Data center carbon dioxide immersed cooling system and method

Abstract

The invention discloses a data center carbon dioxide immersed cooling system and method.The system comprises a centralized energy supply module and a data center module, and the centralized energy supply module comprises a carbon dioxide storage tank, a first compressor, an oil separator, a condenser, an expansion machine, an evaporator and a second compressor; a heat exchange energy storage piece is arranged between the condenser and the evaporator; a pressure compensation piece is arranged at an inlet of the second compressor; the data center module comprises a suspended ceiling pipeline layer, a carbon dioxide incoming pipeline, a honeycomb-shaped nozzle, a floor frame high layer, a carbon dioxide return pipeline, an air return fan and a residual pressure valve; the method comprises the following steps: step 1, carrying out carbon dioxide immersed cooling on the data center; and 2, controlling the temperature of the data center. The device is reasonable in design, carbon dioxide is used for carrying out immersion type cooling on the data center module machine room through the honeycomb-shaped spray head, the carbon dioxide after heat exchange is recycled, efficient refrigeration of the data center can be achieved, and energy consumption of the carbon dioxide can be reduced.

Description

Technical Field

[0001] The invention belongs to the technical field of data center refrigeration, and in particular relates to a carbon dioxide immersion cooling system for a data center. Background Art

[0002] Data centers are regarded as the "brains" of the modern information society, storing, transmitting and processing massive amounts of data at all times. From the short messages we browse every day to the cloud computing services of enterprises, they all rely on the support of data centers. As the scale of data centers continues to grow, the heat dissipation efficiency of traditional air-cooled and water-cooled refrigeration cycles can no longer meet the needs of development, and liquid cooling technology has begun to be gradually applied. The mainstream modes can be divided into three technical paths: cold plate type, immersion type and spray type.

[0003] Traditional data center cooling methods usually use one-way cooling, using water cooling or air cooling to lower the temperature, while the heat taken away by the working fluid will be discharged, ignoring the recycling of waste heat resources, resulting in a waste of resources. The traditional cooling method uses Freon refrigerant as the system working fluid, which will produce hydrofluorocarbon compounds such as Freon during the working process, which seriously damages the ozone layer. The immersion liquid of the liquid cooling technology path is relatively expensive and not conducive to maintenance.

[0004] Therefore, a rationally designed data center carbon dioxide immersion cooling system is needed, which uses honeycomb nozzles to immerse carbon dioxide in the data center module room for cooling, and recycles the carbon dioxide after heat exchange, which can not only achieve efficient cooling of the data center, but also reduce carbon dioxide energy consumption. Summary of the invention

[0005] The technical problem to be solved by the present invention is to provide a data center carbon dioxide immersion cooling system in view of the deficiencies in the above-mentioned prior art. The system has a reasonable design and uses honeycomb nozzles to immerse carbon dioxide in the data center module room for cooling, and the carbon dioxide after heat exchange is recycled, which can not only achieve efficient cooling of the data center, but also reduce carbon dioxide energy consumption.

[0006] In order to solve the above technical problems, the technical solution adopted by the present invention is: a data center carbon dioxide immersion cooling system, characterized in that: it includes a centralized energy supply module and a data center module, the centralized energy supply module includes a carbon dioxide gas storage tank, a first compressor, an oil separator, a condenser, an expander, an evaporator and a second compressor, the first inlet of the condenser is connected to the outlet of the oil separator through a first pipeline, and the inlet of the oil separator is connected to the outlet of the first compressor;

[0007] The first outlet of the condenser is connected to the inlet of the data center module through a second pipeline, a cooling capacity regulating valve is provided on the second pipeline, and the outlet of the data center module is connected to the expander;

[0008] The first inlet of the evaporator is connected to the outlet of the expander through a third pipeline, the first outlet of the evaporator is connected to the inlet of the second compressor through a fourth pipeline, a heat exchange energy storage component is arranged between the condenser and the evaporator; and a pressure compensation component is arranged at the inlet of the second compressor;

[0009] The data center module includes a suspended ceiling pipeline layer arranged on the top of the data center module room, a carbon dioxide supply pipeline laid along the suspended ceiling pipeline layer, a honeycomb nozzle arranged at the bottom of the carbon dioxide supply pipeline, a floor frame layer arranged on the floor of the data center module room, a carbon dioxide return pipeline arranged in the floor frame layer, a plurality of return fans arranged on the carbon dioxide return pipeline, and a residual pressure valve arranged at the outlet end of the carbon dioxide return pipeline, wherein the outlet end of the carbon dioxide return pipeline is the outlet of the data center module; the inlet of the carbon dioxide return pipeline is connected to the indoor of the data center module room.

[0010] In the above-mentioned data center carbon dioxide immersion cooling system, the heat exchange energy storage component includes a cold water storage tank and a hot water storage tank, the second inlet of the condenser is connected to the cold water storage tank, the second outlet of the condenser is connected to the hot water storage tank, a circulating pump is arranged between the second outlet of the condenser and the hot water storage tank, the hot water storage tank is connected to the first inlet of the evaporator, and the first outlet of the evaporator is connected to the cold water storage tank.

[0011] In the above-mentioned data center carbon dioxide immersion cooling system, the pressure compensating component includes a first three-way valve arranged on the first pipeline, a second three-way valve arranged on the fourth pipeline, a pressure compensating pipeline connected between the first three-way valve and the second three-way valve, and a pressure compensating valve arranged on the pressure compensating pipeline.

[0012] In the above-mentioned data center carbon dioxide immersion cooling system, an evaporation pressure regulating valve is provided on the fourth pipeline, and the evaporation pressure regulating valve is located between the second three-way valve and the first outlet of the evaporator;

[0013] The outlet of the second compressor is connected to the carbon dioxide gas storage tank through a fifth pipeline, and a drying filter is arranged on the fifth pipeline;

[0014] A pressure controller is arranged between the inlet of the oil separator and the outlet of the first compressor.

[0015] In the above-mentioned data center carbon dioxide immersion cooling system, an internal temperature sensor is arranged inside the data center module room, and the internal temperature sensor is connected to a computer.

[0016] At the same time, a method for controlling a carbon dioxide refrigeration and temperature control system is provided, characterized in that the method comprises the following steps:

[0017] Step 1: Data Center CO2 Immersion Cooling:

[0018] Step 101, the carbon dioxide gas in the carbon dioxide storage tank enters the first compressor through the first pipeline for compression, the compressed carbon dioxide gas passes through the oil separator, the first three-way valve and the first pipeline into the first inlet of the condenser, and is output as liquid carbon dioxide through heat exchange in the condenser, and the liquid carbon dioxide is transported to the carbon dioxide supply pipeline in the module room of the data center through the second pipeline and the cold capacity regulating valve, and the liquid carbon dioxide in the carbon dioxide supply pipeline is sprayed out through the honeycomb nozzle into the machine room for immersion cooling; wherein, the cold medium in the cold water storage tank enters the second inlet of the condenser for heat exchange with the compressed carbon dioxide gas, and at the same time, the cold water absorbs heat through the condenser to become hot water, and the hot water enters the hot water storage tank for storage under the action of the circulating pump;

[0019] Step 102: The gaseous carbon dioxide at the outlet of the data center module enters the expander to liquefy the gaseous carbon dioxide into liquid carbon dioxide. The liquid carbon dioxide enters the first inlet of the evaporator through the third pipeline. The heat medium in the hot water storage tank enters the first inlet of the evaporator. The heat medium and the liquid carbon dioxide exchange heat twice. The hot water passes through the evaporator to become cold water stored in the cold water storage tank. The liquid carbon dioxide passes through the evaporator to become gaseous carbon dioxide after heat exchange.

[0020] Step 103, the gaseous carbon dioxide after heat exchange enters the second compressor through the fourth pipeline for pressurization, and the pressurized gaseous carbon dioxide is filtered through a drying filter to obtain filtered carbon dioxide gas; wherein, when the gaseous carbon dioxide after heat exchange enters the second compressor through the fourth pipeline for pressurization, the evaporation pressure regulating valve is adjusted or the operating pressure compensation valve is opened to adjust the inlet pressure of the second compressor to meet the design requirements;

[0021] Step 104, the filtered carbon dioxide gas enters the carbon dioxide gas storage tank through the fifth pipeline for collection and recirculation;

[0022] Step 2: Temperature control of data center:

[0023] Step 201: During the process of the honeycomb nozzle spraying atomized water into the computer room for immersion cooling, the internal temperature sensor detects the temperature inside the computer room in real time according to the set sampling time Δt, obtains the internal temperature detection value at each sampling time and sends it to the computer;

[0024] Step 202: The computer determines the internal temperature according to the detected internal temperature value and the internal temperature setting value T set, obtain the temperature deviation value e at the kth moment, the temperature deviation value e at the k-1th moment, and the temperature deviation value e at the k-2th moment; wherein k is a positive integer greater than 2;

[0025] Step 203: According to Δe=ee and Δe=ee, obtain the deviation change Δe at the kth moment and the deviation change Δe at the k-1th moment;

[0026] Step 204: input the temperature deviation value e at the kth moment, the deviation change Δe at the kth moment, and the deviation change Δe at the k-1th moment into N trained RBF neural network models respectively to obtain N groups of proportional adjustment coefficients kp, integral adjustment coefficients ki, and differential adjustment coefficients kd; wherein N is a positive integer;

[0027] Step 205: Input the N groups of proportional adjustment coefficients kp, integral adjustment coefficients ki and differential adjustment coefficients kd into the PID controller in Matlab software for simulation, and obtain the system rise time T of the N groups. r , overshoot O s and steady-state error SSE;

[0028] Step 206: Construct the objective function J = ω1 × T r +ω2×O s +ω3×|SSE|; where J represents fitness, ω1 represents the first weight, and ω1=0.2; ω2 represents the second weight, and ω2=0.3; ω3 represents the third weight, and ω3=0.5;

[0029] Step 207: Set the system rise time T of group N r , overshoot O s Substitute the steady-state error SSE into the objective function, and the trained RBF neural network model corresponding to the minimum fitness value is recorded as the optimal RBF neural network model;

[0030] Step 208: Obtain the control amount of the cooling control valve according to the proportional control coefficient kp, integral control coefficient ki and differential control coefficient kd output by the RBF neural network optimal model until the internal temperature detection value is close to the internal temperature setting value T set .

[0031] Compared with the prior art, the present invention has the following advantages:

[0032] 1. The present invention sets carbon dioxide as a cold source. The system has the advantages of long life and low maintenance cost, and realizes efficient refrigeration effect in compression, condensation, expansion and evaporation cycles, reducing pollution to the environment. Carbon dioxide is used to atomize and immerse the data center module room, avoiding the price disadvantage of the immersion liquid and improving the convenience of daily maintenance.

[0033] 2. The centralized energy supply module and the data center module of the present invention are separately arranged. The relative separation of the centralized energy supply module enables the data center module to save some area for arranging more cabinets, thereby increasing its space utilization efficiency and improving its operating income.

[0034] 3. The present invention uses a honeycomb nozzle arranged on the carbon dioxide pipeline to spray carbon dioxide mist into the computer room until an immersion effect is achieved, thereby cooling the computer room and avoiding liquid cooling. In addition, the carbon dioxide brought out by the refrigeration process at the data center module outlet is evaporated through an evaporator to utilize carbon dioxide as a resource, thereby improving energy efficiency and economic benefits.

[0035] 4. In the present invention, a heat exchange energy storage component is arranged between the condenser and the evaporator to form a loop circulation; the waste heat is stored in a hot water storage tank and the power is transmitted by a circulating pump; the cold medium after passing through the evaporator is stored in a cold water storage tank, which can realize both condensation and evaporation, saving energy.

[0036] In summary, the present invention is reasonably designed, and uses carbon dioxide through honeycomb nozzles to immerse the module room of the data center in cooling, and the carbon dioxide after heat exchange is recycled, which can not only achieve efficient cooling of the data center, but also reduce carbon dioxide energy consumption.

[0037] The technical solution of the present invention is further described in detail below through the accompanying drawings and embodiments. BRIEF DESCRIPTION OF THE DRAWINGS

[0038] Figure 1 It is a structural schematic diagram of the present invention.

[0039] Figure 2 It is a structural schematic diagram of the data center module of the present invention.

[0040] Description of reference numerals:

[0041] 1—CO2 gas storage tank; 2—Pressure controller; 3—First compressor;

[0042] 4—oil separator; 5—first three-way valve; 6—condenser;

[0043] 7—cooling regulating valve; 8—data center module; 8-1—expander;

[0044] 8-2—Carbon dioxide pipeline; 8-3—Honeycomb nozzle;

[0045] 8-6—Carbon dioxide return line; 9—Evaporator;

[0046] 10—evaporation pressure regulating valve; 11—second three-way valve; 12—second compressor;

[0047] 14—drying filter; 15—circulating pump; 16—hot water storage tank;

[0048] 17—cold water storage tank; 18—pressure compensation pipeline; 18-1—pressure compensation valve;

[0049] 35—return air fan; 36—residual pressure valve;

[0050] 37—ceiling pipeline layer; 38—floor frame high layer. DETAILED DESCRIPTION

[0051] like Figure 1 and Figure 2 As shown, a data center carbon dioxide immersion cooling system of the present invention includes a centralized energy supply module and a data center module 8, wherein the centralized energy supply module includes a carbon dioxide gas storage tank 1, a first compressor 3, an oil separator 4, a condenser 6, an expander 8-1, an evaporator 9 and a second compressor 12, wherein a first inlet of the condenser 6 is connected to an outlet of the oil separator 4 through a first pipeline, and an inlet of the oil separator 4 is connected to an outlet of the first compressor 3;

[0052] The first outlet of the condenser 6 is connected to the inlet of the data center module 8 through a second pipeline, a cooling regulating valve 7 is provided on the second pipeline, and the outlet of the data center module 8 is connected to the expander 8-1;

[0053] The first inlet of the evaporator 9 is connected to the outlet of the expander 8-1 through the third pipeline, and the first outlet of the evaporator 9 is connected to the inlet of the second compressor 12 through the fourth pipeline. A heat exchange energy storage component is provided between the condenser 6 and the evaporator 9; and a pressure compensation component is provided at the inlet of the second compressor 12;

[0054] The data center module 8 includes a suspended ceiling pipeline layer 37 arranged on the top of the data center module 8 computer room, a carbon dioxide supply pipeline 8-2 arranged along the suspended ceiling pipeline layer 37, a honeycomb nozzle 8-3 arranged at the bottom of the carbon dioxide supply pipeline 8-2, and a floor frame layer 38 arranged on the floor of the data center module 8 computer room, a carbon dioxide return pipeline 8-6 arranged in the floor frame layer 38, a plurality of return air fans 35 arranged on the carbon dioxide return pipeline 8-6, and a residual pressure valve 36 arranged at the outlet end of the carbon dioxide return pipeline 8-6, the outlet end of the carbon dioxide return pipeline 8-6 is the outlet of the data center module 8; the inlet of the carbon dioxide return pipeline 8-6 is connected to the indoor of the data center module 8 computer room.

[0055] In this embodiment, the heat exchange energy storage component includes a cold water storage tank 17 and a hot water storage tank 16. The second inlet of the condenser 6 is connected to the cold water storage tank 17, and the second outlet of the condenser 6 is connected to the hot water storage tank 16. A circulating pump 15 is arranged between the second outlet of the condenser 6 and the hot water storage tank 16. The hot water storage tank 16 is connected to the first inlet of the evaporator 9, and the first outlet of the evaporator 9 is connected to the cold water storage tank 17.

[0056] In this embodiment, the pressure compensating component includes a first three-way valve 5 arranged on the first pipeline, a second three-way valve 11 arranged on the fourth pipeline, a pressure compensating pipeline 18 connected between the first three-way valve 5 and the second three-way valve 11, and a pressure compensating valve 18-1 arranged on the pressure compensating pipeline 18.

[0057] In this embodiment, an evaporation pressure regulating valve 10 is provided on the fourth pipeline, and the evaporation pressure regulating valve 10 is located between the second three-way valve 11 and the first outlet of the evaporator 9;

[0058] The outlet of the second compressor 12 is connected to the carbon dioxide storage tank 1 through a fifth pipeline, and a drying filter 14 is arranged on the fifth pipeline;

[0059] A pressure controller 2 is arranged between the inlet of the oil separator 4 and the outlet of the first compressor 3 .

[0060] In this embodiment, an internal temperature sensor is provided inside the computer room of the data center module 8, and the internal temperature sensor is connected to a computer.

[0061] In this embodiment, during the specific implementation, a centralized energy supply module and a data center module 8 are set, and the centralized energy supply module is used in an integrated manner, which is convenient for unified management, greatly reduces the operating cost and improves the reliability of energy supply;

[0062] Data center module 8 removes some traditional system refrigeration devices and places them into a centralized energy supply module, making it possible to install more cabinets. At the same time, data center module 8 will adopt carbon dioxide mist immersion cooling, saving the cost of terminal precision air conditioning, and can effectively avoid the occurrence of local hot spots and greatly reduce the PUE value of the data center.

[0063] In this embodiment, during specific implementation, the carbon dioxide gas storage tank 1 has inlet and outlet ports on both sides; the pressure controller 2 is connected to the outlet pipeline of the first compressor 3 to control the exhaust pressure within the standard range to ensure the system operation stability requirements;

[0064] In this embodiment, during specific implementation, a pressure compensation pipeline 18 is provided between the first three-way valve 5 and the second three-way valve 11, and a pressure compensation valve 18-1 is provided on the pressure compensation pipeline 18 to ensure that the system operation is not lower than the minimum return air pressure and meet the design requirements so that the system return air pressure is stable;

[0065] In addition, the evaporation pressure regulating valve 10 is also provided to assist in regulating the return air pressure, and the double auxiliary regulation can avoid an increase in the compressor load, shortening the service life of the equipment, and also avoid an increase in energy consumption.

[0066] In this embodiment, during specific implementation, the cooling capacity regulating valve 7 is arranged on the second pipeline of the condenser 6, and the refrigerant flow is ensured to match the mechanical load through valve throttling control.

[0067] In this embodiment, during specific implementation, the oil separator 4 is used to separate the lubricating oil in the carbon dioxide gas to ensure the operating efficiency of the system.

[0068] In this embodiment, during specific implementation, a heat exchange and energy storage component is arranged between the condenser 6 and the evaporator 9 to form a loop circulation; during actual use, the working medium can be replaced according to the actual needs of the application scenario, and can be replaced with ammonia or other media; the waste heat is stored in the hot water storage tank 16, and the power is transmitted through the circulating pump 15; the cold medium after passing through the evaporator 9 is stored in the cold water storage tank 17, which can realize both condensation and evaporation, saving energy.

[0069] In this embodiment, during specific implementation, both the hot water storage tank 16 and the cold water storage tank 17 are underground cavities, which are often set up underground in the centralized energy supply module for use, and the tank body has an insulation layer, which can effectively reduce energy loss; in some embodiments, a tank body may not be provided, and a water storage pool may be used for storage, and the water pool can be used for practical purposes such as a year-round swimming pool.

[0070] In this embodiment, during specific implementation, the right side of the condenser 6 is connected to a cooling capacity regulating valve 7 to control the flow rate of the carbon dioxide working medium to be within a reasonable range.

[0071] In this embodiment, during the specific implementation, the carbon dioxide supply pipeline 8-2 extends along the ceiling pipeline layer 37 to the upper part of the data center module 8 computer room, and the carbon dioxide is sprayed into the computer room by the honeycomb nozzle 8-3 set on the carbon dioxide supply pipeline 8-2 until the immersion effect is achieved, so as to cool the computer room;

[0072] When the temperature of the carbon dioxide gas in the computer room rises to the set standard value, the return air fan 35 provided on the carbon dioxide return pipe 8-6 below the floor frame high level 38 performs the suction action; the residual pressure valve 36 controls the amount of gas sucked out to ensure that the return air volume is slightly less than the intake air volume, thereby ensuring the carbon dioxide cooling utilization rate and micro-positive pressure operation in the computer room. The gaseous carbon dioxide carrying the residual heat resource is led out of the data center module through the carbon dioxide return pipe 8-6 and sent to the expander 8-1 for liquefaction;

[0073] In this embodiment, during specific implementation, the computer room of the data center module 8 is closed, and the space such as the aisle outside the computer room is a non-closed space in the current embodiment.

[0074] In this embodiment, in actual use, airtight doors need to be installed at the entrance and exit ends of the computer room of the data center module 8 to ensure the indoor pressure balance of the computer room and prevent carbon dioxide from coming into contact with outdoor air and reacting, thereby affecting the operating efficiency of the system; the main airflow direction of the computer room is downward airflow, and the carbon dioxide working medium is introduced into the pipeline 8-2 through the carbon dioxide installed in the ceiling pipeline layer 37, and is atomized and sprayed into the computer room through the honeycomb nozzle 8-3 installed thereon to cool the cabinet. The atomization of the honeycomb nozzle 8-3 is used to increase the spray coverage range to ensure that the concentration of the atomized substance in the space is relatively uniform.

[0075] In this embodiment, in actual use, a drying filter 14 is provided, and the carbon dioxide gas is filtered out of impurities and flows back into the carbon dioxide gas storage tank 1 to complete the system operation cycle.

[0076] In this embodiment, in actual use, the height of the ceiling pipeline layer 37 and the floor frame layer 38 is not less than 70 cm, which is used for the installation of pipelines along the road and at the same time ensures the heat dissipation and maintenance needs of the cabinet grounding side.

[0077] A method for controlling a carbon dioxide immersion cooling system in a data center, the method comprising the following steps:

[0078] Step 1: Data Center CO2 Immersion Cooling:

[0079] Step 101, the carbon dioxide gas in the carbon dioxide storage tank 1 enters the first compressor 3 through the first pipeline for compression, the compressed carbon dioxide gas passes through the oil separator 4, the first three-way valve 5 and the first pipeline into the first inlet of the condenser 6, and is heat-exchanged by the condenser 6 to output liquid carbon dioxide, and the liquid carbon dioxide is transported to the carbon dioxide inlet pipeline 8-2 in the computer room of the data center module 8 through the second pipeline and the cold capacity regulating valve 7, and the liquid carbon dioxide in the carbon dioxide inlet pipeline 8-2 is atomized and sprayed through the honeycomb nozzle 8-3 to enter the computer room for immersion cooling; wherein, the cold medium in the cold water storage tank 17 enters the second inlet of the condenser 3 to perform heat exchange with the compressed carbon dioxide gas, and at the same time, the cold water absorbs heat through the condenser 3 to become hot water, and the hot water enters the hot water storage tank 16 for storage under the action of the circulating pump 15;

[0080] Step 102: The gaseous carbon dioxide at the outlet of the data center module 8 enters the expander 8-1 to liquefy the gaseous carbon dioxide into liquid carbon dioxide. The liquid carbon dioxide enters the first inlet of the evaporator 9 through the third pipeline. The heat medium in the hot water storage tank 16 enters the first inlet of the evaporator 9. The heat medium and the liquid carbon dioxide exchange heat twice. The hot water passes through the evaporator 9 to become cold water stored in the cold water storage tank 17. The liquid carbon dioxide passes through the evaporator 9 to become gaseous carbon dioxide after heat exchange.

[0081] Step 103, the gaseous carbon dioxide after heat exchange enters the second compressor 12 through the fourth pipeline for pressurization, and the pressurized gaseous carbon dioxide is filtered through the drying filter 14 to obtain filtered carbon dioxide gas; wherein, when the gaseous carbon dioxide after heat exchange enters the second compressor 12 through the fourth pipeline for pressurization, the evaporation pressure regulating valve 10 is adjusted or the operating pressure compensation valve 18-1 is opened to adjust the inlet pressure of the second compressor 12 to meet the design requirements;

[0082] Step 104, the filtered carbon dioxide gas enters the carbon dioxide gas storage tank 1 through the fifth pipeline for collection and recirculation;

[0083] Step 2: Temperature control of data center:

[0084] Step 201: During the process of the honeycomb nozzle 8-3 spraying atomized water into the machine room for immersion cooling, the internal temperature sensor detects the temperature inside the machine room in real time according to the set sampling time Δt, obtains the internal temperature detection value at each sampling time and sends it to the computer;

[0085] Step 202: The computer determines the internal temperature according to the detected internal temperature value and the internal temperature setting value T set , obtain the temperature deviation value ek at the kth moment, the temperature deviation value ek-1 at the k-1th moment, and the temperature deviation value ek-2 at the k-2th moment; wherein k is a positive integer greater than 2;

[0086] Step 203, according to Δek=ek-ek-1 and Δek-1=ek-1-ek-2, obtain the deviation change Δek at the kth moment and the deviation change Δek-1 at the k-1th moment;

[0087] Step 204, input the temperature deviation value ek at the kth moment, the deviation change Δek at the kth moment, and the deviation change Δek-1 at the k-1th moment into N trained RBF neural network models respectively, to obtain N groups of proportional adjustment coefficients kp, integral adjustment coefficients ki, and differential adjustment coefficients kd; wherein N is a positive integer;

[0088] Step 205: Input the N groups of proportional adjustment coefficients kp, integral adjustment coefficients ki and differential adjustment coefficients kd into the PID controller in Matlab software for simulation, and obtain the system rise time T of the N groups. r , overshoot O s and steady-state error SSE;

[0089] Step 206: Construct objective function J = ω1 × T r +ω2×O s +ω3×|SSE|; where J represents fitness, ω1 represents the first weight, and ω1=0.2; ω2 represents the second weight, and ω2=0.3; ω3 represents the third weight, and ω3=0.5;

[0090] Step 207: Set the system rise time T of group N r , overshoot O s Substitute the steady-state error SSE into the objective function, and the trained RBF neural network model corresponding to the minimum fitness value is recorded as the optimal RBF neural network model;

[0091] Step 208: Obtain the control amount of the cooling capacity regulating valve 7 according to the proportional regulation coefficient kp, integral regulation coefficient ki and differential regulation coefficient kd output by the RBF neural network optimal model until the internal temperature detection value approaches the internal temperature setting value T set .

[0092] In this embodiment, during specific implementation, the specific process of obtaining N trained RBF neural network models is as follows:

[0093] Step A, according to the method of step 201 to step 203, the temperature deviation value ek at the kth moment, the deviation change Δek at the kth moment, the deviation change Δek-1 at the k-1th moment, and the proportional adjustment coefficient kp, integral adjustment coefficient ki and differential adjustment coefficient kd corresponding to the cooling capacity regulating valve 7 at the kth moment in the historical simulation process are obtained, which is a set of training data;

[0094] Step B: According to the method of step A, multiple groups of training data are obtained to form a training set;

[0095] Step C, constructing N RBF neural network models; wherein the N RBF neural network models each include an input layer, a hidden layer and an output layer, the number of neurons in the input layer is 3, the number of neurons in the hidden layer is 5, the number of neurons in the output layer is 3, and N is a positive integer;

[0096] Step D: Input the training sets into N RBF neural network models for training respectively to obtain N trained RBF neural network models.

[0097] In this embodiment, during specific implementation, the value of N is 50 to 100 and can be adjusted as needed.

[0098] In this embodiment, during the specific implementation, the constructed N RBF neural network models are trained respectively to obtain N trained RBF neural network models; in order to consider the randomness of the grid initialization of the RBF neural network model, random training is performed through N RBF neural network models, and the optimal RBF neural network model is obtained in combination with the subsequent objective function selection, thereby improving the accuracy of the subsequent proportional adjustment coefficient kp, integral adjustment coefficient ki and differential adjustment coefficient kd, thereby improving the PID control accuracy.

[0099] In this embodiment, during specific implementation, the sampling time Δt and the internal temperature setting value T set It is set according to actual requirements and is not specifically limited.

[0100] In this embodiment, when the internal temperature detection value is close to the internal temperature setting value T set That is, the internal temperature detection value and the internal temperature setting value T set The deviation is minimal.

[0101] In summary, the present invention is reasonably designed, and uses carbon dioxide through honeycomb nozzles to immerse the module room of the data center in cooling, and the carbon dioxide after heat exchange is recycled, which can not only achieve efficient cooling of the data center, but also reduce carbon dioxide energy consumption.

[0102] The above description is only a preferred embodiment of the present invention and does not limit the present invention in any way. Any simple modification, change and equivalent structural change made to the above embodiment based on the technical essence of the present invention still falls within the protection scope of the technical solution of the present invention.

Claims

1. A data center carbon dioxide immersion cooling system, characterized in that: The invention comprises a centralized energy supply module and a data center module (8), wherein the centralized energy supply module comprises a carbon dioxide storage tank (1), a first compressor (3), an oil separator (4), a condenser (6), an expander (8-1), an evaporator (9) and a second compressor (12), wherein a first inlet of the condenser (6) is connected to an outlet of the oil separator (4) via a first pipeline, and an inlet of the oil separator (4) is connected to an outlet of the first compressor (3); The first outlet of the condenser (6) is connected to the inlet of the data center module (8) via a second pipeline, a cooling capacity regulating valve (7) is provided on the second pipeline, and the outlet of the data center module (8) is connected to the expander (8-1); The first inlet of the evaporator (9) is connected to the outlet of the expander (8-1) through a third pipeline, the first outlet of the evaporator (9) is connected to the inlet of the second compressor (12) through a fourth pipeline, a heat exchange energy storage component is provided between the condenser (6) and the evaporator (9); and a pressure compensation component is provided at the inlet of the second compressor (12); The data center module (8) comprises a suspended ceiling pipeline layer (37) arranged on the top of the data center module (8) machine room, a carbon dioxide supply pipeline (8-2) arranged along the suspended ceiling pipeline layer (37), a honeycomb nozzle (8-3) arranged at the bottom of the carbon dioxide supply pipeline (8-2), a floor frame layer (38) arranged on the floor of the data center module (8) machine room, a carbon dioxide return pipeline (8-6) arranged in the floor frame layer (38), a plurality of return fans (35) arranged on the carbon dioxide return pipeline (8-6), and a residual pressure valve (36) arranged at the outlet end of the carbon dioxide return pipeline (8-6), wherein the outlet end of the carbon dioxide return pipeline (8-6) is the outlet of the data center module (8); and the inlet of the carbon dioxide return pipeline (8-6) is connected to the interior of the data center module (8) machine room.

2. A data center carbon dioxide immersion cooling system according to claim 1, characterized in that: The heat exchange energy storage component comprises a cold water storage tank (17) and a hot water storage tank (16); the second inlet of the condenser (6) is connected to the cold water storage tank (17); the second outlet of the condenser (6) is connected to the hot water storage tank (16); a circulation pump (15) is provided between the second outlet of the condenser (6) and the hot water storage tank (16); the hot water storage tank (16) is connected to the first inlet of the evaporator (9); and the first outlet of the evaporator (9) is connected to the cold water storage tank (17).

3. A data center carbon dioxide immersion cooling system according to claim 1, characterized in that: The pressure compensating component comprises a first three-way valve (5) arranged on a first pipeline, a second three-way valve (11) arranged on a fourth pipeline, a pressure compensating pipeline (18) connected between the first three-way valve (5) and the second three-way valve (11), and a pressure compensating valve (18-1) arranged on the pressure compensating pipeline (18).

4. A data center carbon dioxide immersion cooling system according to claim 3, characterized in that: The fourth pipeline is provided with an evaporation pressure regulating valve (10), and the evaporation pressure regulating valve (10) is located between the second three-way valve (11) and the first outlet of the evaporator (9); The outlet of the second compressor (12) is connected to the carbon dioxide storage tank (1) via a fifth pipeline, and a drying filter (14) is provided on the fifth pipeline; A pressure controller (2) is arranged between the inlet of the oil separator (4) and the outlet of the first compressor (3).

5. A data center carbon dioxide immersion cooling system according to claim 1, characterized in that: An internal temperature sensor is arranged inside the computer room of the data center module (8), and the internal temperature sensor is connected to a computer.

6. A method for controlling the data center carbon dioxide immersion cooling system according to claim 5, characterized in that: The method comprises the following steps: Step 1: Data Center CO2 Immersion Cooling: Step 101, the carbon dioxide gas in the carbon dioxide storage tank (1) enters the first compressor (3) through the first pipeline for compression, the compressed carbon dioxide gas passes through the oil separator (4), the first three-way valve (5) and the first pipeline into the first inlet of the condenser (6), and is heat exchanged through the condenser (6) to output liquid carbon dioxide, the liquid carbon dioxide is transported to the carbon dioxide inlet pipeline (8-2) in the computer room of the data center module (8) through the second pipeline and the cooling capacity regulating valve (7), the liquid carbon dioxide in the carbon dioxide inlet pipeline (8-2) is sprayed through the honeycomb nozzle (8-3) and enters the computer room for immersion cooling; wherein, the cold medium in the cold water storage tank (17) enters the second inlet of the condenser (3) to perform heat exchange with the compressed carbon dioxide gas, and at the same time, the cold water absorbs heat through the condenser (3) to become hot water, and the hot water enters the hot water storage tank (16) for storage under the action of the circulating pump (15); Step 102: The gaseous carbon dioxide at the outlet of the data center module (8) enters the expander (8-1) to liquefy the gaseous carbon dioxide into liquid carbon dioxide. The liquid carbon dioxide enters the first inlet of the evaporator (9) through the third pipeline. The heat medium in the hot water storage tank (16) enters the first inlet of the evaporator (9). The heat medium and the liquid carbon dioxide exchange heat twice. The hot water passes through the evaporator (9) to become cold water and is stored in the cold water storage tank (17). The liquid carbon dioxide passes through the evaporator (9) to become gaseous carbon dioxide after heat exchange. Step 103, the gaseous carbon dioxide after heat exchange enters the second compressor (12) through the fourth pipeline for pressurization, and the pressurized gaseous carbon dioxide is filtered through the drying filter (14) to obtain filtered carbon dioxide gas; wherein, when the gaseous carbon dioxide after heat exchange enters the second compressor (12) through the fourth pipeline for pressurization, the evaporation pressure regulating valve (10) is adjusted or the operating pressure compensation valve (18-1) is opened to adjust the inlet pressure of the second compressor (12) to meet the design requirements; Step 104, the filtered carbon dioxide gas enters the carbon dioxide gas storage tank (1) through the fifth pipeline for collection and recirculation; Step 2: Temperature control of data center: Step 201: During the process of the honeycomb nozzle (8-3) spraying atomized water into the machine room for immersion cooling, the internal temperature sensor detects the temperature inside the machine room in real time according to the set sampling time Δt, obtains the internal temperature detection value at each sampling time and sends it to the computer; Step 202: The computer determines the internal temperature according to the detected internal temperature value and the internal temperature setting value T set , obtain the temperature deviation value e(k) at the kth moment, the temperature deviation value e(k-1) at the k-1th moment, and the temperature deviation value e(k-2) at the k-2th moment; wherein k is a positive integer greater than 2; Step 203: According to Δe(k)=e(k)-e(k-1) and Δe(k-1)=e(k-1)-e(k-2), obtain the deviation change Δe(k) at the kth moment and the deviation change Δe(k-1) at the k-1th moment; Step 204: input the temperature deviation value e(k) at the kth moment, the deviation change Δe(k) at the kth moment, and the deviation change Δe(k-1) at the k-1th moment into N trained RBF neural network models respectively to obtain N groups of proportional adjustment coefficients kp, integral adjustment coefficients ki, and differential adjustment coefficients kd; wherein N is a positive integer; Step 205: Input the N groups of proportional adjustment coefficients kp, integral adjustment coefficients ki and differential adjustment coefficients kd into the PID controller in Matlab software for simulation, and obtain the system rise time T of the N groups. r , overshoot O s and steady-state error SSE; Step 206: Construct objective function J = ω1 × T r +ω2×O s +ω3×|SSE|; where J represents fitness, ω1 represents the first weight, and ω1=0.2; ω2 represents the second weight, and ω2=0.3; ω3 represents the third weight, and ω3=0.5; Step 207: Set the system rise time T of group N r , overshoot O s Substitute the steady-state error SSE into the objective function, and the trained RBF neural network model corresponding to the minimum fitness value is recorded as the optimal RBF neural network model; Step 208: Obtain the control amount of the cooling capacity regulating valve (7) according to the proportional regulation coefficient kp, integral regulation coefficient ki and differential regulation coefficient kd output by the RBF neural network optimal model, until the internal temperature detection value approaches the internal temperature setting value T set .

Images