO-modified sand cat swarm algorithm to obtain an ISCSO-Bi LSTM liquid cooling system temperature prediction model; Inputting the training monitoring data into the I SCSO-Bi LSTM liquid cooling system temperature prediction model for training to obtain the initial liquid cooling system operating temperature; Based on the initial liquid cooling system operating temperature, the liquid cooling uniform cooling system is temperature controlled in real time, and the controlled data center temperature is obtained; The temperature drop rate of the data center is calculated according to the data center temperature. If the temperature drop rate is greater than a preset rate, a target liquid cooling system operating temperature is generated to perform target temperature control on the liquid cooling uniform cooling system.

8. A method for implementing a data center liquid cooling uniform cooling system as claimed in claim 1, It is characterized in that The method comprises the following steps: The Bi LSTM bidirectional long short-term memory network is used to establish the initial Bi LSTM liquid cooling system temperature prediction model; Set the parameters of the SCSO sand cat swarm algorithm and use the K-means clustering algorithm to initialize the population of the parameters of the sand cat swarm algorithm; Substitute the objective function to obtain the fitness value of each sand cat, and determine the mutation probability of the current iteration of the SCSO sand cat swarm algorithm and the parameters of the evolution process according to the fitness value; The mutation probability is judged, and if it is less than the mutation probability, random walk is performed, and further evolution attack and boundary correction are performed to obtain the I SCSO improved sand cat swarm algorithm; The I SCSO-Bi LSTM liquid cooling system temperature prediction model is obtained by optimizing the hyperparameters of the initial Bi LSTM liquid cooling system temperature prediction model using the sand cat swarm algorithm improved by I SCSO.

9. A method for implementing a data center liquid cooling uniform cooling system as claimed in claim 1, It is characterized in that The method comprises the following steps: Randomly divide the training monitoring data to obtain a training monitoring data set and a test monitoring data set; In the I SCSO-Bi LSTM liquid cooling system temperature prediction model, the I SCSO improved sand cat swarm algorithm is used to perform hyperparameter optimization on the learning rate, hidden nodes and regularization coefficient; Input the training monitoring data set into the forward LSTM in the I SCSO-Bi LSTM liquid cooling system temperature prediction model for calculation to obtain an output result; The reverse LSTM layer saves the calculated output data and performs weighted superposition with the forward LSTM to obtain the initial liquid cooling system operating temperature.

Images