Data center precision system energy-saving optimization method based on data fusion

Abstract

The invention discloses a data center precision system energy-saving optimization method based on data fusion. The method comprises the following steps: acquiring historical operation data corresponding to a target system at a historical operation moment; the historical operation data comprises system operation parameters corresponding to the target system; constructing an operation relation graph corresponding to the target system based on each system operation parameter; performing feature information extraction on the operation relation graph to obtain operation feature information corresponding to the historical operation data; obtaining a target control model, and obtaining current operation data corresponding to the current operation moment of the target system; the target control model is obtained by training based on the operation feature information and a parameter control relationship corresponding to the target system; and inputting the current operation data into the target control model to obtain operation state control information corresponding to the target system. The operation state of the target system is controlled and adjusted in combination with the historical operation data, the accuracy of regulation and control of the target system is improved, and energy consumption is optimized.

Description

Technical Field

[0001] The present invention belongs to the field of computer technology, and more specifically, relates to a data center precision system energy-saving optimization method based on data fusion. Background Art

[0002] With the development of information technology, more and more large-scale integrated equipment is applied to all walks of life and is inextricably linked to social life. Since large-scale integrated equipment has a wide range of application scenarios, such as smart cities, industrial Internet of Things and other scenarios, most of them require continuous and uninterrupted execution of work. At this time, how to set up and control the relevant refrigeration system so that it can effectively save operating energy consumption while ensuring the heat dissipation of large-scale integrated equipment has become a problem that needs to be explored.

[0003] In the prior art, the control of the refrigeration system cannot be carried out in real time and intelligently according to its specific operating status, so there is still room for improvement in the control accuracy of the system; correspondingly, due to the low control accuracy, there is still a large amount of redundant energy consumption. Summary of the invention

[0004] In order to solve the deficiencies in the prior art, the purpose of the present invention is to solve the above-mentioned defects and further propose an energy-saving optimization method for a data center precision system based on data fusion.

[0005] The present invention adopts the following technical solution: The first aspect of the present invention discloses a data center precision system energy-saving optimization method based on data fusion, the method comprising: Acquire historical operation data corresponding to the target system at the historical operation time; the historical operation data includes various system operation parameters corresponding to the target system; Constructing an operation relationship diagram corresponding to the target system based on the operation parameters of each system; Extracting characteristic information from the operation relationship diagram to obtain operation characteristic information corresponding to the historical operation data; Acquire a target control model, and acquire current operation data corresponding to the target system at the current operation time; the target control model is trained based on the operation characteristic information and the parameter control relationship corresponding to the target system; The current operation data is input into the target control model to obtain the operation status control information corresponding to the target system.

[0006] In some possible embodiments, constructing an operation relationship diagram corresponding to the target system based on the operation parameters of each system includes: Determine a corresponding node in the operation relationship diagram based on each operation parameter; Acquire the connection relationship between the components in the target system and the parameter control relationship corresponding to the target system; Based on the connection relationship, determine the node type corresponding to each node; Based on the various nodes, the connection relationships, the node types corresponding to the various nodes, and the parameter control relationships corresponding to the target system, an operation relationship diagram corresponding to the target system is constructed.

[0007] In some possible embodiments, constructing an operation relationship graph corresponding to the target system based on the nodes, the connection relationships, the node types corresponding to the nodes, and the parameter change relationships corresponding to the target system includes: Based on the nodes and the connection relationships, determine the edge direction corresponding to each edge in the operation relationship graph; Based on the parameter control relationship between each node and the target system, determining an edge capacity value corresponding to each edge; the edge capacity value represents an energy transfer threshold between components in the target system; An operation relationship graph corresponding to the target system is constructed based on the nodes, the node types corresponding to the nodes, the edge directions corresponding to the edges, and the edge capacity values.

[0008] In some possible embodiments, extracting characteristic information from the operation relationship diagram to obtain operation characteristic information corresponding to the historical operation data includes: Obtaining the number of traversals of the running relationship graph; Based on the number of traversals, respectively determine the height of the source node, the height of the sink node, and the height of the intermediate node; Calculating the edge flow value corresponding to the edge related to the source node in the operation relationship graph; the edge related to the source node is an edge formed between the source node and an adjacent node of the source node; Constructing a first matrix based on the height of the source node, the height of the sink node, and the height of the intermediate node; Constructing a second matrix based on the edge flow values ​​corresponding to the edges related to each source node; The first matrix and the second matrix are determined as the operation characteristic information.

[0009] In some possible embodiments, determining the height of the source node, the height of the sink node, and the height of the intermediate node based on the number of traversals respectively includes: Determine the number of nodes in the operation relationship graph as the height of the source node; Determine that the height of the sink node is 0; If the traversal count is 1, the height of the intermediate node is determined to be 0, and the traversal count is updated; If the traversal times is greater than 1, the height of the intermediate node is determined based on the node height in the previous traversal process, and the traversal times is updated.

[0010] In some possible embodiments, the method further includes: If the traversal times is greater than 1, performing path detection on each edge in the running state graph to obtain a path detection result; If the path detection result indicates that there is at least one edge in the operating status graph whose starting node is the source node and whose ending node is the sink node, the height of the intermediate node is updated.

[0011] In some possible embodiments, the method further includes: Acquire a graph neural network and sample training data; the sample training data includes historical operation data corresponding to the target system at multiple historical operation moments, and a parameter control relationship corresponding to the target system; The sample training data is input into the graph neural network so that the graph neural network performs model training based on the sample training data and obtains the target control model corresponding to the target system.

[0012] The second aspect of the present invention discloses a data center precision system energy-saving optimization device based on data fusion, the device comprising: A first acquisition module is used to acquire historical operation data corresponding to the target system at the historical operation time; the historical operation data includes various system operation parameters corresponding to the target system; A directed graph construction module, used to construct an operation relationship graph corresponding to the target system based on the operation parameters of each system; An information extraction module, used to extract characteristic information from the operation relationship diagram to obtain operation characteristic information corresponding to the historical operation data; A second acquisition module is used to acquire a target control model and current operation data corresponding to the target system at the current operation time; the target control model is trained based on the operation characteristic information and the parameter control relationship corresponding to the target system; The information output module is used to input the current operation data into the target control model to obtain the operation status control information corresponding to the target system.

[0013] The third aspect of the present invention discloses a device, which includes a processor and a memory, wherein the memory stores at least one instruction and at least one program, and the at least one instruction and the at least one program are loaded and executed by the processor to implement the data fusion-based data center precision system energy-saving optimization method as described above.

[0014] The fourth aspect of the present invention discloses a computer storage medium, in which at least one instruction and at least one program are stored. The at least one instruction and the at least one program are loaded and executed by a processor to implement the energy-saving optimization method for a precision system of a data center based on data fusion as described above.

[0015] Compared with the prior art, the present invention has the following advantages: By obtaining the historical operation data corresponding to the target system at the historical operation time, and constructing the operation relationship diagram corresponding to the historical operation data, and then obtaining the operation characteristic information corresponding to the historical operation data according to the operation relationship diagram, and inputting the operation characteristic information into the target control model, combined with the current operation data, the operation status control information corresponding to the target system is obtained. Combined with the real operation data of the target system at the historical operation time, the operation status information of the target system is determined, and the operation status of the target system is intelligently adjusted and controlled according to the real operation data, which can improve the accuracy of the operation of the target system, optimize the operation status of the target system, and reduce unnecessary energy waste. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] Figure 1 It is a flow chart corresponding to the energy-saving optimization method for a data center precision system based on data fusion provided by an embodiment of the present invention; Figure 2 It is a first process schematic diagram corresponding to the construction of the operation relationship diagram provided in an embodiment of the present invention; Figure 3 It is a second process diagram corresponding to the construction of the operation relationship diagram provided in the embodiment of the present invention; Figure 4 It is a schematic diagram of a corresponding process for determining operation characteristic information provided by an embodiment of the present invention; Figure 5 It is a schematic diagram of a corresponding process of determining a node height provided by an embodiment of the present invention; Figure 6 It is a schematic diagram of the process flow corresponding to the node height update provided by an embodiment of the present invention; Figure 7 is a schematic diagram of a process flow corresponding to target model training provided by an embodiment of the present invention; Figure 8 It is a structural schematic diagram corresponding to the system control device provided in an embodiment of the present invention. DETAILED DESCRIPTION

[0017] The following will be combined with the drawings in the embodiments of this specification to clearly and completely describe the technical solutions in the embodiments of this specification. Obviously, the described embodiments are only part of the embodiments of this specification, not all of the embodiments. Based on the embodiments in this specification, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0018] It should be noted that the terms "first", "second", etc. in the specification and claims of the present invention and the above-mentioned drawings are used to distinguish similar objects, and are not necessarily used to describe a specific order or sequence. It should be understood that the data used in this way can be interchangeable where appropriate, so that the embodiments of the present invention described herein can be implemented in an order other than those illustrated or described herein. In addition, the terms "including" and "having" and any variations thereof are intended to cover non-exclusive inclusions, for example, a process, method, system, product, or server that includes a series of steps or units is not necessarily limited to those steps or units clearly listed, but may include other steps or units that are not clearly listed or inherent to these processes, methods, products, or devices.

[0019] Various exemplary embodiments, features and aspects of the present disclosure will be described in detail below with reference to the accompanying drawings. The same reference numerals in the accompanying drawings represent elements with the same or similar functions. Although various aspects of the embodiments are shown in the accompanying drawings, the drawings are not necessarily drawn to scale unless otherwise specified.

[0020] The word “exemplary” is used exclusively herein to mean “serving as an example, example, or illustration.” Any embodiment described herein as “exemplary” is not necessarily to be construed as preferred or advantageous over other embodiments.

[0021] The term "and / or" herein is only a description of the association relationship of the associated objects, indicating that there may be three relationships. For example, A and / or B can represent: A exists alone, A and B exist at the same time, and B exists alone. In addition, the term "at least one" herein represents any combination of at least two of any one or more of a plurality of. For example, including at least one of A, B, and C can represent including any one or more elements selected from the set consisting of A, B, and C.

[0022] In addition, in order to better illustrate the present disclosure, numerous specific details are given in the following specific embodiments. It should be understood by those skilled in the art that the present disclosure can also be implemented without certain specific details. In some examples, methods, means, components and circuits well known to those skilled in the art are not described in detail in order to highlight the subject matter of the present disclosure.

[0023] Figure 1 This is a flow chart corresponding to the energy-saving optimization method for a data center precision system based on data fusion provided by an embodiment of the present invention. The execution subject may be any terminal or electronic device, such as a computer, that can implement the energy-saving optimization method for a data center precision system based on data fusion; please refer to Figure 1 In one embodiment, a data center precision system energy saving optimization method based on data fusion includes the following steps: Step S101: Acquire historical operation data corresponding to the target system at the historical operation time; the historical operation data includes various system operation parameters corresponding to the target system; The target system can be a precision system in a data center, and the application scenario can be any instrument or equipment with cooling requirements, or any other scenario that requires the application of a precision system in a data center; a precision system in a data center can include the following subcomponents: a control system, a chilled water system, a cooling tower, an air distribution system, and backup equipment, etc.; the chilled water system is used to provide a cooling source, and may include chillers, chilled water pumps, pipes, etc.; the air distribution system can be used to achieve functions such as underfloor air supply and hot and cold channel closure; backup equipment is used to ensure the normal operation of the system in the event of equipment failure.

[0024] A running time can be set as a fixed system running time, and the historical running time can be any running time before the current running time, or any multiple running times; the data generated or used by the target system at the historical running time is the historical running data, and the historical running time can include various system operating parameters corresponding to the target system. The system operating parameters can be obtained through various sensors, and can include at least one or more of the following: temperature, humidity, liquid flow, etc., and specifically can also be data such as the flow of chilled water, air temperature, and the rate of change of air humidity. The operating status of the target system at the corresponding historical running time can be determined by the changes in the parameters in the two historical running data.

[0025] Step S102: constructing an operation relationship diagram corresponding to the target system based on the operation parameters of each system; The energy consumption of the target system's operating state can be optimized by constructing an operating relationship diagram. The operating relationship diagram can characterize the energy transfer direction between various components in the data center precision system, as well as the energy transfer threshold. Therefore, the energy change of the target system can be intuitively reflected based on the operating relationship diagram. By analyzing the operating relationship diagram, the intuitiveness and convenience of the target system regulation can be improved. In a specific embodiment, Figure 2 As shown, the construction of the running relationship graph may include: Step S201: determining a corresponding node in the operation relationship diagram based on each operation parameter; Each system operation parameter included in the historical operation data corresponding to each historical operation time can be determined as a node in the operation relationship graph. Since the operation relationship graph is a directed graph, the direction of the edge formed by two adjacent nodes also needs to be determined.

[0026] Step S202: obtaining the connection relationship between the components in the target system and the parameter control relationship corresponding to the target system; The parameter control relationship may include any two structurally connected sub-components in the target system. The parameter control relationship is used to feedback the coupling relationship between the two sub-components. Therefore, it can be known that different parameter control relationships exist between different components, for example: The parameter control relationship between the chilled water system and the air distribution system is shown in the following formula:

[0027]

[0028]

[0029] in, Pipeline The heat transfer coefficient and pipe length, They represent the flow rate and flow rate change rate of chilled water respectively. is the specific heat capacity of water, Respectively represent the air temperature and temperature change rate, They represent the temperature of the chilled water in the chilled water system and the rate of temperature change, For pipeline The rate of temperature change, is the heat transfer coefficient of the heat exchanger, is the change rate of heat exchanger fouling thermal resistance, is the heat transfer area of ​​the heat exchanger, is the temperature of the chilled water in the chilled water system, is the heat change rate of the heat exchanger, They represent the air flow rate and flow rate change rate respectively. The rate of change of air humidity, It represents the influence coefficient of air humidity on specific heat capacity, is the rate of temperature change of the air distribution system.

[0030] The parameter control relationship between the cooling tower and the chiller is shown in the following formula:

[0031]

[0032]

[0033] in, Pipeline The heat transfer coefficient and pipe length, They represent the flow rate and flow rate change rate of cooling water in the condenser of the cooling tower, Respectively represent the cooling water temperature and temperature change rate, For pipeline The rate of temperature change, is the temperature change rate of the condenser, are the heat transfer coefficient and heat transfer coefficient change rate of the condenser, are the initial heat load and heat load change rate of the condenser, is the rate of change of cooling tower efficiency, The influence coefficient of reaction temperature on cooling tower efficiency, is the temperature change rate of the cooling water in the condenser of the cooling tower.

[0034] The parameter control relationship between the air distribution system and the backup equipment is shown in the following formula:

[0035] in, are the heat load and heat load change rate of the standby equipment, is the temperature change rate of the standby equipment, are the initial efficiencies of the air distribution system and backup equipment, respectively.

[0036] The above parameter control relationships are described as examples in the embodiments of the present invention, and the target system is not limited to only include the above three parameter control relationships. The parameter control relationships can be inferred according to changes in various parameters of the target system during actual application.

[0037] Step S203: determining the node type corresponding to each node based on the connection relationship; Node types may include source nodes, sink nodes, and intermediate nodes; among all nodes included in the running relationship graph, the source node is a node with an in-degree of 0, and the sink node is a node with an out-degree of 0; among all nodes in the running relationship graph, the remaining nodes except the source node and the sink node are intermediate nodes.

[0038] The operation relationship diagram can be a directed graph model. In the constructed directed graph model, it can be determined whether the node belongs to a source point or a sink point according to the energy propagation direction of the precision system of the data center. It can be understood that a directed graph model can include multiple source points, for example, usually the data in the chilled water system is used as the source point node in the node; a directed graph model can include multiple sink points, for example, the data in the air distribution system is used as the sink node in the node, and more intermediate nodes, for example, the data in the cooling tower and the chiller is used as the intermediate node in the node.

[0039] Step S204: constructing an operation relationship diagram corresponding to the target system based on the nodes, the connection relationships, the node types corresponding to the nodes, and the parameter control relationships corresponding to the target system.

[0040] Furthermore, if Figure 3 As shown, step S204 may include: Step S301: determining the edge direction corresponding to each edge in the operation relationship graph based on the nodes and the connection relationship; Since the direction of each node and edge in the operation relationship diagram is needed to characterize the direction of energy propagation between various components in the data center precision system, the direction corresponding to each edge can be determined based on the connection relationship between the nodes and components.

[0041] Step S302: determining an edge capacity value corresponding to each edge based on the parameter control relationship corresponding to each node and the target system; the edge capacity value represents an energy transfer threshold between components in the target system; In a specific embodiment, since the parameter control relationship needs to comply with the law of conservation of energy, after determining the direction of the nodes and edges, the edge capacity value of each edge needs to be calculated. Since the edge is formed based on two nodes, and the nodes can be considered to correspond to sub-components in the target system, in order to make the energy transfer of the target system during operation comply with requirements such as safe operation and the law of conservation of energy, the edge capacity value of each edge needs to be calculated to determine the energy transfer threshold between sub-components.

[0042] Step S303: constructing an operation relationship graph corresponding to the target system based on the nodes, the node types corresponding to the nodes, the edge directions corresponding to the edges, and the edge capacity values.

[0043] The node type is used to determine the position of the node in the operation relationship graph; the operation relationship graph corresponding to the target system is formed by combining the node, node type, edge direction and edge capacity value.

[0044] Next, we will continue based on Figure 1This paper expounds the energy-saving optimization method of data center precision system based on data fusion.

[0045] Step S103: extracting characteristic information from the operation relationship diagram to obtain operation characteristic information corresponding to the historical operation data; The purpose of the present invention is to optimize the energy consumption of the target system during operation by adjusting and controlling the target system. Therefore, it is necessary to convert the work of the compressor in the chilled water system into the heat energy output by the air distribution system as much as possible; then, based on the idea of ​​relabeling, the characteristic information of the historical operation data can be extracted.

[0046] like Figure 4 As shown, step S103 may include: Step S401: obtaining the number of traversal of the operation relationship graph; In order to realize intelligent control of the operating status of the target system and ensure the accuracy of the target system regulation, after each traversal of the operation relationship graph, there may be a need to update the node height to update the operation relationship graph. Therefore, before determining the node height, it is necessary to first determine the number of traversals of the operation relationship graph.

[0047] Step S402: Based on the traversal times, respectively determine the height of the source node, the height of the sink node, and the height of the intermediate node; After the operation relationship graph, i.e., the directed graph, is constructed, feature information of the historical operation data can be extracted based on the relabeling algorithm and the operation relationship graph; specifically, Figure 5 As shown, step S402 may include: Step S501: Determine the number of nodes in the operation relationship graph as the height of the source node; Since the source node is a node with an in-degree of 0, its node height in the running relationship graph can be set to the number of all nodes in the running relationship graph.

[0048] Step S502: Determine that the height of the sink node is 0; Since the sink node is a node with an out-degree of 0, its node height in the running relationship graph can be set to 0.

[0049] Step S503: If the traversal times is 1, determine that the height of the intermediate node is 0, and update the traversal times; It can be considered that the node height of the intermediate node in the initial state is 0. Since the node height of the intermediate node may be updated with the number of traversals, the number of traversals is updated each time after the node height of the intermediate nodes included in the running relationship graph is determined, that is, the number of traversals = the number of traversals + 1.

[0050] Step S504: If the traversal times is greater than 1, the height of the intermediate node is determined based on the node height in the previous traversal process, and the traversal times is updated.

[0051] If the number of traversals is greater than 1, it means that before this traversal of the running relationship graph, there was a historical running time that also traversed the running relationship graph; therefore, the node height of the current intermediate node can be determined based on the intermediate node height corresponding to the previous historical running time, and the number of traversals can be updated.

[0052] In addition, the node height of the intermediate node may be updated as the number of traversals increases; Figure 6 As shown, the node height updating method of the intermediate node also includes: Step S601: If the traversal times is greater than 1, path detection is performed on each edge in the running state graph to obtain a path detection result; When the number of traversals is greater than 1, based on the single-source shortest path algorithm, determine whether there is a path from any source point to any sink point in the remaining network; the remaining network can be a graph consisting of the remaining nodes and edges after the adjacent nodes of the source node are screened out in the operating relationship graph; the edge flow of each edge in the remaining network is equal to the edge capacity in the operating relationship graph minus the edge flow, and the ratio of the energy transmitted by the subcomponent corresponding to each node in the operating relationship graph under the actual operating state to the edge capacity can be determined as the edge flow.

[0053] Step S602: If the path detection result indicates that there is at least one edge in the running state graph whose starting node is the source node and whose ending node is the sink node, the height of the intermediate node is updated, wherein the node height of the updated intermediate node may be 0.

[0054] In a specific embodiment, extracting feature information from historical operation data based on the operation relationship diagram and label re-sticking idea may include the following steps: Step A1: Set the height of all source nodes to n and the height of all sink nodes to 0. n is the number of nodes in the operation relationship graph; if it is the first execution, set the height of all other nodes to 0; otherwise, determine the height of the remaining nodes based on the node height of the previous historical operation data.

[0055] Step A12: For each source node, set the flow of the edge from the source point to the first node as the capacity of the edge; wherein the first node is any adjacent point of the source point in the operation relationship graph; Step A13: If it is the first execution, jump to step A14; otherwise, jump to step A15.

[0056] Step A14: For any node Execute the following PUSH operations or RELABEL operations until they cannot be executed, ending this round of steps.

[0057] The specific PUSH operation is: if the node overflows, then find the node Neighboring points , and the height of the adjacent point is equal to the height of the node minus 1, then the flow between the node and the adjacent point is modified to , as shown below:

[0058] in, Representation Node To Node The capacity of the edge, The node before the PUSH operation is performed To Node The flow of the edge, For Node of overflow flow.

[0059] The RELABEL operation is as follows: if the node Overflow, and all its adjacent points If the heights of are greater than or equal to the node, the height of the node is modified as shown in the following formula:

[0060] in, and Respectively represent nodes The height and node height.

[0061] If the flow rate input to the node is greater than the flow rate output by the node, the node is determined to be overflowed.

[0062] Step A15: Otherwise, only perform the PUSH operation on all nodes u until it cannot be executed. Based on the single-source shortest path algorithm, determine whether there is a path from any source point to any sink point in the remaining network. If not, end this round of steps; otherwise, reset the height of all other nodes to 0 and jump to step A14.

[0063] It can be understood that step A11 to step A15 are continuously iterated, that is, step A11 to step A15 are executed for each piece of historical operation data. What is different from the traditional idea of ​​label re-sticking is that the graph network of traditional label re-sticking itself is relatively complex, and its time performance is O(M*V*V*E), where M, V and E are the number of samples, the number of nodes and the number of edges respectively; however, the idea of ​​label overlap can be borrowed to improve the above time complexity. In the present invention, its corresponding time performance is O(W*V*V*E+M*E), where W is much smaller than M, which can be understood as the type of sample (that is, using different height information to divide the type of sample).

[0064] Step S403: Calculate the edge flow value corresponding to the edge related to the source node in the operation relationship graph; the edge related to the source node is the edge formed between the source node and the adjacent node of the source node; The edge related to the source node can be an edge formed by the source node and any adjacent node adjacent to the source node; the edge flow value can be the ratio of the energy transmitted by the subcomponent corresponding to each node in the operation relationship diagram under the actual operation state to the edge capacity.

[0065] Step S404: constructing a first matrix based on the height of the source node, the height of the sink node and the height of the intermediate node; Step S405: constructing a second matrix based on the edge flow values ​​corresponding to the edges related to each source node; The first matrix can be a transfer matrix, and the second matrix can be an adjacency matrix. In the adjacency matrix P, the element p_uv=f(u,v) in the uth row and vth column. In the transfer matrix H, the element h_uv=uh-vh in the uth row and vth column. The node height and edge flow value of each node can be input into the graph neural network, and the first matrix and the second matrix can be calculated based on the graph neural network.

[0066] Specifically, the generation process of the transfer matrix depends on the height information of the nodes in the directed graph. To determine the transfer relationship between nodes through the height information, we can first create a zero matrix. In the process of generating the transfer matrix, we traverse each node of the directed graph and determine its successor node based on the height information of the node. For each successor node, we set the corresponding element of the transfer matrix to 1, indicating that there is a transfer relationship between the nodes. Then, we recursively find the successor node of the successor node and set the corresponding element of the transfer matrix to 1, and so on.

[0067] Edge flow values, also known as edge weight values, play a key role in constructing the adjacency matrix. The adjacency matrix is ​​used to represent the connection relationship between nodes in a directed graph, while the edge weight information provides the weight of the edge connecting the nodes. By mapping the edge weight information to the corresponding position of the adjacency matrix, we can accurately represent the connection relationship and weight between nodes in the adjacency matrix.

[0068] Step S406: Determine that the first matrix and the second matrix are the operation characteristic information.

[0069] The first matrix and the second matrix obtained according to the operation relationship diagram are the operation characteristic information.

[0070] Step S104: obtaining a target control model and obtaining current operation data corresponding to the target system at the current operation time; the target control model is trained based on the operation characteristic information and the parameter control relationship corresponding to the target system; Using the target control model to process the current operating data to generate operating status control information that matches the target system can improve the accuracy of target system adjustment and control, thereby ensuring good energy consumption optimization effects.

[0071] In addition, if Figure 7 As shown, the method may also include: Step S701: Acquire a graph neural network and sample training data; the sample training data includes historical operation data corresponding to the target system at multiple historical operation moments, and parameter control relationships corresponding to the target system, and may also include operation feature information at multiple historical operation moments; Step S702: Input the sample training data into the graph neural network so that the graph neural network performs model training based on the sample training data and obtains the target control model corresponding to the target system.

[0072] The graph neural network can be trained based on the feature information of historical operation data and the current control strategy. For example, the transfer matrix and adjacency matrix in the feature information are used as the input layer and input into the graph neural network. In each layer of the graph neural network, the model updates the feature representation of the node according to the transfer matrix and the adjacency matrix. In the last layer of the graph neural network, the final feature representation of the node can be obtained, that is, the feature matrix. Among them, the transfer matrix can be used as part of the node feature, and the adjacency matrix is ​​used to describe the connection relationship between nodes.

[0073] For graph neural networks, graph neural networks can be a The convolutional model of the layer, is the number of graph convolutional layers, if The larger the value of , the better the training effect and the longer the iteration time.

[0074] The expression for calculating the characteristic matrix is: , .

[0075] In the formula, Indicates The transpose of the sample data, , is the number of sample data. For the The feature map matrix of the layer, , Indicates The weight matrix of the layer is obtained through graph neural network training. As model parameters of graph neural networks.

[0076] In order to realize the intelligent output of the operating status control information corresponding to the target system and ensure the accuracy of the control information generation, the graph neural network can be trained using the real operating data related to the target system to obtain the corresponding target control model.

[0077] In addition, in practical applications, the energy-saving optimization process of the target system must comprehensively consider a series of practical factors, including the specifications and performance of the hardware equipment, the degree of aging of the equipment, the failure rate, and the operating status. In addition, external environmental factors such as temperature, humidity, wind speed, etc., which affect system performance, must also be considered. These variables may affect the prediction accuracy and optimization effect of the model. Therefore, in order to more comprehensively evaluate and optimize the energy efficiency of the system, the parameters corresponding to the above factors can be combined for analysis to build a more accurate prediction model, which can not only enhance the adaptability of the model to complex real-world scenarios, but also achieve more efficient resource allocation and fault warning, and maximize the energy-saving optimization effect.

[0078] Step S105: input the current operation data into the target control model to obtain the operation status control information corresponding to the target system.

[0079] The operation status control information is used to adjust and control the operation status of the target system; the operation status information can be an energy-saving optimization strategy, specifically the operation data corresponding to the refrigeration room system after energy-saving optimization, including the temperature, flow, specific heat capacity, heat transfer coefficient and other physical quantity data involved in the parameter control relationship, and also includes the specific optimization strategy for each subsystem in the confluence node, such as adjusting the temperature of the chilled water, optimizing the cooling water flow, adjusting the wind temperature of the air distribution system, etc. Intelligently adjusting and controlling the operation status of the target system based on the real operation data can improve the accuracy of the operation of the target system, optimize the operation status of the target system, and reduce unnecessary energy consumption.

[0080] The data center precision system energy-saving optimization device based on data fusion provided by the present invention is described below. The system control device described below and the data center precision system energy-saving optimization method based on data fusion described above can refer to each other.

[0081] like Figure 8 As shown, a data center precision system energy-saving optimization device based on data fusion includes: The first acquisition module 810 is used to acquire the historical operation data corresponding to the target system at the historical operation time; the historical operation data includes various system operation parameters corresponding to the target system; A directed graph construction module 820, configured to construct an operation relationship graph corresponding to the target system based on the operation parameters of each system; An information extraction module 830 is used to extract characteristic information from the operation relationship diagram to obtain operation characteristic information corresponding to the historical operation data; The second acquisition module 840 is used to acquire a target control model and current operation data corresponding to the target system at the current operation time; the target control model is trained based on the operation characteristic information and the parameter control relationship corresponding to the target system; The information output module 850 is used to input the current operation data into the target control model to obtain the operation status control information corresponding to the target system.

[0082] In some other embodiments, the directed graph construction module 820 includes: A node determination module, used to determine a corresponding node in the operation relationship diagram based on each operation parameter; A relationship acquisition module, used to acquire the connection relationship between the components in the target system and the parameter control relationship corresponding to the target system; A node type determination module, used to determine the node type corresponding to each node based on the connection relationship; The first graph construction module is used to construct an operation relationship graph corresponding to the target system based on the nodes, the connection relationships, the node types corresponding to the nodes, and the parameter control relationships corresponding to the target system.

[0083] In some other embodiments, the first graphics construction module includes: A direction determination module, used to determine the edge direction corresponding to each edge in the operation relationship graph based on the nodes and the connection relationship; A threshold determination module, configured to determine an edge capacity value corresponding to each edge based on the parameter control relationship corresponding to each node and the target system; the edge capacity value represents an energy transfer threshold between components in the target system; The second graph construction module is used to construct an operation relationship graph corresponding to the target system based on the nodes, the node types corresponding to the nodes, the edge directions corresponding to the edges, and the edge capacity values.

[0084] In some other embodiments, the information extraction module 830 includes: A traversal times determination module, used to obtain the traversal times of the operation relationship graph; A node height determination module, used to determine the height of the source node, the height of the sink node and the height of the intermediate node respectively based on the number of traversals; An edge flow value determination module, used to calculate the edge flow value corresponding to the edge related to the source node in the operation relationship graph; the edge related to the source node is an edge formed between the source node and an adjacent node of the source node; A first matrix construction module, used to construct a first matrix based on the height of the source node, the height of the sink node and the height of the intermediate node; A second matrix construction module, used for constructing a second matrix based on the edge flow value corresponding to each edge related to the source node; The information determination module is used to determine that the first matrix and the second matrix are the operation characteristic information.

[0085] In some other embodiments, the node height determination module includes: A first height determination module, used to determine the number of nodes in the operation relationship graph as the height of the source node; A second height determination module, used to determine that the height of the sink node is 0; A third height determination module, configured to determine that the height of the intermediate node is 0 if the traversal number is 1, and update the traversal number; The fourth height determination module is used to determine the height of the intermediate node based on the node height in the previous traversal process if the traversal number is greater than 1, and to update the traversal number.

[0086] In some other embodiments, the device further comprises: A path detection module, configured to perform path detection on each edge in the running state graph to obtain a path detection result if the traversal number is greater than 1; A height updating module is used to update the height of the intermediate node if the path detection result indicates that the starting node of at least one edge in the running status graph is the source node and the ending node is the sink node.

[0087] In some other embodiments, the device further comprises: A data acquisition module, used to acquire a graph neural network and sample training data; the sample training data includes historical operation data corresponding to the target system at multiple historical operation moments, and a parameter control relationship corresponding to the target system; A model training module is used to input the sample training data into the graph neural network so that the graph neural network performs model training based on the sample training data and obtains the target control model corresponding to the target system.

[0088] The device and method embodiments in the device embodiment are based on the same inventive concept and are used to implement the above-mentioned data center precision system energy-saving optimization method based on data fusion.

[0089] An embodiment of the present invention also provides an electronic device, comprising: a processor and a memory, wherein the memory stores at least one instruction, at least one program, a code set or an instruction set, and the at least one instruction, the at least one program, the code set or the instruction set is loaded and executed by the processor to implement the data fusion-based data center precision system energy-saving optimization method as described in any one of the method embodiments.

[0090] The embodiment of the present invention also provides a storage medium, and the computer-readable storage medium can be a tangible device that holds and stores instructions used by an instruction execution device. The computer-readable storage medium can be an electrical storage device, a magnetic storage device, an optical storage device, an electromagnetic storage device, a semiconductor storage device, or any suitable combination of the above. More specific examples (non-exhaustive list) of computer-readable storage media include: portable computer disks, hard disks, random access memories (RAM), read-only memories (ROM), erasable programmable read-only memories (EPROM or flash memory), static random access memories (SRAM), portable compact disk read-only memories (CD-ROM), digital versatile disks (DVD), memory sticks, floppy disks, mechanical encoding devices, such as punch cards or protruding structures in grooves on which instructions are stored, and any suitable combination of the above. The computer-readable storage medium used here is not interpreted as a transient signal itself, such as a radio wave or other freely propagating electromagnetic wave, an electromagnetic wave propagated by a waveguide or other transmission medium (e.g., a light pulse by an optical fiber cable), or an electrical signal transmitted by a wire.

[0091] The computer-readable program instructions described herein can be downloaded from a computer-readable storage medium to each computing / processing device, or downloaded to an external computer or external storage device via a network, such as the Internet, a local area network, a wide area network, and / or a wireless network. The network can include copper transmission cables, optical fiber transmission, wireless transmission, routers, firewalls, switches, gateway computers, and / or edge servers. The network adapter card or network interface in each computing / processing device receives the computer-readable program instructions from the network and forwards the computer-readable program instructions for storage in the computer-readable storage medium in each computing / processing device.

[0092] The computer program instructions for performing the operations of the present disclosure may be assembly instructions, instruction set architecture (ISA) instructions, machine instructions, machine-related instructions, microcode, firmware instructions, state setting data, or source code or object code written in any combination of one or more programming languages, including object-oriented programming languages ​​such as Smalltalk, C++, etc., and conventional procedural programming languages ​​such as "C" language or similar programming languages. The computer-readable program instructions may be executed entirely on the user's computer, partially on the user's computer, as a separate software package, partially on the user's computer and partially on a remote computer, or entirely on a remote computer or server. In the case of a remote computer, the remote computer may be connected to the user's computer through any type of network, including a local area network (LAN) or a wide area network (WAN), or may be connected to an external computer (e.g., using an Internet service provider to connect through the Internet). In some embodiments, by using the state information of the computer-readable program instructions to personalize an electronic circuit, such as a programmable logic circuit, a field programmable gate array (FPGA), or a programmable logic array (PLA), the electronic circuit may execute the computer-readable program instructions, thereby implementing various aspects of the present disclosure.

[0093] Various aspects of the present disclosure are described herein with reference to flowcharts and / or block diagrams of methods, devices (systems) and computer program products according to embodiments of the present disclosure. It should be understood that each box in the flowchart and / or block diagram and the combination of boxes in the flowchart and / or block diagram can be implemented by computer-readable program instructions.

[0094] These computer-readable program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, or other programmable data processing device, thereby producing a machine, so that when these instructions are executed by the processor of the computer or other programmable data processing device, a device that implements the functions / actions specified in one or more boxes in the flowchart and / or block diagram is generated. These computer-readable program instructions can also be stored in a computer-readable storage medium, and these instructions cause the computer, programmable data processing device, and / or other equipment to work in a specific manner, so that the computer-readable medium storing the instructions includes a manufactured product, which includes instructions for implementing various aspects of the functions / actions specified in one or more boxes in the flowchart and / or block diagram.

[0095] Computer-readable program instructions may also be loaded onto a computer, other programmable data processing apparatus, or other device so that a series of operating steps are performed on the computer, other programmable data processing apparatus, or other device to produce a computer-implemented process, thereby causing the instructions executed on the computer, other programmable data processing apparatus, or other device to implement the functions / actions specified in one or more boxes in the flowchart and / or block diagram.

[0096] The flow chart and block diagram in the accompanying drawings show the possible architecture, function and operation of the system, method and computer program product according to multiple embodiments of the present disclosure. In this regard, each square frame in the flow chart or block diagram can represent a part of a module, program segment or instruction, and a part of the module, program segment or instruction includes one or more executable instructions for realizing the specified logical function. In some alternative implementations, the functions marked in the square frame can also occur in a sequence different from that marked in the accompanying drawings. For example, two continuous square frames can actually be executed substantially in parallel, and they can sometimes be executed in reverse order, depending on the functions involved. It should also be noted that each square frame in the block diagram and / or flow chart, and the combination of the square frames in the block diagram and / or flow chart can be implemented with a dedicated hardware-based system that performs the specified function or action, or can be implemented with a combination of special-purpose hardware and computer instructions.

[0097] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention rather than to limit it. Although the present invention has been described in detail with reference to the above embodiments, those skilled in the art should understand that the specific implementation modes of the present invention can still be modified or replaced by equivalents, and any modification or equivalent replacement that does not depart from the spirit and scope of the present invention should be covered within the scope of protection of the claims of the present invention.

Claims

1. A data center precision system energy saving optimization method based on data fusion, characterized in that: The method comprises: Acquire historical operation data corresponding to the target system at the historical operation time; the historical operation data includes various system operation parameters corresponding to the target system; Constructing an operation relationship diagram corresponding to the target system based on the operation parameters of each system; Extracting characteristic information from the operation relationship diagram to obtain operation characteristic information corresponding to the historical operation data; Acquire a target control model, and acquire current operation data corresponding to the target system at the current operation time; the target control model is trained based on the operation characteristic information and the parameter control relationship corresponding to the target system; The current operation data is input into the target control model to obtain the operation status control information corresponding to the target system.

2. According to claim 1, a data center precision system energy saving optimization method based on data fusion is characterized in that: The constructing an operation relationship diagram corresponding to the target system based on the operation parameters of each system includes: Determine a corresponding node in the operation relationship diagram based on each system operation parameter; Acquire the connection relationship between the components in the target system and the parameter control relationship corresponding to the target system; Based on the connection relationship, determine the node type corresponding to each node; Based on the various nodes, the connection relationships, the node types corresponding to the various nodes, and the parameter control relationships corresponding to the target system, an operation relationship diagram corresponding to the target system is constructed.

3. The method for optimizing energy saving of a data center precision system based on data fusion according to claim 2 is characterized in that: The constructing an operation relationship diagram corresponding to the target system based on the nodes, the connection relationship, the node type corresponding to the nodes, and the parameter change relationship corresponding to the target system includes: Based on the nodes and the connection relationships, determine the edge direction corresponding to each edge in the operation relationship graph; Based on the parameter control relationship between each node and the target system, determining an edge capacity value corresponding to each edge; the edge capacity value represents an energy transfer threshold between components in the target system; An operation relationship graph corresponding to the target system is constructed based on the nodes, the node types corresponding to the nodes, the edge directions corresponding to the edges, and the edge capacity values.

4. The method for optimizing energy saving of a data center precision system based on data fusion according to claim 1, characterized in that: The node types include source nodes, sink nodes and intermediate nodes; The extracting characteristic information from the operation relationship diagram to obtain operation characteristic information corresponding to the historical operation data includes: Obtaining the number of traversals of the running relationship graph; Based on the number of traversals, respectively determine the height of the source node, the height of the sink node, and the height of the intermediate node; Calculating the edge flow value corresponding to the edge related to the source node in the operation relationship graph; the edge related to the source node is an edge formed between the source node and an adjacent node of the source node; Constructing a first matrix based on the height of the source node, the height of the sink node, and the height of the intermediate node; Constructing a second matrix based on the edge flow values ​​corresponding to the edges related to each source node; The first matrix and the second matrix are determined as the operation characteristic information.

5. The method for optimizing energy saving of a data center precision system based on data fusion according to claim 4 is characterized in that: The determining, based on the number of traversals, the height of the source node, the height of the sink node, and the height of the intermediate node respectively includes: Determine the number of nodes in the operation relationship graph as the height of the source node; Determine that the height of the sink node is 0; If the traversal count is 1, the height of the intermediate node is determined to be 0, and the traversal count is updated; If the traversal times is greater than 1, the height of the intermediate node is determined based on the node height in the previous traversal process, and the traversal times is updated.

6. The method for optimizing energy saving of a data center precision system based on data fusion according to claim 5, characterized in that: The method further comprises: If the traversal times is greater than 1, performing path detection on each edge in the running state graph to obtain a path detection result; If the path detection result indicates that there is at least one edge in the operating status graph whose starting node is the source node and whose ending node is the sink node, the height of the intermediate node is updated.

7. The method for optimizing energy saving of a data center precision system based on data fusion according to claim 1, characterized in that: The method further comprises: Acquire a graph neural network and sample training data; the sample training data includes historical operation data corresponding to the target system at multiple historical operation moments, and a parameter control relationship corresponding to the target system; The sample training data is input into the graph neural network so that the graph neural network performs model training based on the sample training data and obtains the target control model corresponding to the target system.

8. A data center precision system energy-saving optimization device based on data fusion, characterized in that: The device comprises: A first acquisition module is used to acquire historical operation data corresponding to the target system at the historical operation time; the historical operation data includes various system operation parameters corresponding to the target system; A directed graph construction module, used to construct an operation relationship graph corresponding to the target system based on the operation parameters of each system; An information extraction module, used to extract characteristic information from the operation relationship diagram to obtain operation characteristic information corresponding to the historical operation data; A second acquisition module is used to acquire a target control model and current operation data corresponding to the target system at the current operation time; the target control model is trained based on the operation characteristic information and the parameter control relationship corresponding to the target system; The information output module is used to input the current operation data into the target control model to obtain the operation status control information corresponding to the target system.

9. An electronic device, characterized in that: The device includes a processor and a memory, wherein the memory stores at least one instruction and at least one program, and the at least one instruction and the at least one program are loaded and executed by the processor to implement the data center precision system energy-saving optimization method based on data fusion as described in any one of claims 1-7.

10. A computer storage medium, characterized in that: The computer storage medium stores at least one instruction and at least one program, and the at least one instruction and the at least one program are loaded and executed by the processor to implement the data fusion-based energy-saving optimization method for a data center precision system as described in any one of claims 1-7.

Images