Method and device for regulating flow of cooling liquid, storage medium and electronic device

Abstract

This application provides a method, apparatus, storage medium, and electronic device for regulating coolant flow rate. The method includes: determining node flow rate allocation coefficients for m server nodes based on node information of m server nodes at a first moment; and determining branch flow rate allocation coefficients for n cooling branches based on branch information of n cooling branches of each server node at a first moment; controlling the node solenoid valves of the m server nodes according to the node flow rate allocation coefficients within a first control cycle to control the first flow rate of coolant flowing through the m server nodes via the node solenoid valves; and controlling the branch solenoid valves of the n cooling branches according to the branch flow rate allocation coefficients of each server node within the first control cycle to control the second flow rate of coolant flowing through the n cooling branches via the branch solenoid valves.

Description

Technical Field

[0001] This application relates to the field of computers, and more specifically, to a method and apparatus for regulating coolant flow, a storage medium, and an electronic device. Background Technology

[0002] With the rapid development of technologies such as big data, artificial intelligence (AI), and high-performance computing (HPC), server hardware has evolved from single processors to multi-core processors, and from single nodes to multi-node cluster systems. As the number of computing nodes increases, power consumption issues also become more severe.

[0003] Every server generates a significant amount of heat during operation, which must be effectively dissipated to ensure system stability and performance. While traditional air cooling technology has been widely used for decades, its heat dissipation efficiency is increasingly insufficient for today's high-density, high-power server systems. Liquid cooling technology, by using liquid as a heat transfer medium, offers higher thermal conductivity and more uniform heat distribution than air, thus significantly improving heat dissipation efficiency.

[0004] Liquid cooling systems currently typically employ a uniform cooling scheme, setting a fixed coolant flow rate for all nodes. Once the system starts operating, coolant is supplied according to this pre-set fixed flow rate, without considering the system's current power consumption and cooling requirements. While this coolant flow rate may meet system needs at low power consumption levels, it leads to wasted cooling resources. At high power consumption levels, the coolant flow rate cannot adapt, posing a risk to system heat dissipation. Furthermore, rack-mount server systems have multiple server nodes with varying power consumption and coolant requirements. A fixed coolant flow rate cannot effectively regulate the instantaneous coolant demand of each node, failing to achieve proper coolant distribution.

[0005] Regarding the relevant technologies, current liquid cooling systems adopt a unified cooling scheme, but the power consumption of each server node in a rack server system is inconsistent, and the coolant demand varies. The fixed coolant flow rate cannot reasonably adjust the instantaneous coolant demand of each node, and the technical problem of not being able to achieve reasonable distribution of coolant has not yet been solved. Summary of the Invention

[0006] This application provides a method and apparatus for regulating coolant flow rate, a storage medium, and an electronic device to at least solve the problem in the related art where current liquid cooling systems use a unified cooling scheme, but the power consumption of each server node in the rack server system is inconsistent, the coolant demand varies, and a fixed coolant flow rate cannot reasonably adjust the instantaneous coolant demand of each node, thus failing to achieve reasonable distribution of coolant.

[0007] According to one embodiment of this application, a method for regulating coolant flow rate is provided, comprising: determining node flow rate allocation coefficients for m server nodes based on node information of m server nodes at a first moment, and determining branch flow rate allocation coefficients for n cooling branches of each server node based on branch information of n cooling branches at the first moment, wherein m and n are both positive integers, the node information includes: node power consumption information, node coolant inlet and outlet temperature difference, and the branch information includes: branch power consumption information, branch coolant inlet and outlet temperature difference; controlling the node solenoid valves of the m server nodes according to the m node flow rate allocation coefficients within a first control cycle, so as to control the flow rate of a first flow rate of coolant flowing through the m server nodes through the m node solenoid valves, wherein the first control cycle corresponds to the first moment; controlling the branch solenoid valves of the n cooling branches according to the n branch flow rate allocation coefficients of each server node within the first control cycle, so as to control the flow rate of a second flow rate of coolant flowing through the n cooling branches through the n branch solenoid valves.

[0008] In an exemplary embodiment, determining the node traffic allocation coefficient of the m server nodes based on the node information of the m server nodes at a first time moment includes: calculating the average node power consumption of the m server nodes at the first time moment based on the node power consumption information of the m server nodes at the first time moment, and calculating the average node temperature difference of the m server nodes at the first time moment based on the node coolant inlet and outlet temperature difference of the m server nodes at the first time moment; using the formula Calculate the flow distribution coefficients for the m nodes, where Rm is the flow distribution coefficient of the mth server node, Pm is the power consumption information of the mth server node, Pc is the average power consumption of the node, ΔTm is the temperature difference between the inlet and outlet of the coolant of the mth server node, and ΔTc is the average temperature difference of the node.

[0009] In an exemplary embodiment, controlling the node solenoid valves of the m server nodes according to the m node flow distribution coefficients within a first control cycle, so as to control the flow rate of the first flow of coolant flowing through the m server nodes through the m node solenoid valves, includes: normalizing the m node flow distribution coefficients Rm to obtain m normalized node flow distribution coefficients Rxm; converting the m normalized node flow distribution coefficients Rxm into m digital signals Dm, where Dm = Rxm * 255; and sending the m digital signals Dm to a DAC chip to instruct the DAC chip to distribute the first flow of coolant through the m server nodes. A digital signal Dm is converted into m analog voltage signals Am; the m analog voltage signals Am sent by the DAC chip are received, and the m analog voltage signals Am are input to an amplifier to instruct the amplifier to adjust the magnitude of the signal current according to the m analog voltage signals, so as to obtain m signal currents Cm; the m signal currents Cm are respectively input to m node solenoid valves to instruct the m node solenoid valves to adjust the valve opening size of the m node solenoid valves in the first control cycle according to the m signal currents Cm, and control the flow rate of the first flow of coolant flowing through the m server nodes through the valve opening size.

[0010] In an exemplary embodiment, after controlling the node solenoid valves of the m server nodes according to the m node flow allocation coefficients within a first control cycle, the method further includes: acquiring node information of the m server nodes at a second time moment, wherein the second time moment is within the first control cycle; calculating a first power consumption difference and a first temperature difference of the m server nodes, wherein the first power consumption difference is the difference between the node power consumption information of the server node at the second time moment and the average power consumption of the node, and the first temperature difference is the difference between the node coolant inlet / outlet temperature difference of the server node at the second time moment and the average temperature difference of the node; Based on the node power consumption difference threshold, the m first power consumption differences are divided into multiple first power consumption fuzzy sets; based on the node temperature difference threshold, the m first temperature difference differences are divided into multiple first temperature difference fuzzy sets; and based on the node allocation coefficient threshold, the m node traffic allocation coefficients are divided into multiple first coefficient fuzzy sets. If it is determined that a target server node among the m server nodes meets the first-level cooling warning rule, a node cooling failure warning is issued for the target server node. The first-level cooling warning rule is determined based on the multiple first power consumption fuzzy sets, the multiple first temperature difference fuzzy sets, and the multiple first coefficient fuzzy sets.

[0011] In an exemplary embodiment, determining the flow distribution coefficient of the n cooling branches based on the branching information of the n cooling branches of each server node at the first time moment includes: calculating the average power consumption of the n cooling branches at the first time moment based on the branching power consumption information of the n cooling branches at the first time moment, and calculating the average temperature difference of the n cooling branches at the first time moment based on the inlet and outlet temperature difference of the coolant in the n cooling branches at the first time moment; using the formula Calculate n branch flow distribution coefficients, where Rn is the branch flow distribution coefficient of the nth cooling branch, Pn is the branch power consumption information of the nth cooling branch, Pk is the average power consumption of the branch, ΔTn is the branch coolant inlet and outlet temperature difference of the nth cooling branch, and ΔTk is the average temperature difference of the branch.

[0012] In an exemplary embodiment, the step of controlling the branch solenoid valves of n cooling branches within the first control cycle according to the n branch flow allocation coefficients of each server node, so as to control the flow rate of the second flow of coolant flowing through the n cooling branches through the n branch solenoid valves, includes: normalizing the n branch flow allocation coefficients Rn to obtain n normalized branch flow allocation coefficients Rxn; converting the n normalized branch flow allocation coefficients Rxn into n digital signals Dn, where Dn = Rxn * 255; and sending the n digital signals Dn to the DAC chip to instruct the DAC chip. The n digital signals Dn are converted into n analog voltage signals An; the n analog voltage signals An sent by the DAC chip are received, and the n analog voltage signals An are input to the amplifier to instruct the amplifier to adjust the magnitude of the signal current according to the n analog voltage signals, so as to obtain n signal currents Cn; the n signal currents Cn are respectively input to the n branch solenoid valves to instruct the n branch solenoid valves to adjust the valve opening size of the n branch solenoid valves in the first control cycle according to the n signal currents Cn, and control the flow rate of the second flow of coolant flowing through the n cooling branches by the valve opening size.

[0013] In an exemplary embodiment, after controlling the branch solenoid valves of the n cooling branches within the first control cycle according to the n branch flow allocation coefficients of each server node, the method further includes: obtaining the branch information of the n cooling branches at a third time, wherein the third time is located within the first control cycle; calculating the second power consumption difference and the second temperature difference of the n cooling branches, wherein the second power consumption difference is the difference between the branch power consumption information of the cooling branch at the third time and the average power consumption of the branch, and the second temperature difference is the difference between the branch coolant inlet / outlet temperature difference of the cooling branch at the third time and the average temperature difference of the branch. The difference; based on the branch power consumption difference threshold, the n second power consumption differences are divided into multiple second power consumption fuzzy sets; based on the branch temperature difference threshold, the n second temperature difference differences are divided into multiple second temperature difference fuzzy sets; and based on the branch allocation coefficient threshold, the n branch flow allocation coefficients are divided into multiple second coefficient fuzzy sets; when it is determined that there is a target cooling branch among the n cooling branches that meets the secondary cooling warning rule, a branch cooling failure warning is issued for the target cooling branch, wherein the secondary cooling warning rule is determined based on the multiple second power consumption fuzzy sets, the multiple second temperature difference fuzzy sets, and the multiple second coefficient fuzzy sets.

[0014] According to another embodiment of this application, a coolant flow rate control device is provided, comprising: a determining module, configured to determine the node flow rate allocation coefficient of m server nodes based on node information of m server nodes at a first moment, and to determine the branch flow rate allocation coefficient of n cooling branches of each server node based on branch information of n cooling branches at the first moment, wherein m and n are both positive integers, the node information includes: node power consumption information, node coolant inlet and outlet temperature difference, and the branch information includes: branch power consumption information, branch coolant inlet and outlet temperature difference; and a first control module, configured to... The node solenoid valves of the m server nodes are controlled according to the m node flow distribution coefficients within a first control cycle, so as to control the flow rate of the first flow of coolant flowing through the m server nodes through the m node solenoid valves, wherein the first control cycle corresponds to the first moment; the second control module is used to control the branch solenoid valves of n cooling branches within the first control cycle according to the n branch flow distribution coefficients of each server node, so as to control the flow rate of the second flow of coolant flowing through the n cooling branches through the n branch solenoid valves.

[0015] According to yet another embodiment of this application, a computer-readable storage medium is also provided, wherein a computer program is stored therein, and the computer program is configured to perform the steps in any of the above method embodiments when it is run.

[0016] According to yet another embodiment of this application, an electronic device is also provided, including a memory and a processor, wherein the memory stores a computer program, and the processor is configured to run the computer program to perform the steps in any of the above method embodiments.

[0017] According to yet another embodiment of this application, a computer program product is also provided, including a computer program that, when executed by a processor, implements the steps of the methods described in various embodiments of this application.

[0018] This application proposes a real-time control method for coolant flow. First, based on the node information of m server nodes at a first moment, the node flow allocation coefficients for these m server nodes are determined. Then, based on the branch information of n cooling branches of each server node at the first moment, the branch flow allocation coefficients for each cooling branch are determined. The node information includes node power consumption and coolant inlet / outlet temperature difference, and the branch information includes branch power consumption and coolant inlet / outlet temperature difference. Next, based on the m node flow allocation coefficients, the node solenoid valves of these m server nodes are controlled within the first control cycle, thereby controlling the coolant flow rate through each server node. Then, within each server node… Based on n branch flow distribution coefficients, the branch solenoid valves of the n cooling branches are controlled within the first control cycle, thereby controlling the flow rate of coolant through each cooling branch. Using this scheme, the node solenoid valves and branch solenoid valves installed within each server node are adjusted according to the cooling requirements of different server nodes and different cooling branches, thus achieving a reasonable distribution of coolant. This solves the problem in related technologies where current liquid cooling systems use a uniform cooling scheme, but the power consumption of each server node in a rack server system is inconsistent, resulting in different coolant requirements. A fixed coolant flow rate cannot reasonably adjust the instantaneous coolant demand of each node, failing to achieve reasonable coolant distribution. Attached Figure Description

[0019] Figure 1 This is a hardware structure block diagram of a computer terminal for a method of regulating coolant flow according to an embodiment of this application.

[0020] Figure 2 This is a flowchart of a method for regulating coolant flow rate according to an embodiment of this application;

[0021] Figure 3 This is an optional whole-rack liquid cooling system and hardware connection block diagram according to an embodiment of this application;

[0022] Figure 4This is an optional intranode liquid cooling system and hardware connection block diagram according to an embodiment of this application;

[0023] Figure 5 This is an optional intelligent control flowchart of a rack server liquid cooling system according to an embodiment of this application;

[0024] Figure 6 This is a structural block diagram of a coolant flow rate control device according to an embodiment of this application. Detailed Implementation

[0025] The embodiments of this application will be described in detail below with reference to the accompanying drawings and examples.

[0026] It should be noted that the terms "first," "second," etc., in the specification, claims, and drawings of this application are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence.

[0027] The methods and embodiments provided in this application can be executed on a computer terminal or similar computing device. Taking running on a computer terminal as an example, Figure 1 This is a hardware structure block diagram of a computer terminal for a method of regulating coolant flow according to an embodiment of this application. Figure 1 As shown, a computer terminal may include one or more ( Figure 1 Only one is shown in the diagram. A processor 102 (which may include, but is not limited to, a microprocessor MCU or a programmable logic device FPGA, etc.) and a memory 104 for storing data are also shown. The computer terminal may further include a transmission device 106 for communication functions and an input / output device 108. Those skilled in the art will understand that... Figure 1 The structure shown is for illustrative purposes only and does not limit the structure of the computer terminal described above. For example, the computer terminal may also include components that are more complex than those described above. Figure 1 The more or fewer components shown, or having the same Figure 1 The different configurations shown.

[0028] The memory 104 can be used to store computer programs, such as application software programs and modules, like the computer program corresponding to the coolant flow control method in this embodiment. The processor 102 executes various functional applications and data processing by running the computer program stored in the memory 104, thus implementing the above-described method. The memory 104 may include high-speed random access memory and non-volatile memory, such as one or more magnetic storage devices, flash memory, or other non-volatile solid-state memory. In some instances, the memory 104 may further include memory remotely located relative to the processor 102, and these remote memories can be connected to a computer terminal via a network. Examples of such networks include, but are not limited to, the Internet, corporate intranets, local area networks, mobile communication networks, and combinations thereof.

[0029] The transmission device 106 is used to receive or send data via a network. Specific examples of the network described above may include a wireless network provided by a communication provider for the computer terminal. In one example, the transmission device 106 includes a Network Interface Controller (NIC), which can connect to other network devices via a base station to communicate with the Internet. In another example, the transmission device 106 may be a Radio Frequency (RF) module used for wireless communication with the Internet.

[0030] The following explains the technical terms used in this application:

[0031] BMC: Baseboard Management Controller;

[0032] DAC: Digital-to-Analog Converter Chip;

[0033] 12C: Inter-Integrated Circuit, I2C bus or integrated circuit interconnect bus;

[0034] GPU: graphics processing unit;

[0035] SW: PCIe Switch, PCIe expansion chip;

[0036] CPU: Central Processing Unit.

[0037] This embodiment provides a method for regulating coolant flow rate. Figure 2 This is a flowchart of a method for regulating coolant flow rate according to an embodiment of this application, as follows: Figure 2As shown, the method includes the following steps:

[0038] Step S202: Determine the node flow allocation coefficient of the m server nodes based on the node information of the m server nodes at the first moment, and determine the branch flow allocation coefficient of the n cooling branches based on the branch information of the n cooling branches of each server node at the first moment, where m and n are both positive integers. The node information includes: node power consumption information and node coolant inlet and outlet temperature difference. The branch information includes: branch power consumption information and branch coolant inlet and outlet temperature difference.

[0039] It should be noted that the connection method between the whole rack liquid cooling system and multiple server nodes is as follows: Figure 3 As shown, the whole rack liquid cooling system includes a water inlet passage and a water return passage. Each server node is connected to the water inlet passage through a solenoid valve. The flow rate of coolant flowing through the corresponding server node can be controlled by the valve opening size of the solenoid valve. Each server node is equipped with a BMC, which communicates with the whole rack control system.

[0040] Step S204: According to the flow distribution coefficients of the m nodes, the node solenoid valves of the m server nodes are controlled in the first control cycle, so as to control the flow rate of the first flow of the coolant flowing through the m server nodes through the node solenoid valves, wherein the first control cycle corresponds to the first moment.

[0041] Step S206: Based on the n branch flow distribution coefficients of each server node, control the branch solenoid valves of the n cooling branches within the first control cycle, so as to control the flow rate of the second flow of the coolant flowing through the n cooling branches through the n branch solenoid valves.

[0042] The above scheme first determines the node flow allocation coefficients of the m server nodes based on their node information at the first moment, and then determines the branch flow allocation coefficients of each of the n cooling branch circuits of each server node based on their branch information at the first moment. The node information includes node power consumption and coolant inlet / outlet temperature difference, and the branch information includes branch power consumption and coolant inlet / outlet temperature difference. Then, based on the m node flow allocation coefficients, the node solenoid valves of the m server nodes are controlled within the first control cycle, thereby controlling the coolant flow rate through each server node. Finally, within each server node, the flow rates of the n branch circuits are... The allocation coefficient controls the branch solenoid valves of n cooling branches during the first control cycle, thereby controlling the flow rate of coolant through each cooling branch. Using this scheme, the node solenoid valves and branch solenoid valves located within each server node are adjusted according to the cooling requirements of different server nodes and different cooling branches, thus achieving a reasonable distribution of coolant. This solves the problem in related technologies where current liquid cooling systems use a uniform cooling scheme, but the power consumption of each server node in a rack server system is inconsistent, resulting in different coolant requirements. A fixed coolant flow rate cannot reasonably adjust the instantaneous coolant demand of each node, failing to achieve reasonable coolant distribution.

[0043] In an exemplary embodiment, determining the node traffic allocation coefficient of the m server nodes based on the node information of the m server nodes at a first time moment includes: calculating the average node power consumption of the m server nodes at the first time moment based on the node power consumption information of the m server nodes at the first time moment, and calculating the average node temperature difference of the m server nodes at the first time moment based on the node coolant inlet and outlet temperature difference of the m server nodes at the first time moment; using the formula Calculate the flow distribution coefficients for the m nodes, where Rm is the flow distribution coefficient of the mth server node, Pm is the power consumption information of the mth server node, Pc is the average power consumption of the node, ΔTm is the temperature difference between the inlet and outlet of the coolant of the mth server node, and ΔTc is the average temperature difference of the node.

[0044] The rack control system BMC communicates with each node control system BMC via multiple I2C SWITCH chips, polling to obtain the node power consumption information Pm and the node inlet and outlet water temperature difference ΔTm (i.e., the inlet and outlet temperature difference of the coolant at the node) reported by each node control system BMC (e.g., P1 and ΔT1 for server node 1, P2 and ΔT2 for server node 2, and Pm and ΔTm for server node M).

[0045] Then, the rack control system BMC calculates the average power consumption of each node in the rack based on the power consumption information reported by the BMC of each node control system (i.e., the aforementioned node power consumption information). Average temperature difference at nodes of computer cabinet Traffic distribution coefficients at each node of the computer cabinet (Node 1 is R1, Node 2 is R2, and Node M is Rm).

[0046] In this embodiment, the cooling requirements of each server node at the first moment are determined based on the actual power consumption and the temperature difference between the inlet and outlet water at the first moment. Then, the node flow allocation coefficient of each server node is determined based on the cooling requirements of each server node. Subsequently, the coolant flow rate that meets the cooling requirements of the server node can be allocated according to the node flow allocation coefficient.

[0047] Further, the step of controlling the node solenoid valves of the m server nodes according to the m node flow distribution coefficients in the first control cycle, so as to control the flow rate of the first flow of coolant flowing through the m server nodes through the m node solenoid valves, includes: normalizing the m node flow distribution coefficients Rm to obtain m normalized node flow distribution coefficients Rxm; converting the m normalized node flow distribution coefficients Rxm into m digital signals Dm, where Dm = Rxm * 255; and sending the m digital signals Dm to the DAC chip to instruct the DAC chip to distribute the m digital signals... The signal Dm is converted into m analog voltage signals Am; the m analog voltage signals Am sent by the DAC chip are received, and the m analog voltage signals Am are input to the amplifier to instruct the amplifier to adjust the magnitude of the signal current according to the m analog voltage signals, so as to obtain m signal currents Cm; the m signal currents Cm are respectively input to the m node solenoid valves to instruct the m node solenoid valves to adjust the valve opening size of the m node solenoid valves in the first control cycle according to the m signal currents Cm, and control the flow rate of the first flow of coolant flowing through the m server nodes through the valve opening size.

[0048] The flow distribution coefficients of each node are normalized to obtain the normalized flow distribution coefficients Rxm. The cabinet control system (BMC) outputs different digital signals Dm (Dm = Rxm * 255) to the DAC chip. The DAC converts the digital signal Dm into the corresponding analog voltage signal Am (for example, a digital value of 0 may be converted to 0V, and a digital value of 255 may be converted to the maximum output voltage (such as 5V)). The DAC outputs a continuous analog voltage signal to the amplifier. The analog voltage signal output by the DAC needs to be amplified to adjust the current. The amplifier adjusts the current magnitude according to the change in the input voltage signal and sends the adjusted current Cm to the corresponding downstream node solenoid valve. The node solenoid valve controls the flow rate of the current node according to the received signal current magnitude.

[0049] like Figure 3 As shown, after obtaining the node traffic distribution coefficients of each server node, the whole rack control system processes the data through devices such as I2CSW, BMC, DAC chips and amplifiers, and converts it into corresponding signal currents, which are then input into the corresponding solenoid valves to complete the control of the solenoid valves.

[0050] The rack control system BMC continuously polls each node BMC to obtain power consumption information and repeats the above process to achieve primary control of the coolant flow rate of each node in the entire rack.

[0051] In this embodiment, the node solenoid valves of each server node are controlled according to the calculated node flow distribution coefficient, thereby inputting an appropriate amount of coolant to the server node according to its cooling needs to complete the cooling work and achieve reasonable distribution of coolant.

[0052] Based on the above steps, after controlling the node solenoid valves of the m server nodes according to the flow distribution coefficients of the m nodes in the first control cycle, the method further includes: obtaining the node information of the m server nodes at a second time moment, wherein the second time moment is within the first control cycle; calculating the first power consumption difference and the first temperature difference of the m server nodes, wherein the first power consumption difference is the difference between the node power consumption information of the server node at the second time moment and the average power consumption of the node, and the first temperature difference is the difference between the node coolant inlet / outlet temperature difference of the server node at the second time moment and the average temperature difference of the node; according to The node power consumption difference threshold divides the m first power consumption differences into multiple first power consumption fuzzy sets, the node temperature difference threshold divides the m first temperature difference differences into multiple first temperature difference fuzzy sets, and the node allocation coefficient threshold divides the m node traffic allocation coefficients into multiple first coefficient fuzzy sets. If it is determined that there is a target server node among the m server nodes that meets the first-level cooling warning rule, a node cooling failure warning is issued for the target server node. The first-level cooling warning rule is determined based on the multiple first power consumption fuzzy sets, the multiple first temperature difference fuzzy sets, and the multiple first coefficient fuzzy sets.

[0053] The cabinet control system (BMC) also uses fuzzy control to perform failure early warning based on the node power consumption Pm, node temperature difference ΔTm, and the node flow distribution coefficient Rm calculated above, as reported by the BMC of each node control system. Specifically, it calculates the difference between the node power consumption and the node average power consumption (Pm-Pc) (i.e., the first power consumption difference mentioned above), and the difference between the node inlet and outlet water temperature difference (i.e., the node coolant inlet and outlet temperature difference mentioned above) and the node average inlet and outlet water temperature difference (i.e., the node average temperature difference mentioned above) (ΔTm-ΔTc) (i.e., the first temperature difference mentioned above). It then divides each power consumption difference and temperature difference into fuzzy sets: negative large, negative medium, negative small, zero, positive small, positive medium, and positive large (i.e., multiple first power consumption fuzzy sets and multiple first temperature difference fuzzy sets mentioned above). For example: If the power consumption difference between node 1 and the average power consumption of the node is P1-Pc < 0, and |P1-Pc| > 0.7*Pc, then the power consumption difference of this node is classified as "negatively large". Here, 0 and 0.7*Pc are the threshold values ​​for the power consumption difference of the node. If the temperature difference between node 2 and the average temperature difference of the node is ΔT1-ΔTc > 0, and 0.3*ΔTc < |ΔT1-ΔTc| < 0.7*Pc, then the temperature difference of this node is classified as "moderate". Here, 0, 0.3*ΔTc, and 0.7*ΔTc are the threshold values ​​for the temperature difference of the node. The node flow allocation coefficient Rm is divided into fuzzy sets with 30% and 70% (i.e., node allocation coefficient thresholds): small, medium, and large (i.e., multiple fuzzy sets of first coefficients). The cabinet control system (BMC) monitors the fuzzy set to which the power consumption difference, temperature difference, and flow distribution coefficient of each node belong. It issues failure warnings based on the set rules (i.e., the first-level cooling warning rules). For example, if the power consumption difference of node m is "positive large", the temperature difference is "negative small", and the flow distribution coefficient of node m is "large", then the node may have a cooling failure. The entire system BMC should issue a node failure warning, and maintenance personnel need to inspect the liquid cooling system of that node.

[0054] This embodiment combines the actual power consumption, actual temperature difference, and node traffic allocation coefficient of each server node to perform cooling monitoring and early warning. When cooling abnormalities are detected, timely repairs are carried out to avoid system damage and greater losses.

[0055] Optionally, determining the flow distribution coefficient of the n cooling branches based on the branching information of the n cooling branches of each server node at the first moment includes: calculating the average power consumption of the n cooling branches at the first moment based on the branching power consumption information of the n cooling branches at the first moment, and calculating the average temperature difference of the n cooling branches at the first moment based on the inlet and outlet temperature difference of the coolant in the n cooling branches at the first moment; using the formula Calculate n branch flow distribution coefficients, where Rn is the branch flow distribution coefficient of the nth cooling branch, Pn is the branch power consumption information of the nth cooling branch, Pk is the average power consumption of the branch, ΔTn is the branch coolant inlet and outlet temperature difference of the nth cooling branch, and ΔTk is the average temperature difference of the branch.

[0056] After the rack control system BMC completes the primary regulation of the coolant flow rate of the nodes (i.e., completes the flow regulation of the server nodes), the BMC of each node control system performs secondary regulation of the coolant flow rate inside the node based on the power consumption of each branch circuit and the temperature difference between the inlet and outlet water of each branch circuit.

[0057] The liquid cooling system within the server node and its corresponding connection method are as follows: Figure 4 As shown, the node control system (BMC) connects multiple I2C switches to expand the I2C pathway. I2C switch0 connects to 8 PCIe switch chips, I2C switch1 connects to 8 Retime chips, and I2C switch2 connects to 8 GPUs. The BMC directly connects to 2 CPUs and communicates with the PCIe switch, Retime, and GPUs through 3 I2C switches to obtain the current power consumption of each chip. The node's internal cooling system is divided into 3 branches: the motherboard branch, the switch board branch, and the GPU branch. The BMC obtains the inlet and outlet water temperature differences for these 3 branches. The flow distribution coefficient for the motherboard branch is calculated based on CPU power consumption information and the motherboard branch temperature difference. The flow coefficient for the switch board branch is calculated based on PCIe switch chip power consumption information and the switch board branch temperature difference. The flow rate for the GPU board branch is calculated based on the Retime chip, GPU chip power consumption information, and GPU board branch error. The node control system (BMC) calculates the branch flow distribution coefficient based on the power consumption information of each branch and the inlet and outlet water temperature difference. The calculation method for the branch flow distribution coefficient is similar to that for the node flow distribution coefficient.

[0058] In this embodiment, within the server node, multiple chips are further divided into multiple cooling circuits. Then, based on the power consumption and inlet / outlet water temperature difference data of each cooling circuit, the cooling requirements of each cooling circuit are determined, thereby determining the flow distribution coefficient of each cooling circuit. This allows for further distribution of coolant flow within the server node.

[0059] Optionally, the step of controlling the branch solenoid valves of the n cooling branches within the first control cycle according to the n branch flow allocation coefficients of each server node, so as to control the flow rate of the second flow of coolant flowing through the n cooling branches through the n branch solenoid valves, includes: normalizing the n branch flow allocation coefficients Rn to obtain n normalized branch flow allocation coefficients Rxn; converting the n normalized branch flow allocation coefficients Rxn into n digital signals Dn, where Dn = Rxn * 255; and sending the n digital signals Dn to the DAC chip to instruct the DAC chip to control the second flow of coolant flowing through the n cooling branches. n digital signals Dn are converted into n analog voltage signals An; the n analog voltage signals An sent by the DAC chip are received, and the n analog voltage signals An are input to the amplifier to instruct the amplifier to adjust the magnitude of the signal current according to the n analog voltage signals, so as to obtain n signal currents Cn; the n signal currents Cn are respectively input to the n branch solenoid valves to instruct the n branch solenoid valves to adjust the valve opening size of the n branch solenoid valves in the first control cycle according to the n signal currents Cn, and control the flow rate of the second flow of coolant flowing through the n cooling branches by the valve opening size.

[0060] First, the n branch flow distribution coefficients Rn are normalized to obtain n normalized branch flow distribution coefficients Rxn. Then, Rxn is converted into n digital signals Dn. Optionally, the digital signals can be 8-bit. The corresponding 8-bit digital signals are output to the DAC chip. The DAC converts the digital signals into corresponding analog voltage signals and sends them to the amplifier chip. The amplifier adjusts the current according to the change in the input voltage signal and sends the adjusted current to the corresponding internal solenoid valve. The internal solenoid valve controls the coolant flow of each path in the node according to the current magnitude of the received signal, completing the two-stage regulation.

[0061] Through the above embodiments, the node control system monitors the power consumption information of each branch and the temperature difference between the inlet and outlet water of each branch within the node, and calculates the branch flow distribution coefficient according to the calculation method, and controls the solenoid valves of each flow channel within the node to distribute the coolant flow in each area of ​​the node, thereby achieving two-level regulation.

[0062] Optionally, after controlling the branch solenoid valves of the n cooling branches within the first control cycle according to the n branch flow allocation coefficients of each server node, the method further includes: obtaining the branch information of the n cooling branches at a third time, wherein the third time is within the first control cycle; calculating the second power consumption difference and the second temperature difference of the n cooling branches, wherein the second power consumption difference is the difference between the branch power consumption information of the cooling branch at the third time and the average power consumption of the branch, and the second temperature difference is the difference between the branch coolant inlet and outlet temperature difference of the cooling branch at the third time and the average temperature difference of the branch; Based on the power consumption difference threshold, n second power consumption differences are divided into multiple second power consumption fuzzy sets; based on the temperature difference threshold, n second temperature difference differences are divided into multiple second temperature difference fuzzy sets; and based on the distribution coefficient threshold, n branch flow distribution coefficients are divided into multiple second coefficient fuzzy sets. If it is determined that a target cooling branch among the n cooling branches meets the secondary cooling warning rule, a branch cooling failure warning is issued for the target cooling branch. The secondary cooling warning rule is determined based on the multiple second power consumption fuzzy sets, the multiple second temperature difference fuzzy sets, and the multiple second coefficient fuzzy sets.

[0063] In addition, the node control system (BMC) also uses fuzzy control methods to provide failure warnings based on the power consumption, temperature difference, and calculated flow distribution coefficients of each branch. When the information of a certain branch meets the warning rules, the node control system (BMC) issues a branch failure warning and reports the warning information to the entire system (BMC).

[0064] Specifically, the fuzzy warning method for each cooling branch is based on the same principle as the fuzzy warning method for the server node. First, the branch information for these n cooling branches at the third time step is obtained, and the second power consumption difference and the second temperature difference are calculated. The second power consumption difference is the difference between the branch power consumption information at the third time step and the average power consumption of the branch. The second temperature difference is the difference between the branch coolant inlet / outlet temperature difference and the average temperature difference of the branch at the third time step. Then, based on the branch power consumption difference threshold, the n second power consumption differences are divided into multiple second power consumption fuzzy sets; based on the branch temperature difference threshold, the n second temperature difference differences are divided into multiple second temperature difference fuzzy sets; and based on the branch allocation coefficient threshold, the n branch flow allocation coefficients are divided into multiple second coefficient fuzzy sets. Finally, branch cooling failure warnings are issued for these multiple branches according to the secondary cooling warning rules.

[0065] In this embodiment, the node control system (BMC) uses fuzzy control to provide early warning of failure of the branch liquid cooling system based on the power consumption of each branch, the temperature difference between the inlet and outlet water of the branch, and the flow distribution coefficient of the branch.

[0066] In one optional embodiment, this application provides a flowchart of intelligent control of a rack server liquid cooling system, such as... Figure 5 As shown, the specific steps include:

[0067] Step S501: Begin.

[0068] Step S502: The cabinet control system BMC communicates with the node control system BMC to obtain the node power consumption and the temperature difference between the inlet and outlet water of the node.

[0069] Step S503: The cabinet control system BMC calculates the node flow distribution coefficient and outputs different digital signals Dm to the DAC chip.

[0070] Step S504: The DAC converts the digital signal Dm into the corresponding analog voltage signal Am and sends it to the amplifier.

[0071] Step S505: The amplifier adjusts the current magnitude according to the change in the input voltage signal and outputs it to the node solenoid valve.

[0072] Step S506: The node solenoid valve controls the flow rate of the current node path according to the received signal current, thus completing the first-level regulation.

[0073] While executing steps S504-S506, the BMC will simultaneously execute the following steps S507-S510:

[0074] Step S507: The cabinet control system (BMC) monitors the node cooling system based on node power consumption, temperature difference, and flow distribution coefficient using fuzzy control theory.

[0075] Step S508: Determine whether the node cooling has failed. If it has failed, proceed to step S509; if it has not failed, proceed to step S510.

[0076] Step S509: The cabinet control system issues a failure warning for the node cooling system.

[0077] Step S510: Continuously monitor each node.

[0078] Step S511: The node control system BMC polls to obtain the power consumption information of each branch and the temperature difference between the inlet and outlet water of the branch, calculates the branch flow distribution system, and outputs an 8-bit digital signal to the DAC chip in the node.

[0079] Step S512: The DAC within the node converts the digital signal into a corresponding analog voltage signal and sends it to the amplifier chip within the node.

[0080] Step S513: The amplifier inside the node adjusts the current according to the change in the input voltage signal and outputs it to the solenoid valve inside the node (i.e., the aforementioned branch solenoid valve).

[0081] Step S514: The solenoid valves inside the node control the coolant flow rate of each passage within the node according to the magnitude of the current received from the signal, thus completing the secondary regulation.

[0082] While executing steps S512-S514, the BMC will simultaneously execute the following steps S515-S518:

[0083] Step S515: The node control system BMC monitors the status of the branch cooling system based on the branch power consumption, temperature difference, and flow distribution coefficient using fuzzy control theory.

[0084] Step S516: Determine if the branch cooling is ineffective. If it is ineffective, proceed to step S517; if it is not ineffective, proceed to step S518.

[0085] Step S517: The rack control system issues a failure warning for the node cooling system and reports it to the rack control system BMC.

[0086] Step S518: The node control system BMC continuously monitors the branch cooling system.

[0087] Step S519, End.

[0088] Through the above embodiments, this application proposes a method and system for intelligent flow control and failure early warning of liquid-cooled servers. This is achieved by placing node solenoid valves at each server node in a rack-mounted server system to control the coolant flow at each node, and by placing branch solenoid valves in each main branch flow channel within the server node to control the distribution of coolant flow in each branch flow channel within the node. The solenoid valves are controlled by the magnitude of the electrical signal current connected to the rack control system. When the current increases, the solenoid valve controls the coolant flow in the current flow channel to increase; when the current decreases, the solenoid valve controls the coolant flow in the current flow channel to decrease.

[0089] Through two-level control and early warning, intelligent allocation and monitoring of the rack cooling system are achieved. This ensures that the heat dissipation requirements of high-power servers at all levels are met, avoiding overheating, which could affect system stability and performance and shorten lifespan; and avoiding excessive cooling during low-power periods, which would lead to wasted cooling resources, increased system energy consumption, and increased operating costs.

[0090] Through the above description of the embodiments, those skilled in the art can clearly understand that the methods according to the above embodiments can be implemented by means of software plus necessary general-purpose hardware platforms. Of course, they can also be implemented by hardware, but in many cases the former is a better implementation method. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product is stored in a storage medium (such as ROM / RAM, magnetic disk, optical disk) and includes several instructions to cause a terminal device (which may be a mobile phone, computer, server, or network device, etc.) to execute the methods described in the various embodiments of this application.

[0091] This embodiment also provides a coolant flow rate control device. Figure 6 This is a structural block diagram of a coolant flow rate control device according to an embodiment of this application, as shown below. Figure 6 As shown, the device includes:

[0092] The determining module 62 is used to determine the node flow allocation coefficient of the m server nodes based on the node information of the m server nodes at the first moment, and to determine the branch flow allocation coefficient of the n cooling branches based on the branch information of the n cooling branches of each server node at the first moment, wherein m and n are both positive integers, the node information includes: node power consumption information, node coolant inlet and outlet temperature difference, and the branch information includes: branch power consumption information, branch coolant inlet and outlet temperature difference;

[0093] The first control module 64 is used to control the node solenoid valves of the m server nodes in a first control cycle according to the node flow distribution coefficients of the m nodes, so as to control the flow rate of the first flow of the coolant flowing through the m server nodes through the node solenoid valves, wherein the first control cycle corresponds to the first moment.

[0094] The second control module 66 is used to control the branch solenoid valves of the n cooling branches within the first control cycle according to the n branch flow distribution coefficients of each server node, so as to control the flow rate of the second flow of the coolant flowing through the n cooling branches through the n branch solenoid valves.

[0095] Using the aforementioned device, firstly, the node flow distribution coefficients of the m server nodes are determined based on their node information at the first moment. Then, the branch flow distribution coefficients of each of the n cooling branch circuits of each server node are determined based on their branch information at the first moment. The node information includes node power consumption and coolant inlet / outlet temperature difference, while the branch information includes branch power consumption and coolant inlet / outlet temperature difference. Next, the node solenoid valves of the m server nodes are controlled within the first control cycle based on the m node flow distribution coefficients, thereby controlling the coolant flow rate through each server node. Finally, within each server node, the flow distribution coefficients of the n branch circuits are determined. The allocation coefficient controls the branch solenoid valves of n cooling branches during the first control cycle, thereby controlling the flow rate of coolant through each cooling branch. Using this scheme, the node solenoid valves and branch solenoid valves located within each server node are adjusted according to the cooling requirements of different server nodes and different cooling branches, thus achieving a reasonable distribution of coolant. This solves the problem in related technologies where current liquid cooling systems use a uniform cooling scheme, but the power consumption of each server node in a rack server system is inconsistent, resulting in different coolant requirements. A fixed coolant flow rate cannot reasonably adjust the instantaneous coolant demand of each node, failing to achieve reasonable coolant distribution.

[0096] Optionally, the determining module 62 is further configured to calculate the average power consumption of the m server nodes at the first time based on the node power consumption information of the m server nodes at the first time, and to calculate the average temperature difference of the m server nodes at the first time based on the node coolant inlet and outlet temperature difference of the m server nodes at the first time; using the formula Calculate the flow distribution coefficients for the m nodes, where Rm is the flow distribution coefficient of the mth server node, Pm is the power consumption information of the mth server node, Pc is the average power consumption of the node, ΔTm is the temperature difference between the inlet and outlet of the coolant of the mth server node, and ΔTc is the average temperature difference of the node.

[0097] Optionally, the first control module 64 is further configured to normalize the m node flow allocation coefficients Rm to obtain m normalized node flow allocation coefficients Rxm; convert the m normalized node flow allocation coefficients Rxm into m digital signals Dm, where Dm = Rxm * 255; send the m digital signals Dm to the DAC chip to instruct the DAC chip to convert the m digital signals Dm into m analog voltage signals Am; receive the m analog voltage signals Am sent by the DAC chip and input the m analog voltage signals Am to the amplifier to instruct the amplifier to adjust the magnitude of the signal current according to the m analog voltage signals to obtain m signal currents Cm; input the m signal currents Cm to the m node solenoid valves respectively to instruct the m node solenoid valves to adjust the valve opening size of the m node solenoid valves in the first control cycle according to the m signal currents Cm, and control the flow rate of the first flow of coolant flowing through the m server nodes through the valve opening size.

[0098] Furthermore, the aforementioned first control module 64 is also used to acquire node information of the m server nodes at a second time moment, wherein the second time moment is within the first control cycle; calculate the first power consumption difference and the first temperature difference difference of the m server nodes, wherein the first power consumption difference is the difference between the node power consumption information of the server node at the second time moment and the average power consumption of the node, and the first temperature difference difference is the difference between the node coolant inlet / outlet temperature difference of the server node at the second time moment and the average temperature difference of the node; divide the m first power consumption differences into multiple first power consumption fuzzy sets according to the node power consumption difference threshold, divide the m first temperature difference differences into multiple first temperature difference fuzzy sets according to the node temperature difference threshold, and divide the m node flow allocation coefficients into multiple first coefficient fuzzy sets according to the node allocation coefficient threshold; and, if it is determined that there is a target server node among the m server nodes that meets the first-level cooling warning rule, perform a node cooling failure warning on the target server node, wherein the first-level cooling warning rule is determined based on the multiple first power consumption fuzzy sets, the multiple first temperature difference fuzzy sets, and the multiple first coefficient fuzzy sets.

[0099] Optionally, the determining module 62 is further configured to calculate the average power consumption of the n cooling branches at the first time based on the branch power consumption information of the n cooling branches at the first time, and to calculate the average temperature difference of the n cooling branches at the first time based on the branch coolant inlet and outlet temperature difference of the n cooling branches at the first time; using the formula Calculate n branch flow distribution coefficients, where Rn is the branch flow distribution coefficient of the nth cooling branch, Pn is the branch power consumption information of the nth cooling branch, Pk is the average power consumption of the branch, ΔTn is the branch coolant inlet and outlet temperature difference of the nth cooling branch, and ΔTk is the average temperature difference of the branch.

[0100] Optionally, the second control module 66 is further configured to normalize the n branch flow distribution coefficients Rn to obtain n normalized branch flow distribution coefficients Rxn; convert the n normalized branch flow distribution coefficients Rxn into n digital signals Dn, where Dn = Rxn * 255; send the n digital signals Dn to the DAC chip to instruct the DAC chip to convert the n digital signals Dn into n analog voltage signals An; receive the n analog voltage signals An sent by the DAC chip and input the n analog voltage signals An to the amplifier to instruct the amplifier to adjust the magnitude of the signal current according to the n analog voltage signals to obtain n signal currents Cn; input the n signal currents Cn to the n branch solenoid valves respectively to instruct the n branch solenoid valves to adjust the valve opening size of the n branch solenoid valves in the first control cycle according to the n signal currents Cn, and control the flow rate of the second flow of the coolant flowing through the n cooling branches by the valve opening size.

[0101] Optionally, the second control module 66 is further configured to acquire the branch information of the n cooling branches at a third time, wherein the third time is within the first control cycle; calculate the second power consumption difference and the second temperature difference of the n cooling branches, wherein the second power consumption difference is the difference between the branch power consumption information of the cooling branch at the third time and the average power consumption of the branch, and the second temperature difference is the difference between the branch coolant inlet / outlet temperature difference of the cooling branch at the third time and the average temperature difference of the branch; and divide the n second power consumption differences into [a specific category] based on the branch power consumption difference threshold. Multiple second power consumption fuzzy sets are used; n second temperature difference values ​​are divided into multiple second temperature difference fuzzy sets according to the branch temperature difference value threshold; and n branch flow allocation coefficients are divided into multiple second coefficient fuzzy sets according to the branch allocation coefficient threshold; when it is determined that there is a target cooling branch among the n cooling branches that meets the secondary cooling warning rule, a branch cooling failure warning is issued for the target cooling branch, wherein the secondary cooling warning rule is determined based on the multiple second power consumption fuzzy sets, the multiple second temperature difference fuzzy sets, and the multiple second coefficient fuzzy sets.

[0102] It should be noted that the above modules can be implemented by software or hardware. For the latter, they can be implemented in the following ways, but are not limited to: all the above modules are located in the same processor; or, the above modules are located in different processors in any combination.

[0103] Embodiments of this application also provide a computer-readable storage medium storing a computer program, wherein the computer program is configured to execute the steps in any of the above method embodiments when run.

[0104] In one exemplary embodiment, the aforementioned computer-readable storage medium may include, but is not limited to, various media capable of storing computer programs, such as a USB flash drive, read-only memory (ROM), random access memory (RAM), portable hard disk, magnetic disk, or optical disk.

[0105] Embodiments of this application also provide an electronic device, including a memory and a processor, wherein the memory stores a computer program and the processor is configured to run the computer program to perform the steps in any of the above method embodiments.

[0106] In one exemplary embodiment, the electronic device may further include a transmission device and an input / output device, wherein the transmission device is connected to the processor and the input / output device is connected to the processor.

[0107] Embodiments of this application also provide a computer program product, including a non-volatile computer-readable storage medium storing the computer program product, wherein the computer program, when executed by a processor, implements the steps of the methods described in various embodiments of this application.

[0108] Specific examples in this embodiment can be found in the examples described in the above embodiments and exemplary implementations, and will not be repeated here.

[0109] Obviously, those skilled in the art should understand that the modules or steps of this application described above can be implemented using general-purpose computing devices. They can be centralized on a single computing device or distributed across a network of multiple computing devices. They can be implemented using computer-executable program code, and thus can be stored in a storage device for execution by a computing device. In some cases, the steps shown or described can be performed in a different order than those presented here, or they can be fabricated as separate integrated circuit modules, or multiple modules or steps can be fabricated as a single integrated circuit module. Thus, this application is not limited to any particular combination of hardware and software.

[0110] The above description is merely a preferred embodiment of this application and is not intended to limit this application. Various modifications and variations can be made to this application by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the principles of this application should be included within the protection scope of this application.

Claims

1. A method for regulating coolant flow rate, characterized in that, include: The node flow allocation coefficients of the m server nodes are determined based on the node information of the m server nodes at the first moment, and the branch flow allocation coefficients of the n cooling branches of each server node are determined based on the branch information of the n cooling branches at the first moment, where m and n are both positive integers. The node information includes: node power consumption information and node coolant inlet and outlet temperature difference. The branch information includes: branch power consumption information and branch coolant inlet and outlet temperature difference. According to the node flow distribution coefficients of m nodes, the node solenoid valves of the m server nodes are controlled in the first control cycle to control the flow rate of the first flow of coolant flowing through the m server nodes through the node solenoid valves, wherein the first control cycle corresponds to the first moment. Based on the n branch flow distribution coefficients of each server node, the branch solenoid valves of the n cooling branches are controlled within the first control cycle to control the flow rate of the second flow of the coolant flowing through the n cooling branches through the n branch solenoid valves. The step of determining the node traffic allocation coefficients of the m server nodes based on their node information at the first moment includes: Based on the power consumption information of the m server nodes at the first moment, using the formula... Calculate the average power consumption of the m server nodes at the first time point, and based on the temperature difference between the inlet and outlet coolant of the m server nodes at the first time point, use the formula... Calculate the average temperature difference of the m server nodes at the first moment, where Pm is the node power consumption information of the m-th server node, Pc is the average power consumption of the node, ΔTm is the temperature difference between the inlet and outlet of the coolant of the m-th server node, and ΔTc is the average temperature difference of the node. Through formula Calculate the traffic allocation coefficients for the m nodes, where Rm is the node traffic allocation coefficient for the m-th server node; The method further includes, after controlling the node solenoid valves of the m server nodes according to the node traffic allocation coefficients in the first control cycle: Obtain the node information of the m server nodes at a second time point, wherein the second time point is within the first control period; Calculate the first power consumption difference and the first temperature difference of the m server nodes, wherein the first power consumption difference is the difference between the node power consumption information of the server node at the second time and the average power consumption of the node, and the first temperature difference is the difference between the node coolant inlet and outlet temperature difference of the server node at the second time and the average temperature difference of the node. Based on the node power consumption difference threshold, m first power consumption differences are divided into multiple first power consumption fuzzy sets; based on the node temperature difference threshold, m first temperature difference differences are divided into multiple first temperature difference fuzzy sets; and based on the node allocation coefficient threshold, m node flow allocation coefficients are divided into multiple first coefficient fuzzy sets. The multiple first power consumption fuzzy sets include: negative large, negative medium, negative small, zero, positive small, positive medium, and positive large; the multiple first temperature difference fuzzy sets include: negative large, negative medium, negative small, zero, positive small, positive medium, and positive large; and the multiple first coefficient fuzzy sets include: small, medium, and large. If it is determined that there is a target server node among the m server nodes that meets the first-level cooling warning rule, a node cooling failure warning is issued for the target server node. The first-level cooling warning rule is determined based on the plurality of first power consumption fuzzy sets, the plurality of first temperature difference fuzzy sets, and the plurality of first coefficient fuzzy sets.

2. The method for regulating coolant flow rate according to claim 1, characterized in that, The step of controlling the node solenoid valves of the m server nodes according to the flow distribution coefficients of the m nodes within a first control cycle, so as to control the flow rate of the first flow of coolant flowing through the m server nodes through the node solenoid valves, includes: The m node traffic allocation coefficients Rm are normalized to obtain m normalized node traffic allocation coefficients Rxm; The m normalized node traffic allocation coefficients Rxm are converted into m digital signals Dm, where Dm = Rxm 255; The m digital signals Dm are sent to the DAC chip to instruct the DAC chip to convert the m digital signals Dm into m analog voltage signals Am; The system receives the m analog voltage signals Am sent by the DAC chip and inputs the m analog voltage signals Am to the amplifier to instruct the amplifier to adjust the magnitude of the signal current according to the m analog voltage signals, thereby obtaining m signal currents Cm; The m signal currents Cm are respectively input to the m node solenoid valves to instruct the m node solenoid valves to adjust the valve opening size of the m node solenoid valves in the first control cycle according to the m signal currents Cm, and control the flow rate of the first flow of coolant flowing through the m server nodes by the valve opening size.

3. The method for regulating coolant flow rate according to claim 1, characterized in that, The step of determining the flow distribution coefficients for the n cooling branches based on the branching information of the n cooling branches of each server node at the first moment includes: The average power consumption of the n cooling branches at the first time is calculated based on the branch power consumption information of the n cooling branches at the first time, and the average temperature difference of the n cooling branches at the first time is calculated based on the branch coolant inlet and outlet temperature difference of the n cooling branches at the first time. Through formula Calculate n branch flow distribution coefficients, where Rn is the branch flow distribution coefficient of the nth cooling branch, Pn is the branch power consumption information of the nth cooling branch, Pk is the average power consumption of the branch, ΔTn is the branch coolant inlet and outlet temperature difference of the nth cooling branch, and ΔTk is the average temperature difference of the branch.

4. The method for regulating coolant flow rate according to claim 3, characterized in that, The step of controlling the branch solenoid valves of the n cooling branches within the first control cycle according to the n branch flow distribution coefficients of each server node, so as to control the flow rate of the second flow of the coolant flowing through the n cooling branches through the n branch solenoid valves, includes: The n branch flow allocation coefficients Rn are normalized to obtain n normalized branch flow allocation coefficients Rxn; The n normalized branch flow allocation coefficients Rxn are converted into n digital signals Dn, where Dn = Rxn 255; The n digital signals Dn are sent to the DAC chip to instruct the DAC chip to convert the n digital signals Dn into n analog voltage signals An; The system receives the n analog voltage signals An sent by the DAC chip and inputs the n analog voltage signals An to the amplifier to instruct the amplifier to adjust the magnitude of the signal current according to the n analog voltage signals, so as to obtain n signal currents Cn; The n signal currents Cn are respectively input to the n branch solenoid valves to instruct the n branch solenoid valves to adjust the valve opening size of the n branch solenoid valves within the first control cycle according to the n signal currents Cn, and control the flow rate of the second flow of coolant flowing through the n cooling branches by the valve opening size.

5. The method for regulating coolant flow rate according to claim 3, characterized in that, After controlling the branch solenoid valves of the n cooling branches within the first control cycle according to the n branch flow allocation coefficients of each server node, the method further includes: Obtain the branch information of the n cooling branches at the third time, wherein the third time is within the first control cycle; Calculate the second power consumption difference and the second temperature difference for the n cooling branches, wherein the second power consumption difference is the difference between the branch power consumption information of the cooling branch at the third time and the average power consumption of the branch, and the second temperature difference is the difference between the branch coolant inlet and outlet temperature difference of the cooling branch at the third time and the average temperature difference of the branch. The n second power consumption differences are divided into multiple second power consumption fuzzy sets according to the branch power consumption difference threshold, the n second temperature difference differences are divided into multiple second temperature difference fuzzy sets according to the branch temperature difference threshold, and the n branch flow allocation coefficients are divided into multiple second coefficient fuzzy sets according to the branch allocation coefficient threshold. If it is determined that there is a target cooling branch among the n cooling branches that meets the secondary cooling warning rule, a branch cooling failure warning is issued for the target cooling branch. The secondary cooling warning rule is determined based on the plurality of second power consumption fuzzy sets, the plurality of second temperature difference fuzzy sets and the plurality of second coefficient fuzzy sets.

6. A device for regulating coolant flow rate, characterized in that, include: The determination module is used to determine the node flow allocation coefficient of the m server nodes based on the node information of the m server nodes at the first moment, and to determine the branch flow allocation coefficient of the n cooling branches of each server node based on the branch information of the n cooling branches at the first moment, wherein m and n are both positive integers, the node information includes: node power consumption information, node coolant inlet and outlet temperature difference, and the branch information includes: branch power consumption information, branch coolant inlet and outlet temperature difference; The first control module is used to control the node solenoid valves of the m server nodes in a first control cycle according to the node flow distribution coefficients of the m nodes, so as to control the flow rate of the first flow of the coolant flowing through the m server nodes through the node solenoid valves, wherein the first control cycle corresponds to the first moment. The second control module is used to control the branch solenoid valves of the n cooling branches within the first control cycle according to the n branch flow distribution coefficients of each server node, so as to control the flow rate of the second flow of the coolant flowing through the n cooling branches through the n branch solenoid valves. The determining module is further configured to, based on the node power consumption information of the m server nodes at the first moment, use a formula... Calculate the average power consumption of the m server nodes at the first time point, and based on the temperature difference between the inlet and outlet coolant of the m server nodes at the first time point, use the formula... Calculate the average temperature difference of the m server nodes at the first time point, where Pm is the node power consumption information of the m-th server node, Pc is the average power consumption of the node, ΔTm is the coolant inlet / outlet temperature difference of the m-th server node, and ΔTc is the average temperature difference of the node; using the formula... Calculate the traffic allocation coefficients for the m nodes, where Rm is the node traffic allocation coefficient for the m-th server node; The first control module is further configured to: acquire node information of the m server nodes at a second time moment, wherein the second time moment is within the first control period; calculate a first power consumption difference and a first temperature difference of the m server nodes, wherein the first power consumption difference is the difference between the node power consumption information of the server node at the second time moment and the average power consumption of the node, and the first temperature difference is the difference between the node coolant inlet / outlet temperature difference of the server node at the second time moment and the average temperature difference of the node; divide the m first power consumption differences into multiple first power consumption fuzzy sets according to a node power consumption difference threshold, and divide the m first temperature difference differences into multiple first temperature difference fuzzy sets according to a node temperature difference difference threshold; and Based on the node allocation coefficient threshold, the traffic allocation coefficients of the m nodes are divided into multiple first coefficient fuzzy sets. The multiple first power consumption fuzzy sets include: negative large, negative medium, negative small, zero, positive small, positive medium, and positive large. The multiple first temperature difference fuzzy sets include: negative large, negative medium, negative small, zero, positive small, positive medium, and positive large. The multiple first coefficient fuzzy sets include: small, medium, and large. If it is determined that a target server node among the m server nodes meets the first-level cooling warning rule, a node cooling failure warning is issued for the target server node. The first-level cooling warning rule is determined based on the multiple first power consumption fuzzy sets, the multiple first temperature difference fuzzy sets, and the multiple first coefficient fuzzy sets.

7. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores a computer program, wherein the computer program, when executed by a processor, implements the steps of the method according to any one of claims 1 to 5.

8. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the processor executes the computer program, it implements the steps of the method according to any one of claims 1 to 5.

Images