AI-based Energy Consumption Data Management Platform for Liquid Cooling Systems

Abstract

The present invention relates to the technical field of liquid cooling system management, and particularly relates to an energy consumption data management platform for a liquid cooling system based on artificial intelligence. The energy consumption data management platform for the liquid cooling system includes a data acquisition module for acquiring the actual energy consumption data of each server in a server group under each historical liquid cooling cycle; a first partitioning module for obtaining initial clustering clusters; a second partitioning module for reassigning the energy consumption data points in the initial clustering clusters to obtain respective target clustering clusters; and a liquid cooling system management module for reassigning and managing the servers controlled by the CDU of the liquid cooling system according to the servers corresponding to the energy consumption data points in the target clustering clusters. Moreover, the present invention can reduce the energy consumption waste during the cooling of a centralized liquid cooling system, improve the energy consumption utilization rate and energy utilization rate of the centralized liquid cooling system, and realize the balanced management of the energy consumption and resources of the centralized liquid cooling system.

Description

Technical Field

[0001] The present invention relates to the technical field of liquid cooling system management, and particularly to an energy consumption data management platform for a liquid cooling system based on artificial intelligence. Background Art

[0002] For the servers in a data center, a large amount of heat is generated during their operation. If heat dissipation cannot be carried out in time, the increase in device temperature will not only cause the chip to downclock, reduce the computing speed, trigger a protective shutdown, but also accelerate the aging of electronic components and increase the risk of hardware failures. Therefore, in order to ensure the stable operation of the data center, a liquid cooling system is usually used to dissipate heat from the servers in the data center. And considering the technical difficulty and overall cost, the liquid cooling system used in the data center is usually a centralized liquid cooling system. The centralized liquid cooling system means that the servers in the same cabinet or multiple cabinets in the data center are usually controlled by one CDU in the centralized liquid cooling system, and the servers controlled by the same CDU are connected in parallel in the same water cycle, and then a large refrigeration machine and a pump are used to uniformly regulate the water temperature and flow rate of the entire water cycle.

[0003] However, in the prior art, when determining the servers controlled or served by the CDU in the liquid cooling system, the temperature difference between the servers is not considered, that is, the energy consumption difference required for dissipating heat or cooling the servers is not considered. And the temperature of the circulating liquid is determined by the server with the highest temperature among all the parallel servers. Therefore, in the actual operation process, there will be phenomena of energy consumption waste and resource waste, resulting in low energy consumption utilization rate and resource utilization rate of the liquid cooling system. For example, if only one server needs to be cooled among the servers controlled by a certain CDU and the other servers do not need to be cooled, but at this time, in order to cool the server that needs to be cooled, all the water circulation branches controlled by this CUD will be cooled. However, this process will cause energy consumption waste and resource waste in the branches corresponding to the servers that do not need to be cooled during the cooling process. Therefore, how to manage the servers controlled by the CDU of the liquid cooling system to improve the energy consumption utilization rate and resource utilization rate of the liquid cooling system has become an urgent problem to be solved. Summary of the Invention

[0004] In order to solve the above problems, the present invention provides an energy consumption data management platform for a liquid cooling system based on artificial intelligence, and the specific technical solutions adopted are as follows:

[0005] An embodiment of the present invention provides an energy consumption data management platform for a liquid cooling system based on artificial intelligence. The power grid service data management service terminal includes:

[0006] A data acquisition module, configured to acquire the actual energy consumption data of each server in the server group during each historical liquid cooling cycle. The liquid cooling cycle is completed by a liquid cooling system, and the server group is the server group served by the liquid cooling system;

[0007] A first partitioning module, configured to record the sequence composed of the actual energy consumption data of the server during all historical liquid cooling cycles as the actual energy consumption data sequence of the corresponding server, and obtain the energy consumption data points corresponding to the server according to the mean of the actual energy consumption data sequence and the number of data greater than a preset energy consumption threshold in the actual energy consumption data sequence. Cluster the energy consumption data points according to the distance between the energy consumption data points to obtain initial clustering clusters;

[0008] A second partitioning module, configured to reallocate the energy consumption data points in the initial clustering clusters according to the total number of energy consumption data points in the initial clustering clusters and the number of servers covered by the CDU in the liquid cooling system to obtain each target clustering cluster;

[0009] A liquid cooling system management module, configured to reallocate and manage the servers controlled by the CDU of the liquid cooling system according to the servers corresponding to the energy consumption data points in the target clustering clusters.

[0010] Beneficial effects: The present invention includes a data acquisition module, configured to acquire the actual energy consumption data of each server in the server group during each historical liquid cooling cycle; a first partitioning module, configured to obtain the energy consumption data points corresponding to the server according to the mean of the actual energy consumption data sequence and the number of data greater than a preset energy consumption threshold in the actual energy consumption data sequence, and cluster the energy consumption data points according to the distance between the energy consumption data points to obtain initial clustering clusters; a second partitioning module, configured to reallocate the energy consumption data points in the initial clustering clusters according to the total number of energy consumption data points in the initial clustering clusters and the number of servers covered by the CDU in the liquid cooling system to obtain each target clustering cluster; a liquid cooling system management module, configured to reallocate and manage the servers controlled by the CDU of the liquid cooling system according to the servers corresponding to the energy consumption data points in the target clustering clusters. Moreover, the present invention can reduce the energy consumption waste and resource waste during the cooling process of the centralized liquid cooling system, improve the energy consumption utilization rate and energy utilization rate of the centralized liquid cooling system, and achieve the balanced management of the energy consumption and resources of the centralized liquid cooling system. Description of the Drawings

[0011] In order to more clearly illustrate the technical solutions and advantages in the embodiments of the present invention or the prior art, the following will briefly introduce the drawings required for use in the description of the embodiments or the prior art. Obviously, the following drawings are only some embodiments of the present invention. For those of ordinary skill in the art, other drawings can be obtained based on these drawings without creative efforts.

[0012] Figure 1 This is the structural block diagram of an energy consumption data management platform for a liquid cooling system based on artificial intelligence according to the present invention;

[0013] Figure 2 This is the schematic diagram of the liquid cooling system according to the present invention. Detailed implementation manners

[0014] Next, the technical solutions in the embodiments of the present invention will be clearly and completely described in conjunction with the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art belong to the scope protected by the embodiments of the present invention.

[0015] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by those skilled in the art belonging to the present invention.

[0016] This embodiment provides an energy consumption data management platform for a liquid cooling system based on artificial intelligence, which is described in detail as follows:

[0017] As Figure 1 shown, an energy consumption data management platform for a liquid cooling system based on artificial intelligence provided in this embodiment includes:

[0018] A data acquisition module 01, configured to acquire the actual energy consumption data of each server in the server group under each historical liquid cooling cycle.

[0019] The main purpose of this embodiment is to re - allocate the servers controlled by the CDU (Cold Distribution Unit) of the liquid cooling system according to the energy consumption required by the servers during multiple historical cycles, so as to improve the energy consumption utilization rate of the centralized liquid cooling system and achieve the balanced management of the energy consumption of the centralized liquid cooling system; and for the convenience of understanding, the subsequent description of this embodiment will be based on the re - allocation process of the servers controlled by the CDU of the centralized liquid cooling system used in any data center, that is, all servers appearing subsequently in this embodiment belong to the same data center, and all CDUs appearing subsequently are CDUs of the same centralized liquid cooling system. The CDU of the centralized liquid cooling system is a key component of the centralized liquid cooling system. The CDU of the centralized liquid cooling system serves multiple servers. The schematic diagram of the centralized liquid cooling system is as Figure 2As shown; in addition, the data center in this embodiment is an equalized cooling data center or a symmetric heat dissipation architecture data center, and the number of servers covered by the CDU of the liquid cooling system for cooling the equalized cooling data center or the symmetric heat dissipation architecture data center is the same. For example, data centers in application scenarios such as hyperscale cloud data centers and AI training clusters all belong to the equalized cooling data center or the symmetric heat dissipation architecture data center.

[0020] In this embodiment, first, the cluster composed of all servers controlled by the centralized liquid cooling system in the data center is denoted as the server group. Then, the actual energy consumption data of each server in the server group under each historical liquid cooling cycle is obtained. The actual energy consumption data is also the actual liquid cooling energy consumption of the chiller required to cool the circulation branch of each server during each historical liquid cooling cycle. The circulation branch of the server refers to a specific pipeline branch in the liquid cooling system for the coolant to flow and complete heat transfer. Its core function is to efficiently export the heat generated by the server through liquid circulation. And the liquid cooling cycles performed before the current moment are all called historical liquid cooling cycles. A liquid cooling cycle of a server refers to a liquid cooling cycle process of controlling the CDU corresponding to the server, or a liquid cooling cycle of a server refers to a closed circulation path formed by the coolant between the external cooling equipment (such as cooling towers and chillers) and the CDU. Moreover, a liquid cooling cycle of the CDU also refers to the closed circulation path of the coolant between the external cooling equipment and the CDU. The core function is to transfer the heat generated inside the data center to the external environment through liquid-liquid heat exchange to complete the final heat dissipation. Thus, it can be seen that a liquid cooling cycle of the server controlled by the CDU refers to the whole process of the coolant completing "flowing - absorbing heat - dissipating heat - returning" in the primary side loop of the corresponding CDU, that is, a liquid cooling cycle of the server controlled by the CDU refers to the process in which the coolant starts from the corresponding CDU, flows through the external cooling equipment to complete heat dissipation, and then returns to the CDU.

[0021] In addition, it should be noted that, in order to avoid energy waste and ensure the cooling effect, when designing a centralized liquid cooling system for a data center, the number of servers covered by each CDU of the centralized liquid cooling system is generally determined comprehensively in combination with factors such as server power consumption density, cooling requirements, and system architecture design. That is, in actual applications, the number of servers covered by the CDU needs to be set according to the configuration scenario of the actual business scenario. For example, if the application scenario of the data center in this embodiment is an AI training cluster, which mainly conducts large model training or scientific computing, then the number of servers covered by the CDU in this embodiment is generally set to 8. However, when designing the CDU interface, redundancy reservation is generally required, and the redundant reserved interface is generally reserved as a standby loop. If the number of servers covered by a certain CDU is 8, then this CDU is generally configured with 10 interfaces, and the remaining 2 interfaces are reserved for expansion or redundancy.

[0022] For the convenience of understanding, the following description will be given by taking the specific acquisition process of the actual energy consumption data of the b-th server in the a-th historical liquid cooling cycle in the server group as an example. The a-th historical liquid cooling cycle also refers to the a-th historical liquid cooling cycle of the CDU that controls or serves the b-th server during the operation historical time period of the centralized liquid cooling system. The operation historical time period of the centralized liquid cooling system refers to the operation time period before the current moment. Then the specific acquisition process of the actual energy consumption data of the b-th server in the a-th historical liquid cooling cycle is as follows:

[0023] First, obtain the cold liquid flow rate, cold liquid density, and cold liquid specific heat capacity in the circulation branch of the b-th server during the a-th historical liquid cooling cycle. The cold liquid refers to the cooling medium or the liquid used for cooling in the liquid cooling system. In this embodiment, the cold liquid is water, the specific heat capacity of water is 4.2×10³ J / (kg·°C), the density of water is 1×10³ kg / m³, and the cold liquid flow rate in the circulation branch refers to the water flow rate in the circulation branch, which is collected by a flow sensor installed on the circulation branch pipeline. Then, obtain the temperature value of the b-th server before the start of the a-th historical liquid cooling cycle by the CDU serving the b-th server, and denote it as Z1. The temperature value of the server can be collected by a temperature sensor configured inside the corresponding server, or the cold plate temperature of the server can be used as the temperature value of the server. The cold plate temperature of the server is usually collected by a temperature sensor installed on the cold plate of the server. Then, obtain the CPU safe temperature value of the b-th server, and denote it as Z0. The CPU safe temperature value of the server refers to the maximum temperature value that can be tolerated. Once the CPU safe temperature value of the server is exceeded or reached, cooling is required, and it is determined whether Z1 is greater than Z0. If so, the result of Z1 - Z0 is used as the temperature that needs to be reduced for the b-th server during the a-th historical liquid cooling cycle. Otherwise, 0 is used as the temperature that needs to be reduced for the b-th server during the a-th historical liquid cooling cycle. After that, obtain the overall cycle time of the CDU serving the b-th server during the a-th historical liquid cooling cycle, and denote it as the overall cycle time of the b-th server during the a-th historical liquid cooling cycle. Immediately afterwards, obtain the refrigeration energy efficiency ratio of the centralized liquid cooling system, and denote the reciprocal of the refrigeration energy efficiency ratio of the centralized liquid cooling system as the characteristic ratio. The refrigeration energy efficiency ratio is generally determined by dividing the heat removed by the refrigerator by the energy consumed by the refrigerator. In this embodiment, according to the prior art, the refrigeration energy efficiency ratio is set to 3.6. Then, multiply the cold liquid flow rate, cold liquid density, cold liquid specific heat capacity, the temperature that needs to be reduced for the b-th server during the a-th historical liquid cooling cycle, the characteristic ratio, and the overall cycle time of the b-th server during the a-th historical liquid cooling cycle during the a-th historical liquid cooling cycle of the b-th server, and the result is used as the actual energy consumption data of the b-th server during the a-th historical liquid cooling cycle, that is, the actual energy consumption data of the b-th server during the a-th historical liquid cooling cycle is , where is the cold liquid specific heat capacity in the circulation branch of the b-th server, is the temperature that needs to be reduced for the b-th server during the a-th historical liquid cooling cycle, f is the cold liquid flow rate in the circulation branch of the b-th server, t is the overall cycle time of the b-th server during the a-th historical liquid cooling cycle, is the cold liquid density in the circulation branch of the b-th server, is the refrigeration energy efficiency ratio of the liquid cooling system; and the calculation process of the actual energy consumption data is determined according to the energy calculation process.

[0024] In addition, in this embodiment, the purpose of obtaining the actual energy consumption data is to allocate servers with similar actual energy consumption characteristics to the same CDU when performing server reallocation subsequently, so as to improve the energy consumption utilization rate.

[0025] Therefore, through the above process, this embodiment can obtain the actual energy consumption data of each server in the server group under each historical liquid cooling cycle.

[0026] The first partitioning module 02 is configured to record the sequence composed of the actual energy consumption data of the server under all historical liquid cooling cycles as the actual energy consumption data sequence of the corresponding server, and obtain the energy consumption data points corresponding to the server according to the mean value of the actual energy consumption data sequence and the number of data greater than the preset energy consumption threshold in the actual energy consumption data sequence, and cluster the energy consumption data points according to the distances between the energy consumption data points to obtain initial clustering clusters.

[0027] Since the actual energy consumption data under a single cycle is difficult to characterize the energy consumption characteristics of the server, this embodiment will next characterize the energy consumption characteristics of the server according to the actual energy consumption data of the server under multiple historical cycles and perform clustering to obtain initial clustering clusters. That is, this embodiment first sorts the actual energy consumption data of each server under all historical liquid cooling cycles in the order of the liquid cooling cycle time, and records the sorted sequence as the actual energy consumption data sequence of the corresponding server; and as other real-time methods, this embodiment can also record the sequence composed of the actual energy consumption data of each server under the most recent N historical liquid cooling cycles as the actual energy consumption data sequence of the corresponding server, and the value of N generally needs to be set by the implementer according to the actual historical cycle times and experimental statistics. Generally, the value of N is set to 50, that is, the actual energy consumption requirements under 50 historical liquid cooling cycles can better describe the change of the server's energy consumption requirements in the near future.

[0028] After obtaining the actual energy consumption data sequence of the server, this embodiment will next obtain the energy consumption data points corresponding to each server according to the mean value of the actual energy consumption data sequence of the server and the number of data greater than the preset energy consumption threshold in the actual energy consumption data sequence of the server, and the specific process of obtaining the energy consumption data points corresponding to the server is as follows:

[0029] First, obtain the mean of the actual energy consumption data sequences of each server, and denote it as the historical actual energy consumption data of the corresponding server. Then, in the actual energy consumption data sequence of each server, count the total number of actual energy consumption data greater than the preset energy consumption threshold of the corresponding server, and denote it as the actual cooling frequency of the corresponding server. For example, if there are n actual energy consumption data greater than the preset energy consumption threshold of a certain server in the actual energy consumption data sequence of the server, then the actual cooling frequency of the server is n. In addition, the preset energy consumption threshold in this embodiment is 0, because when the temperature that needs to be reduced during the historical liquid cooling cycle of the server is 0, it indicates that the actual energy consumption value required during the cycle is also 0, so the calculated actual energy consumption data is 0.

[0030] Immediately construct a two-dimensional mapping space. The abscissa of the two-dimensional mapping space is the actual cooling frequency, and the ordinate axis is the historical actual energy consumption data. Then map the historical actual energy consumption data and the actual energy consumption frequency of each server into the two-dimensional mapping space to obtain the energy consumption data points corresponding to the corresponding servers. The abscissa value of the energy consumption data point corresponding to the server is the actual cooling frequency of the corresponding server, and the ordinate value is the historical actual energy consumption data of the corresponding server, that is, one server corresponds to one energy consumption data point.

[0031] After obtaining the energy consumption data points corresponding to the servers, calculate the distances between the energy consumption data points. For the i-th energy consumption data point and the j-th energy consumption data point in the two-dimensional mapping space, the distance between the i-th energy consumption data point and the j-th energy consumption data point refers to the Euclidean distance between the i-th energy consumption data point and the j-th energy consumption data point, where i is not equal to j. Then, according to the distances between the energy consumption data points, perform K-means clustering on all the energy consumption data points in the two-dimensional mapping space, and denote all the clustering clusters obtained as initial clustering clusters. One energy consumption data point in the initial clustering cluster represents one server. And when performing K-means clustering, the clustering metric distance is the Euclidean distance between the energy consumption data points. In this embodiment, the number of cluster centers during K-means clustering is the same as the number of CDUs in the centralized liquid cooling system. The number of CDUs needs to be set according to the total number of servers covered by the centralized liquid cooling system and the number of servers covered by one CDU. The total number of servers covered by the liquid cooling system is the number of servers in the server group in this embodiment. For example, if the number of servers covered by one CDU is A1 and the number of servers in the server group is A0, then the number of CDUs in the centralized liquid cooling system is generally the ceiling value of the result obtained by dividing A0 by A1.

[0032] Therefore, through the above process, this embodiment can obtain the initial clustering clusters.

[0033] The second partitioning module 03 is configured to reassign the energy consumption data points in the initial clustering clusters according to the total number of energy consumption data points in the initial clustering clusters and the number of servers covered by the CDU in the liquid cooling system, so as to obtain each target clustering cluster.

[0034] Since the data center in this embodiment is a balanced cooling data center or a symmetric heat dissipation architecture data center, it is required that the number of servers covered or served by each CDU in the centralized liquid cooling system in this embodiment is the same. However, the total number of energy consumption data points in the obtained initial clustering clusters may be greater than or less than the number of servers covered by the set CDU. Therefore, in the following, the energy consumption data points in the initial clustering clusters need to be reassigned according to the total number of energy consumption data points in the initial clustering clusters and the number of servers covered by the CDU in the liquid cooling system to obtain each target clustering cluster. The specific process of reassigning the energy consumption data points in the initial clustering clusters according to the total number of energy consumption data points in the initial clustering clusters and the number of servers covered by the CDU in the liquid cooling system to obtain each target clustering cluster is as follows:

[0035] First, the number of servers covered by a single CDU in the centralized liquid cooling system is used as the quantity judgment threshold; then it is judged whether the total number of energy consumption data points in the initial clustering cluster is greater than the quantity judgment threshold. If so, it indicates that there are redundant energy consumption data points in the corresponding initial clustering cluster, and the redundant energy consumption data points in such initial clustering clusters need to be assigned to other initial clustering clusters. Then, such initial clustering clusters are recorded as surplus clustering clusters; then it is continued to judge whether the total number of energy consumption data points in the initial clustering cluster is less than the quantity judgment threshold. If so, it indicates that there is a lack of energy consumption data points in the corresponding initial clustering cluster, and such initial clustering clusters need to receive the assigned energy consumption data points. Then, such initial clustering clusters are recorded as lack clustering clusters; then it is continued to judge whether the total number of energy consumption data points in the initial clustering cluster is equal to the quantity judgment threshold. If so, it indicates that the number of energy consumption data points in the corresponding initial clustering cluster is appropriate, and the corresponding initial clustering cluster does not need to receive other energy consumption data points, nor does it need to reassign the energy consumption data points in the corresponding initial clustering cluster to other clusters. Then, such initial clustering clusters are recorded as target clustering clusters.

[0036] Next, in this embodiment, it is necessary to re - allocate the obtained surplus clustering clusters and lack clustering clusters to obtain target clustering clusters. And in this embodiment, according to the distance between each energy - consumption data point in the surplus clustering clusters and the clustering center point of the lack clustering clusters, and the actual energy - consumption data sequence of the servers corresponding to each energy - consumption data point in the surplus clustering clusters, the re - allocation of the surplus clustering clusters and the lack clustering clusters is carried out. Then, in this embodiment, the process of re - allocating the surplus clustering clusters and the lack clustering clusters to obtain target clustering clusters according to the distance between each energy - consumption data point in the surplus clustering clusters and the clustering center point of the lack clustering clusters, and the actual energy - consumption data sequence of the servers corresponding to each energy - consumption data point in the surplus clustering clusters is as follows:

[0037] First, according to the distance between each energy - consumption data point in the surplus clustering clusters and the clustering center point of the lack clustering clusters, and the difference between adjacent actual energy - consumption data in the actual energy - consumption data sequence of the servers corresponding to each energy - consumption data point in the surplus clustering clusters, obtain the degree of being to be allocated corresponding to each energy - consumption data point in each surplus clustering cluster; then, according to the degree of being to be allocated corresponding to each energy - consumption data point in each surplus clustering cluster and the quantity judgment threshold, obtain the sequence of data points to be allocated corresponding to each surplus clustering cluster; after that, according to the sequence of data points to be allocated corresponding to each surplus clustering cluster, re - allocate the surplus clustering clusters and the lack clustering clusters to obtain the allocated clustering clusters corresponding to each surplus clustering cluster and the received clustering clusters corresponding to each lack clustering cluster, and record all the allocated clustering clusters and all the received clustering clusters as target clustering clusters.

[0038] In this embodiment, the specific process of obtaining the degree of being to be allocated corresponding to each energy - consumption data point in each surplus clustering cluster according to the distance between each energy - consumption data point in the surplus clustering clusters and the clustering center point of the lack clustering clusters, and the difference between adjacent actual energy - consumption data in the actual energy - consumption data sequence of the servers corresponding to each energy - consumption data point in the surplus clustering clusters is as follows:

[0039] For the g-th energy consumption data point in any surplus clustering cluster G: First, obtain the server corresponding to the g-th energy consumption data point, and denote the actual energy consumption data sequence of the server corresponding to the g-th energy consumption data point as the sequence to be analyzed. Then, obtain the difference sequence of the sequence to be analyzed, and denote the sum of all differences in the difference sequence as the comprehensive difference. Perform a negative correlation mapping on the comprehensive difference, and denote the mapping result as the credibility of the g-th energy consumption data point. And the h-th difference in the difference sequence is the absolute value of the difference between the h-th data and the (h + 1)-th data in the sequence to be analyzed. The smaller the comprehensive difference, the higher the credibility of the sequence to be analyzed, and then the more credible the initial clustering cluster to which the g-th energy consumption data point belongs obtained by clustering; After that, obtain the Euclidean distance between the cluster center point of each lack clustering cluster and the cluster center point of the surplus clustering cluster G, and denote it as the measurement distance between the corresponding lack clustering cluster and the surplus clustering cluster G. And use the cluster center point of the lack clustering cluster corresponding to the minimum measurement distance as the nearest neighbor cluster center point of the surplus clustering cluster G, that is, the cluster center point of the lack clustering cluster corresponding to the minimum measurement distance is the cluster center point of the lack clustering cluster closest to the surplus clustering cluster G; Immediately calculate the Euclidean distance between the g-th energy consumption data point and the nearest neighbor cluster center point of the surplus clustering cluster G, and denote it as the nearest neighbor cluster distance of the g-th energy consumption data point; Then, obtain the reciprocal of the credibility of the g-th energy consumption data point and the reciprocal of the nearest neighbor cluster distance of the g-th energy consumption data point, and denote the result of multiplying the reciprocal of the credibility of the g-th energy consumption data point by the reciprocal of the nearest neighbor cluster distance of the g-th energy consumption data point as the degree of being to be assigned corresponding to the g-th energy consumption data point in the surplus clustering cluster G, that is, the degree of being to be assigned corresponding to the g-th energy consumption data point is , where R0 is the credibility of the g-th energy consumption data point, R1 is the nearest neighbor cluster distance of the g-th energy consumption data point, the credibility of the g-th energy consumption data point is exp(-D0), D0 is the comprehensive difference, and exp() is the exponential function with the constant e as the base; Moreover, when the degree of being to be assigned corresponding to the g-th energy consumption data point is larger, it indicates that the corresponding energy consumption data point is more suitable to be assigned to the lack clustering cluster. On the contrary, when the degree of being to be assigned corresponding to the g-th energy consumption data point is smaller, it indicates that the corresponding energy consumption data point is more suitable to be retained in the current initial clustering cluster. The larger R0 is, the greater the credibility of the g-th energy consumption data point is, and then the corresponding energy consumption data point is more suitable to be retained in the current initial clustering cluster. The larger R1 is, the farther the g-th energy consumption data point is from the lack clustering cluster, and then the corresponding energy consumption data point is more suitable to be retained in the current initial clustering cluster.

[0040] In this embodiment, the specific process of obtaining the sequence of data points to be allocated corresponding to each surplus clustering cluster according to the degree of allocation to be determined and the quantity judgment threshold corresponding to each energy consumption data point in each surplus clustering cluster is as follows: For any surplus clustering cluster G: First, sort all the energy consumption data points in the surplus clustering cluster G in descending order of the degree of allocation to be determined, and denote the sorted sequence as the sequence of energy consumption data points corresponding to the surplus clustering cluster G; Then, obtain the result of subtracting the quantity judgment threshold from the total number of energy consumption data points in the surplus clustering cluster G, and denote it as the number value of redundant data points corresponding to the surplus clustering cluster G. If the total number of energy consumption data points in the surplus clustering cluster G is 10 and the quantity judgment threshold is 8, then the number value of redundant data points corresponding to the surplus clustering cluster G is 2; After that, in the sequence of energy consumption data points corresponding to the surplus clustering cluster G, obtain the first M consecutive energy consumption data points, and denote the sequence formed by the first M consecutive energy consumption data points obtained as the sequence of data points to be allocated corresponding to the surplus clustering cluster G, where M is the number value of redundant data points corresponding to the surplus clustering cluster G. All the data points in the sequence of data points to be allocated need to be reallocated to the lack clustering clusters, and the data points in the sequence of data points to be allocated corresponding to the surplus clustering cluster G are all the data points to be allocated in the surplus clustering cluster G.

[0041] And the allocation principle in this embodiment is: Determine the order of allocation of surplus clustering clusters according to the number of data points to be allocated, determine the order of receiving data points to be allocated of lack clustering clusters according to the distance from the surplus clustering clusters, determine the number of data points to be allocated that a lack clustering cluster can receive according to the difference between the total number of energy consumption data points in the lack clustering cluster and the quantity judgment threshold, determine the number of data points to be allocated that a surplus clustering cluster needs to allocate according to the number of data points to be allocated in the surplus clustering cluster, and determine the allocation order of the data points to be allocated in the surplus clustering cluster according to the sequence of data points to be allocated corresponding to the surplus clustering cluster, that is, the more the number of data points to be allocated, the earlier the data points to be allocated in the corresponding surplus clustering cluster are allocated, the closer the lack clustering cluster is to the surplus clustering cluster, the earlier it receives the data points to be allocated in the corresponding surplus clustering cluster, the greater the difference between the total number of energy consumption data points in the lack clustering cluster and the quantity judgment threshold, the more data points the corresponding lack clustering cluster needs to receive, the more the number of data points to be allocated in the surplus clustering cluster, the more data points the corresponding surplus clustering cluster needs to allocate, and the earlier the data points in the sequence of data points to be allocated corresponding to the surplus clustering cluster are allocated; In addition, as other real-time methods, other allocation principles can also be set, for example, the fewer the number of data points to be allocated, the earlier the data points to be allocated in the corresponding surplus clustering cluster are allocated, or the order of allocation of surplus clustering clusters is determined according to the principle of random selection.

[0042] In this embodiment, according to the set allocation principle, and then based on the sequence of data points to be allocated corresponding to each surplus cluster, the surplus clusters and the lack clusters are re-allocated. The specific process of obtaining the allocated cluster corresponding to each surplus cluster and the received cluster corresponding to each lack cluster is as follows:

[0043] First, obtain the sequence of lack clusters corresponding to each surplus cluster. The lack clusters in the sequence of lack clusters corresponding to a surplus cluster are arranged according to the distance from the corresponding surplus cluster. That is, in the sequence of lack clusters corresponding to any surplus cluster, the lack cluster ranked earlier is closer to the surplus cluster, or in other words, the measurement distance between the lack cluster ranked earlier and the surplus cluster in the sequence of lack clusters corresponding to any surplus cluster is smaller. Then, obtain the starting reception quantity of each lack cluster. The starting reception quantity of a lack cluster refers to the number of data points missing in the corresponding cluster when the corresponding lack cluster has not started receiving the data points to be allocated. That is, the starting reception quantity of any lack cluster is the result of subtracting the total number of original data points in the lack cluster from the quantity judgment threshold. The original data points in the lack cluster refer to the energy consumption data points divided into the corresponding lack cluster through K-means clustering.

[0044] Then, determine whether the total number of data points to be allocated in the first surplus cluster in the cluster sequence is less than or equal to the starting reception quantity of the first lack cluster in the sequence of lack clusters corresponding to the first surplus cluster. If so, allocate all the data points to be allocated in the first surplus cluster to the first lack cluster in the sequence of lack clusters corresponding to the first surplus cluster, and allocate all the data points to be allocated in the first surplus cluster to the first surplus cluster after the first lack cluster in the sequence of lack clusters corresponding to the first surplus cluster, which is recorded as the allocated cluster corresponding to the first surplus cluster. That is, the allocated cluster corresponding to the first surplus cluster refers to the cluster after all the data points to be allocated in the first surplus cluster are allocated to the lack cluster. Then, update the starting reception quantity of each lack cluster according to the number of data points to be allocated belonging to the first surplus cluster received by the lack cluster to obtain the first updated reception quantity of each lack cluster.

[0045] Next, continue to determine whether the total number of data points to be assigned in the second surplus cluster in the cluster sequence is less than or equal to the first updated reception quantity of the first lack cluster in the lack cluster sequence corresponding to the second surplus cluster. If not, then determine whether the result of adding the first updated reception quantity of the first lack cluster in the lack cluster sequence corresponding to the second surplus cluster and the first updated reception quantity of the second lack cluster in the lack cluster sequence corresponding to the second surplus cluster is greater than or equal to the total number of data points to be assigned in the second surplus cluster. If so, then allocate the first V1 data points to be assigned in the sequence of data points to be assigned corresponding to the second surplus cluster to the first lack cluster in the lack cluster sequence corresponding to the second surplus cluster, and allocate all the data points to be assigned after the V1th data point to be assigned in the sequence of data points to be assigned corresponding to the second surplus cluster to the second lack cluster in the lack cluster sequence corresponding to the second surplus cluster. And after allocating the first V1 data points to be assigned in the sequence of data points to be assigned corresponding to the second surplus cluster to the first lack cluster in the lack cluster sequence corresponding to the second surplus cluster and allocating all the data points to be assigned after the V1th data point to be assigned in the sequence of data points to be assigned corresponding to the second surplus cluster to the second lack cluster in the lack cluster sequence corresponding to the second surplus cluster, the second surplus cluster is denoted as the second surplus cluster with assigned clusters, that is, the assigned cluster corresponding to the second surplus cluster refers to the cluster after all the data points to be assigned in the second surplus cluster are allocated to the lack clusters. Then, update the first updated reception quantity of each lack cluster to obtain the second updated reception quantity of each lack cluster according to the quantity of the data points to be assigned belonging to the second surplus cluster received by the lack clusters.

[0046] Next, continue to determine whether the total number of data points to be assigned in the third surplus cluster in the cluster sequence is less than or equal to the first updated reception quantity of the first lack cluster in the lack cluster sequence corresponding to the third surplus cluster. If not, determine whether the sum of the second updated reception quantity of the first lack cluster in the lack cluster sequence corresponding to the third surplus cluster and the second updated reception quantity of the second lack cluster in the lack cluster sequence corresponding to the third surplus cluster is greater than or equal to the total number of data points to be assigned in the third surplus cluster. If not yet, continue to determine whether the sum of the second updated reception quantity of the first lack cluster in the lack cluster sequence corresponding to the third surplus cluster, the second updated reception quantity of the second lack cluster in the lack cluster sequence corresponding to the third surplus cluster, and the second updated reception quantity of the third lack cluster in the lack cluster sequence corresponding to the third surplus cluster is greater than or equal to the total number of data points to be assigned in the third surplus cluster. If so, assign the first Q1 data points to be assigned at the front of the data point sequence to be assigned corresponding to the third surplus cluster to the first lack cluster in the lack cluster sequence corresponding to the third surplus cluster, assign all the data points to be assigned between the (Q1 + 1)-th data point to be assigned and the Q2-th data point to be assigned in the data point sequence to be assigned corresponding to the third surplus cluster to the second lack cluster in the lack cluster sequence corresponding to the third surplus cluster. The data points assigned to the second lack cluster include the (Q1 + 1)-th data point to be assigned and the Q2-th data point to be assigned. Assign all the data points to be assigned after the Q2-th data point to be assigned in the data point sequence to be assigned corresponding to the third surplus cluster to the third lack cluster in the lack cluster sequence corresponding to the third surplus cluster. Q1 is the second updated reception quantity of the first lack cluster in the lack cluster sequence corresponding to the third surplus cluster, Q2 is the second updated reception quantity of the second lack cluster in the lack cluster sequence corresponding to the second surplus cluster, the Q2-th data point to be assigned belongs to the data point sequence to be assigned corresponding to the third surplus cluster, and mark the cluster after all the data points to be assigned in the third surplus cluster are assigned to the lack clusters as the assigned cluster corresponding to the third surplus cluster. Then, update the second updated reception quantity of each lack cluster to obtain the third updated reception quantity of each lack cluster according to the number of data points to be assigned belonging to the third surplus cluster received by the lack cluster; and complete the assignment of all surplus clusters according to the above process. Moreover, after obtaining the assigned cluster corresponding to the last surplus cluster in the cluster sequence, determine that the assignment is completed.

[0047] In addition, after the allocation is completed, the cluster formed by all the data points to be allocated received by each lacking cluster during the allocation process and all the original data points in the corresponding lacking cluster is denoted as the received cluster corresponding to the lacking cluster. That is, for any lacking cluster, the cluster formed by all the data points to be allocated received by the lacking cluster during the allocation process and all the original data points in the lacking cluster is the received cluster corresponding to the lacking cluster.

[0048] In this embodiment, the specific process of updating the starting reception quantity of the lacking cluster to obtain the first updated reception quantity of the corresponding lacking cluster, updating the first updated reception quantity of the lacking cluster to obtain the second updated reception quantity of the corresponding lacking cluster, and updating the second updated reception quantity of each lacking cluster to obtain the third updated reception quantity of each lacking cluster is as follows:

[0049] For any lack of cluster F: If the lack of cluster F receives the data points to be assigned allocated by the first surplus cluster, then the result of subtracting the number of data points to be assigned belonging to the first surplus cluster received by the lack of cluster F from the starting reception quantity of the lack of cluster F is denoted as the first updated reception quantity of the lack of cluster F. If the lack of cluster F does not receive the data points to be assigned belonging to the first surplus cluster, then the starting reception quantity of the lack of cluster F is used as the first updated reception quantity of the lack of cluster F. If the lack of cluster F receives the data points to be assigned allocated by the second surplus cluster, then the result of subtracting the number of data points to be assigned belonging to the second surplus cluster received by the lack of cluster F from the first updated reception quantity of the lack of cluster F is denoted as the second updated reception quantity of the lack of cluster F. If the lack of cluster F does not receive the data points to be assigned belonging to the second surplus cluster, then the first updated reception quantity of the lack of cluster F is used as the second updated reception quantity of the lack of cluster F. Similarly, if the lack of cluster F receives the data points to be assigned allocated by the third surplus cluster, then the result of subtracting the number of data points to be assigned belonging to the third surplus cluster received by the lack of cluster F from the second updated reception quantity of the lack of cluster F is denoted as the third updated reception quantity of the lack of cluster F. If the lack of cluster F does not receive the data points to be assigned belonging to the third surplus cluster, then the second updated reception quantity of the lack of cluster F is used as the third updated reception quantity of the lack of cluster F. For example, if the starting reception quantity of the lack of cluster F is 3 and the lack of cluster F receives 2 data points to be assigned allocated by the first surplus cluster during the process of assigning the data points to be assigned in the first surplus cluster, then the first updated reception quantity of the lack of cluster F is 1. The purpose of updating the reception quantity of the lack of cluster is to ensure that the data points received by all lack of clusters do not exceed the maximum limit that the lack of clusters can receive.

[0050] Example: If the number of servers covered by the CDU is 8, that is, the quantity judgment threshold is 8. If the number of surplus clustering clusters is 3, namely surplus clustering cluster W1, surplus clustering cluster W2, and surplus clustering cluster W3. If the number of lacking clustering clusters is 3, namely lacking clustering cluster U1, lacking clustering cluster U2, and lacking clustering cluster U3. If the total number of data points to be allocated in surplus clustering cluster W1 is 6, the total number of data points to be allocated in surplus clustering cluster W2 is 4, and the total number of data points to be allocated in surplus clustering cluster W3 is 2. The starting receiving quantity of lacking clustering cluster U1 is 5, the starting receiving quantity of lacking clustering cluster U2 is 4, and the starting receiving quantity of lacking clustering cluster U3 is 3. The sequence of lacking clustering clusters corresponding to surplus clustering cluster W1 is {U2, U1, U3}, the sequence of lacking clustering clusters corresponding to surplus clustering cluster W2 is {U1, U3, U2}, and the sequence of lacking clustering clusters corresponding to surplus clustering cluster W3 is {U3, U2, U1}. Then the allocation process of surplus clustering cluster W1 is as follows: First, allocate the first 4 data points to be allocated in the sequence of data points to be allocated corresponding to surplus clustering cluster W1 to U2, then allocate the last 2 data points to be allocated in the sequence of data points to be allocated corresponding to surplus clustering cluster W1 to U1, and then obtain the first updated receiving quantity of the lacking clustering cluster according to the process of allocating the data points to be allocated in surplus clustering cluster W1. At this time, the first updated receiving quantity of lacking clustering cluster U1 is 3, the first updated receiving quantity of lacking clustering cluster U2 is 0, and the first updated receiving quantity of lacking clustering cluster U3 is 3. The allocation process of surplus clustering cluster W2 is as follows: First, allocate the first 3 data points to be allocated in the sequence of data points to be allocated corresponding to surplus clustering cluster W2 to U1, then allocate the last 1 data point to be allocated in the sequence of data points to be allocated corresponding to surplus clustering cluster W2 to U3, and then obtain the second updated receiving quantity of the lacking clustering cluster according to the process of allocating the data points to be allocated in surplus clustering cluster W2. At this time, the second updated receiving quantity of lacking clustering cluster U1 is 0, the second updated receiving quantity of lacking clustering cluster U2 is 0, and the first updated receiving quantity of lacking clustering cluster U3 is 2. The allocation process of surplus clustering cluster W3 is: Just allocate all the data points to be allocated in the sequence of data points to be allocated corresponding to surplus clustering cluster W3 to U3;The assigned clusters corresponding to the surplus clusters W1, W2, and W3 are the clusters obtained by allocating all the data points to be assigned in the corresponding surplus clusters to the lack clusters. The received cluster corresponding to the lack cluster U1 is composed of all the original data points in the lack cluster U1, the first 3 data points to be assigned in the sequence of data points to be assigned corresponding to the surplus cluster W2, and the last 2 data points to be assigned in the sequence of data points to be assigned corresponding to the surplus cluster W1. The received cluster corresponding to the lack cluster U2 is composed of the first 4 data points to be assigned in the sequence of data points to be assigned corresponding to the surplus cluster W1 and all the original data points in the lack cluster U2. The received cluster corresponding to the lack cluster U3 is composed of all the original data points in the lack cluster U3, the last 1 data point to be assigned in the sequence of data points to be assigned corresponding to the surplus cluster W2, and all the data points to be assigned in the sequence of data points to be assigned corresponding to the surplus cluster W3.

[0051] Therefore, all the target clusters are obtained through the above process in this embodiment.

[0052] The liquid cooling system management module 04 is used to reallocate and manage the servers controlled by the CDU of the liquid cooling system according to the servers corresponding to the energy consumption data points in the target clusters.

[0053] After obtaining the target clusters, all the servers corresponding to the energy consumption data points in the target clusters are connected to the CDU of the same liquid cooling system; that is, the servers corresponding to all the energy consumption data points in the same target cluster are connected to the same CDU of the centralized liquid cooling system using CDU quick connectors. This method of connecting the servers of the same target cluster to the same CDU can reduce the energy consumption waste and energy waste during the cooling process of the centralized liquid cooling system, improve the energy consumption utilization rate and energy utilization rate of the centralized liquid cooling system, and achieve the balanced management of the energy consumption and energy of the centralized liquid cooling system.

[0054] So far, this embodiment has completed the reallocation of the servers controlled by the CDU, reduced the energy consumption waste and energy waste during the cooling process of the centralized liquid cooling system, improved the energy consumption utilization rate and energy utilization rate of the centralized liquid cooling system, and achieved the balanced management of the energy consumption and energy of the centralized liquid cooling system.

[0055] In summary, this embodiment includes a data acquisition module for acquiring the actual energy consumption data of each server in the server group during each historical liquid cooling cycle; a first partitioning module for obtaining energy consumption data points corresponding to the servers according to the mean of the actual energy consumption data sequence and the number of data in the actual energy consumption data sequence that is greater than a preset energy consumption threshold, and clustering the energy consumption data points according to the distances between the energy consumption data points to obtain initial clustering clusters; a second partitioning module for reassigning the energy consumption data points in the initial clustering clusters according to the number of energy consumption data points in the initial clustering clusters and the number of servers covered by the CDU in the liquid cooling system to obtain each target clustering cluster; and a liquid cooling system management module for reassigning and managing the servers controlled by the CDU of the liquid cooling system according to the servers corresponding to the energy consumption data points in the target clustering clusters. Moreover, this embodiment can reduce the energy consumption waste and energy waste during the cooling process of the centralized liquid cooling system, improve the energy consumption utilization rate and energy utilization rate of the centralized liquid cooling system, and achieve the balanced management of the energy consumption and energy of the centralized liquid cooling system.

[0056] The above-described embodiments are only used to illustrate the technical solutions of the present application, rather than to limit them; although the present application has been described in detail with reference to the foregoing embodiments, those of ordinary skill in the art should understand that they can still modify the technical solutions described in the foregoing embodiments, or perform equivalent replacements for some of the technical features; and these modifications or replacements do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the present application in each embodiment, and should all be included in the protection scope of the present application.

Claims

1. An energy consumption data management platform for a liquid cooling system based on artificial intelligence, characterized in that, The energy consumption data management platform of the liquid cooling system includes: A data acquisition module, which is used to acquire the actual energy consumption data of each server in the server group during each historical liquid cooling cycle. The liquid cooling cycle is completed by the liquid cooling system, and the server group is the server group served by the liquid cooling system; A first partitioning module, which is used to record the sequence composed of the actual energy consumption data of the server during all historical liquid cooling cycles as the actual energy consumption data sequence of the corresponding server, and obtain the energy consumption data points corresponding to the server according to the mean value of the actual energy consumption data sequence and the number of data greater than the preset energy consumption threshold in the actual energy consumption data sequence. Cluster the energy consumption data points according to the distance between the energy consumption data points to obtain initial clustering clusters; A second partitioning module, which is used to reallocate the energy consumption data points in the initial clustering clusters according to the total number of energy consumption data points in the initial clustering clusters and the number of servers covered by the CDU in the liquid cooling system to obtain each target clustering cluster; A liquid cooling system management module, which is used to reallocate and manage the servers controlled by the CDU of the liquid cooling system according to the servers corresponding to the energy consumption data points in the target clustering clusters; The method of reallocating the energy consumption data points in the initial clustering clusters to obtain each target clustering cluster includes: Taking the number of servers covered by a single CDU as the quantity judgment threshold; if the total number of energy consumption data points in the initial clustering cluster is greater than the quantity judgment threshold, the corresponding initial clustering cluster is recorded as an excess clustering cluster; if the total number of energy consumption data points in the initial clustering cluster is less than the quantity judgment threshold, the corresponding initial clustering cluster is recorded as a lack clustering cluster; if the total number of energy consumption data points in the initial clustering cluster is equal to the quantity judgment threshold, the corresponding initial clustering cluster is recorded as a target clustering cluster; According to the distance between each energy consumption data point in the excess clustering cluster and the clustering center point of the lack clustering cluster and the difference between adjacent actual energy consumption data in the actual energy consumption data sequence of the server corresponding to each energy consumption data point in the excess clustering cluster, obtain the degree of allocation to be determined corresponding to each energy consumption data point in the excess clustering cluster; according to the degree of allocation to be determined corresponding to each energy consumption data point in the excess clustering cluster and the quantity judgment threshold, obtain the sequence of data points to be allocated corresponding to each excess clustering cluster; according to the sequence of data points to be allocated corresponding to the excess clustering cluster, reallocate the excess clustering cluster and the lack clustering cluster to obtain the allocated clustering cluster corresponding to each excess clustering cluster and the received clustering cluster corresponding to each lack clustering cluster, and record all the allocated clustering clusters and all the received clustering clusters as target clustering clusters. The data points in the sequence of data points to be allocated are data points to be allocated.

2. The energy consumption data management platform of the liquid cooling system based on artificial intelligence according to claim 1, characterized in that The method for acquiring the actual energy consumption data of each server during each historical liquid cooling cycle includes: For any server: Obtain the temperature to be reduced during the a-th historical liquid cooling cycle of the server, and denote it as the temperature change amount. Obtain the reciprocal of the refrigeration energy efficiency ratio of the liquid cooling system, and denote it as the characteristic ratio. Denote the result of multiplying the cold liquid flow rate, cold liquid density, cold liquid specific heat capacity, the temperature change amount, the overall cycle time of the server during the a-th historical liquid cooling cycle, and the characteristic ratio in the circulation branch during the a-th historical liquid cooling cycle of the server as the actual energy consumption data of the server during the a-th historical liquid cooling cycle.

3. The energy consumption data management platform of the liquid cooling system based on artificial intelligence according to claim 1, characterized in that, The method for obtaining the energy consumption data points corresponding to the server includes: Denote the mean value of the actual energy consumption data sequence of the server as the historical actual energy consumption data of the corresponding server, and denote the total number of actual energy consumption data greater than the preset energy consumption threshold in the actual energy consumption data sequence of the server as the actual cooling frequency of the corresponding server. Construct a two-dimensional mapping space, where the abscissa of the two-dimensional mapping space is the actual cooling frequency and the ordinate axis is the historical actual energy consumption data. Map the historical actual energy consumption data and the actual energy consumption frequency of the server into the two-dimensional mapping space to obtain the energy consumption data points corresponding to the corresponding server.

4. The energy consumption data management platform of the liquid cooling system based on artificial intelligence according to claim 3, characterized in that, The method for obtaining the initial clustering clusters includes: Denote the Euclidean distance between any two energy consumption data points in the two-dimensional mapping space as the metric distance between the corresponding two energy consumption data points. According to the Euclidean distance between any two energy consumption data points in the two-dimensional mapping space, use a clustering algorithm to cluster all the energy consumption data points in the two-dimensional mapping space, and denote all the clustering clusters obtained by clustering as the initial clustering clusters.

5. The energy consumption data management platform of the liquid cooling system based on artificial intelligence according to claim 1, characterized in that The method for obtaining the degree of assignment corresponding to each energy consumption data point in the surplus clustering cluster includes: For the g-th energy consumption data point in any surplus clustering cluster G: Denote the actual energy consumption data sequence of the server corresponding to the g-th energy consumption data point as the sequence to be analyzed. Obtain the difference sequence of the sequence to be analyzed, and denote the sum of all differences in the difference sequence as the comprehensive difference. Denote the negative correlation mapping value of the comprehensive difference as the credibility of the g-th energy consumption data point. The h-th difference in the difference sequence is the absolute value of the difference between the h-th data and the (h + 1)-th data in the sequence to be analyzed. Obtain the cluster center point of the lack clustering cluster closest to the surplus clustering cluster G, and denote it as the nearest neighbor cluster center point of the surplus clustering cluster G. Denote the Euclidean distance between the g-th energy consumption data point and the nearest neighbor cluster center point as the nearest neighbor cluster distance of the g-th energy consumption data point. Denote the result of multiplying the reciprocal of the credibility by the reciprocal of the nearest neighbor cluster distance as the degree of assignment corresponding to the g-th energy consumption data point in the surplus clustering cluster G.

6. The energy consumption data management platform of the liquid cooling system based on artificial intelligence according to claim 1, characterized in that The method for obtaining the sequence of data points to be assigned corresponding to the surplus clustering cluster includes: For any surplus clustering cluster: Sort the energy consumption data points in the surplus clustering cluster in descending order according to the degree of pending allocation to obtain the energy consumption data point sequence corresponding to the surplus clustering cluster; Denote the result of subtracting the quantity judgment threshold from the total number of energy consumption data points in the surplus clustering cluster as the redundant data point quantity value corresponding to the surplus clustering cluster; In the energy consumption data point sequence corresponding to the surplus clustering cluster, Denote the sequence composed of the first M energy consumption data points as the pending allocation data point sequence corresponding to the surplus clustering cluster, where M is the redundant data point quantity value of the surplus clustering cluster.

7. The energy consumption data management platform of the liquid cooling system based on artificial intelligence according to claim 1, characterized in that, The method for obtaining the allocated clustering cluster corresponding to the surplus clustering cluster and the received clustering cluster corresponding to the lack clustering cluster includes: Obtain the surplus clustering cluster sequence and the lack clustering cluster sequence corresponding to the surplus clustering cluster. Among the lack clustering clusters corresponding to the surplus clustering cluster, the closer the lack clustering cluster is to the corresponding surplus clustering cluster, the more forward it is ranked. The surplus clustering cluster sequence is arranged according to the number of pending allocation data points within the cluster; Obtain the starting reception quantity of the lack clustering cluster. The starting reception quantity of the lack clustering cluster is the result of subtracting the total number of original data points in the corresponding lack clustering cluster from the quantity judgment threshold. The original data points in the lack clustering cluster refer to the energy consumption data points divided into the lack clustering cluster through clustering. Determine whether the total number of data points to be assigned in the first surplus cluster in the surplus cluster sequence is less than or equal to the starting reception quantity of the first lack cluster in the lack cluster sequence corresponding to the first surplus cluster. If so, allocate all the data points to be assigned in the first surplus cluster to the first surplus cluster after the first lack cluster in the lack cluster sequence corresponding to the first surplus cluster, which is denoted as the assigned cluster corresponding to the first surplus cluster, and update the starting reception quantity of the lack cluster to obtain the first updated reception quantity of the lack cluster. Then continue to determine whether the total number of data points to be assigned in the second surplus cluster in the surplus cluster sequence is less than or equal to the first updated reception quantity of the first lack cluster in the lack cluster sequence corresponding to the second surplus cluster. If not, determine whether the sum of the first updated reception quantity of the first lack cluster in the lack cluster sequence corresponding to the second surplus cluster and the first updated reception quantity of the second lack cluster in the lack cluster sequence corresponding to the second surplus cluster is greater than or equal to the total number of data points to be assigned in the second surplus cluster. If so, allocate the first V1 data points to be assigned in the sequence of data points to be assigned corresponding to the second surplus cluster to the first lack cluster in the lack cluster sequence corresponding to the second surplus cluster, allocate all the data points to be assigned after the V1 -th data point to be assigned in the sequence of data points to be assigned corresponding to the second surplus cluster to the second lack cluster in the lack cluster sequence corresponding to the second surplus cluster, and allocate all the data points to be assigned in the second surplus cluster to the second surplus cluster after the first lack cluster and the second lack cluster in the lack cluster sequence corresponding to the second surplus cluster, which is denoted as the assigned cluster corresponding to the second surplus cluster, and so on, until the allocation is determined to be completed after obtaining the assigned cluster corresponding to the last surplus cluster in the surplus cluster sequence; V1 is the first updated reception quantity of the first lack cluster in the lack cluster sequence corresponding to the second surplus cluster. Denote the cluster formed by all the data points to be assigned received by the lack cluster during the allocation process and all the original data points in the corresponding lack cluster as the received cluster corresponding to the corresponding lack cluster.

8. The energy consumption data management platform of the liquid cooling system based on artificial intelligence according to claim 7, characterized in that, The method for updating the starting reception quantity of the lack cluster to obtain the first updated reception quantity of the lack cluster includes: For any lack of clustering cluster, if the lack of clustering cluster receives the data points to be allocated assigned by the first surplus clustering cluster, then the result of subtracting the number of data points to be allocated belonging to the first surplus clustering cluster received by the lack of clustering cluster from the starting reception quantity of the lack of clustering cluster is recorded as the first updated reception quantity of the lack of clustering cluster. If the lack of clustering cluster has not received the data points to be allocated belonging to the first surplus clustering cluster, then the starting reception quantity of the lack of clustering cluster is used as the first updated reception quantity of the lack of clustering cluster.

9. The energy consumption data management platform of the liquid cooling system based on artificial intelligence according to claim 1, characterized in that, A method for reallocating and managing the servers controlled by the CDU of the liquid cooling system according to the servers corresponding to the energy consumption data points in the target clustering cluster includes: For any target clustering cluster, all the servers corresponding to the energy consumption data points in the target clustering cluster are connected to the CDU of the same liquid cooling system.

Images