Data center unit demand response cooperative control method based on shapley algorithm

Abstract

The present application relates to the technical field of data center unit demand response collaborative control, solves the technical problem that the cooperation between different data center operators under the same voltage level node in the industrial park is limited by management authority, and the time and space transfer of work load cannot be realized, and particularly relates to a data center unit demand response collaborative control method based on Shapley algorithm, which comprises the following steps: obtaining user parameters of each data center in the industrial park, and constructing a characteristic parameter matrix of data center users according to the user parameters; based on the characteristic parameter matrix of the data center users, the interactive work load of the data centers participating in the cooperation in the industrial park is equally spaced clustered according to the processing time delay. The present application can realize the cooperation of different data center operators, integrate the computing power resources of the data centers in the same park, make the work load flexibly transfer on the servers of different data center operators, and achieve the purpose of electricity demand regulation.

Description

Technical Field

[0001] This invention relates to the field of demand response collaborative control technology for data center units, and in particular to a demand response collaborative control method for data center units based on the Shapley algorithm. Background Technology

[0002] Currently, data centers are primarily located in industrial parks, which are equipped with extensive integrated energy facilities and distributed renewable energy sources. Data centers can collaborate with integrated energy operators within the park through demand response initiatives to absorb renewable energy sources such as distributed photovoltaic power within the park.

[0003] Existing technologies primarily model data center workloads to facilitate the temporal or spatial shifting of workloads across multiple or individual data centers under the same data center operator. This alters the timing and location of task processing on servers, thereby adjusting data center power demands. However, this approach emphasizes that data centers belong to the same cloud and operator, with unified operator management driving the temporal and spatial shifting of workloads to regulate power demand. In industrial parks, however, at the same voltage level, data centers are treated as a single entity. Servers within the park belong to different operators, and server usage is restricted by management permissions, making temporal and spatial workload shifting impossible. Summary of the Invention

[0004] To address the shortcomings of existing technologies, this invention provides a collaborative control method for demand response of data center units based on the Shapley algorithm. This method solves the technical problem that, under the same voltage level node in an industrial park, collaboration between different data center operators is limited by management authority, making it impossible to achieve time and space transfer of workloads. This method achieves the goal of enabling collaboration between different data center operators and integrating the computing resources of data centers within the same park.

[0005] To address the aforementioned technical problems, this invention provides the following technical solution: a collaborative control method for demand response of data center units based on the Shapley algorithm, comprising the following processes:

[0006] Obtain user parameters for each data center within the industrial park, and construct a feature parameter matrix for data center users based on these user parameters;

[0007] Based on the feature parameter matrix of data center users, the interactive workloads of data centers participating in the collaboration within the industrial park are clustered at equal intervals according to processing latency.

[0008] From data center collection Obtain data from various data centers Demand response parameters;

[0009] A heterogeneous data center collaboration model is constructed based on demand response parameters to enable collaboration between multiple data centers and integrated energy systems within an industrial park.

[0010] Using a heterogeneous data center collaboration model as a constraint, the operating cost function of data center operators and integrated energy system operators is optimized based on the Shapley algorithm, and the demand response control strategy that maximizes the interests of different data center operators and integrated energy operators is output.

[0011] Furthermore, the user parameters include , , , , , , , , , and ;

[0012] in, This represents the collection of all data centers within the industrial park; This indicates that the industrial park includes data centers. A collection of data centers that collaborate within the same organization can transfer interactive workloads to each other. This indicates the number of collaborative data centers within the industrial park; Indicates the number of data centers; This indicates the total number of interactive workloads in the data center; This indicates the total number of batch processing workloads in the data center; Indicates the first The first data center An interactive workload during a time period The workload to be processed; Indicates the first The first data center An interactive workload during a time period The maximum processing latency; Indicates the first The first data center Individual user batch processing workload; and They represent the first The first data center The processing deadline and submission time for each user's batch workload.

[0013] Furthermore, the feature parameter matrix of the data center users and The definition is as follows:

[0014]

[0015] In the formula, For the first The first data center An interactive workload during a time period The feature parameter matrix; For the first The first data center Characteristic parameters of individual user batch processing workloads.

[0016] Furthermore, the process of clustering the interactive workloads of the collaborative data centers within the industrial park according to processing latency at equal intervals includes the following steps:

[0017] S21. Introduce categories as auxiliary variables. The interactive workload processing latency is divided into several time slots, namely ,in For time period Minimum processing latency for interactive workloads across all participating data centers within the industrial park. For time period Maximum processing latency for all interactive workloads For time intervals;

[0018] S22, Introducing Variables and ,make ,variable As an auxiliary variable, Indicates the time period The first to be calculated Interactive workloads;

[0019] S23, Traverse every data center participating in the collaboration within the industrial park. Feature parameter matrix of each interactive workload Based on the maximum processing latency and The relationship is updated according to the following formula during the time period. The first to be calculated Interactive workloads ,Right now:

[0020]

[0021] In the formula, Indicates the first Lower limit of latency for interactive workloads; Indicates the first Maximum latency for interactive workloads Indicates time period Data center latency for interactive workloads.

[0022] S24. Determine whether the following conditions are met. If they are met, proceed to step S26; otherwise, proceed to step S25.

[0023]

[0024] In the formula, For workload processing speed;

[0025] S25. Determine the time period The first to be calculated Interactive workloads ,Right now:

[0026]

[0027] S26, Order ,make and judge Is it equal to If yes, the process ends; otherwise, return to step S22.

[0028] Furthermore, the demand response parameters include , , , , , ;

[0029] in, Indicates data center Energy efficiency coefficient; and These represent the server's peak power and idle power, respectively. For workload processing speed; This refers to the time interval for data center workload scheduling. Indicates data center The total number of servers.

[0030] Furthermore, the heterogeneous data center collaborative model includes a data center system model, an integrated energy system optimization model, and a cost function.

[0031] Furthermore, the data center system model includes a data center energy consumption model, a workload scheduling model, and a service level agreement model, namely:

[0032] The data center energy consumption model is as follows:

[0033]

[0034] in, Indicates data center During the period Energy consumption; It is a data center During the period The number of active servers; Indicates data center The server during the time period The workload to be processed;

[0035] The workload scheduling model is as follows:

[0036]

[0037] In the formula, and They respectively represent data centers During the period Processing batch workloads and interactive workloads; Indicates data center During the period The first processing Batch processing workload per user; Indicates data center During the period The first processing Interactive workloads;

[0038] The service level agreement model is as follows:

[0039]

[0040] In the formula, and They represent data centers During the period Batch processing and the first The number of servers for interactive workloads;

[0041] The integrated energy system optimization model includes the operating characteristic model of combined heat and power equipment and the power balance model of the integrated energy system, namely:

[0042] The operating characteristic model of a combined heat and power (CHP) unit is as follows:

[0043]

[0044] In the formula, and These represent the time periods of the combined heat and power (CHP) equipment. The electrical and thermal power output to the outside; , , , , These represent the energy conversion efficiency of the cogeneration equipment, the calorific value of natural gas, the heating coefficient, the power generation efficiency of the cogeneration, and the natural gas input, respectively. and The minimum and maximum gas volumes input to the combined heat and power (CHP) equipment; This indicates the maximum ramp power of the combined heat and power (CHP) equipment.

[0045] The integrated energy system power balance model is as follows:

[0046]

[0047] In the formula, Indicates time period Waste heat generated by data centers; For consumers within the park during certain time periods The heat demand; To meet the electricity demand of consumers within the park during time period t; This indicates the output of distributed photovoltaic new energy within the industrial park; Indicates the upper limit of distributed photovoltaic new energy output; This indicates the electrical power supplied by the grid.

[0048] The cost functions for data center operators and integrated energy system operators are:

[0049]

[0050]

[0051] In the formula, and These represent the time periods for data center operators and integrated energy system operators, respectively. Operating costs; and These are the unit price of natural gas and the price of electricity; It is the incentive price given by the power grid; It is the price for consumers to receive heating.

[0052] Furthermore, the process of using a heterogeneous data center collaboration model as a constraint, optimizing the operating cost function of data center operators and integrated energy system operators based on the Shapley algorithm, and outputting a demand response control strategy that maximizes the interests of different data center operators and integrated energy operators includes the following steps:

[0053] S51, Initialization , Order number This refers to an integrated energy operator; when an integrated energy operator collaborates with a data center operator, it aggregates... When integrated energy operators do not cooperate with data center operators, ;in, , This represents the collection of all potential collaborators, including data center operators and integrated energy operators. yes A subset of;

[0054] S52. Analyze whether the following conditions are met;

[0055]

[0056]

[0057] In the formula, This indicates the serial number of the data center operator or integrated energy operator participating in the collaboration; Indicates the first The optimal operating cost for each participant is also the Shapley value to be determined. The cost can be obtained using the cost function model of the heterogeneous data center collaborative model. Indicates the first Participants and sets The optimal total operating cost for all participants; Represents a set The optimal total operating cost for all participants; This represents the operating costs without collaboration. It is an empty set;

[0058] S53. Calculate the Shapley value of the collaborator using the following formula:

[0059]

[0060] In the formula, Indicates participants The Shapley value; and Represents a set and Number of participants

[0061] S54, Based on Shapley value Output demand response control strategy, which includes , , , , ;in, and These represent the time periods of the combined heat and power (CHP) equipment. The electrical and thermal power output to the outside; Indicates time period Waste heat generated by data centers; This indicates the output of distributed photovoltaic new energy within the industrial park; Indicates the amount of natural gas input; This indicates the electrical power supplied by the grid.

[0062] By employing the above technical solution, the present invention provides a data center unit demand response collaborative control method based on the Shapley algorithm, which has at least the following beneficial effects:

[0063] 1. This invention enables collaboration among different data center operators, integrates computing resources within the same park, allows workloads to be flexibly transferred between servers of different data center operators, achieves the purpose of regulating power demand, and enables data center operators and integrated energy providers to cooperate in operation, thereby maximizing the interests of different data center operators and integrated energy providers through integrated energy strategy optimization.

[0064] 2. This invention can clearly identify workload groups with different processing latency requirements through cluster analysis. Based on the real-time characteristics and latency requirements of the workload, the resource allocation strategy can be dynamically adjusted to ensure that critical tasks receive sufficient computing resources. At the same time, it can identify workload groups with similar processing latency requirements, thereby promoting resource sharing and collaborative work among data centers and improving overall processing capacity and efficiency. Attached Figure Description

[0065] The accompanying drawings, which are included to provide a further understanding of this application and form part of this application, illustrate exemplary embodiments of this application and are used to explain this application, but do not constitute an undue limitation of this application. In the drawings:

[0066] Figure 1 This is a flowchart of the data center unit demand response collaborative control method of the present invention. Detailed Implementation

[0067] To make the above-mentioned objects, features, and advantages of the present invention more apparent and understandable, the present invention will be further described in detail below with reference to the accompanying drawings and specific embodiments. This will allow for a full understanding of how the present application uses technical means to solve technical problems and achieve technical effects, and to facilitate its implementation.

[0068] Data centers' participation in demand response is crucial. With the acceleration of digital transformation, data center energy consumption is constantly increasing, and electricity demand is rapidly rising. How to achieve coordinated load control of data centers, participate in grid supply and demand interaction, and optimize operating costs has become a key issue for data center operators. By participating in demand response, data centers can adjust their power consumption patterns during periods of energy shortage, reducing the load pressure on the grid, improving energy efficiency, and lowering operating costs. Furthermore, proactively responding to electricity demand helps ensure the stability of power supply, reduces energy waste, and aligns with sustainable development goals. Data center participation not only benefits the operations of the enterprises themselves but also promotes sustainable energy use throughout society, contributing to environmental protection and resource conservation.

[0069] Currently, data centers are primarily located in industrial parks, which are equipped with extensive integrated energy facilities and distributed renewable energy sources. Data centers can collaborate with integrated energy operators within the park through demand response mechanisms to absorb renewable energy sources such as distributed photovoltaic power. Current technical solutions mainly involve scheduling the workloads processed by the data center to alter the timing and spatial location of computing tasks, thereby adjusting the data center's power demand and matching it with the grid's demand response regulation requirements to absorb renewable energy.

[0070] However, in industrial parks, data centers are treated as a single entity at the same voltage level. The data center servers within the park belong to different operators, and server usage is restricted by management authority, making it impossible to transfer workloads over time and space.

[0071] To achieve time and space shifting of workloads, please refer to Figure 1 This embodiment proposes a collaborative control method for data center unit demand response based on the Shapley algorithm. This method enables collaboration among different data center operators, integrates computing resources within the same campus, and allows workloads to be flexibly transferred between servers of different data center operators, thereby achieving the goal of regulating power demand. The method includes the following steps:

[0072] S1. Obtain user parameters for each data center within the industrial park, and construct a feature parameter matrix for data center users based on the user parameters; in this embodiment, the user parameters include , , , , , , , , , and ,in, This represents the collection of all data centers within the industrial park; This indicates that the industrial park includes data centers. A collection of data centers that collaborate within the same organization can transfer interactive workloads to each other. This indicates the number of collaborative data centers within the industrial park; Indicates the number of data centers; This indicates the total number of interactive workloads in the data center; This indicates the total number of batch processing workloads in the data center; Indicates the first The first data center An interactive workload during a time period The workload to be processed; Indicates the first The first data center An interactive workload during a time period The maximum processing latency; Indicates the first The first data center Batch processing workload per user; and They represent the first The first data center The processing deadline and submission time for each user's batch workload.

[0073] Feature parameter matrix of data center users and The definition is as follows:

[0074]

[0075] In the formula, For the first The first data center An interactive workload during a time period The feature parameter matrix; For the first The first data center Characteristic parameters of individual user batch processing workloads.

[0076] S2. Based on the feature parameter matrix of data center users, the interactive workloads of the data centers participating in the collaboration within the industrial park are clustered at equal intervals according to processing latency; in step S2, the specific process includes the following steps:

[0077] S21. Introduce categories as auxiliary variables. The interactive workload processing latency is divided into several time slots, namely ,in For time period Minimum processing latency for interactive workloads across all participating data centers within the industrial park. For time period Maximum processing latency for all interactive workloads For time intervals;

[0078] S22, Introducing Variables and ,make ,variable As an auxiliary variable, Indicates the time period The first to be calculated Interactive workloads;

[0079] S23, Traverse every data center participating in the collaboration within the industrial park. Feature parameter matrix of each interactive workload Based on the maximum processing latency and The relationship is updated according to the following formula during the time period. The first to be calculated Interactive workloads ,Right now:

[0080]

[0081] In the formula, Indicates the first Lower limit of latency for interactive workloads; Indicates the first Maximum latency for interactive workloads Indicates time period Data center latency for interactive workloads.

[0082] In this embodiment, the maximum processing latency and The relationship, namely: the analysis period Each interactive workload Maximum processing latency Is it between Between; if so, then the first The first data center An interactive workload during a time period Processing workload Clustered as the first Interactive workloads If not, proceed to step S24; otherwise, return to step S21 to perform equal-space clustering on the next data center participating in the collaboration within the industrial park, thereby completing equal-space clustering on all data centers participating in the collaboration within the industrial park.

[0083] S24. Determine whether the following conditions are met. If they are met, proceed to step S26; otherwise, proceed to step S25.

[0084]

[0085] In the formula, For workload processing speed;

[0086] S25. Determine the time period The first to be calculated Interactive workloads ,Right now:

[0087]

[0088] S26, Order ,make and judge Is it equal to If yes, the process ends; otherwise, return to step S22.

[0089] This embodiment achieves multiple technical effects by clustering interactive workloads of collaborative data centers within an industrial park according to their processing latency at equal intervals. This optimizes resource allocation, improves service quality, enhances system stability, optimizes cost-effectiveness, promotes collaborative work, and improves decision support capabilities. Ultimately, this helps improve the overall performance and operational efficiency of the data center, providing more efficient, stable, and reliable data services to enterprises and users within the industrial park.

[0090] Specifically, cluster analysis can clearly identify workload groups with different processing latency requirements, thereby dynamically adjusting resource allocation strategies based on the real-time characteristics and latency requirements of the workload to ensure that critical tasks receive sufficient computing resources. At the same time, it can identify workload groups with similar processing latency requirements, thereby promoting resource sharing and collaborative work among data centers and improving overall processing capacity and efficiency.

[0091] S3, from data center aggregation Obtain data from various data centers Demand response parameters, including , , , , , The demand response parameters can be obtained by either reporting them directly by the data center operator or extracting them from the data center's load scheduling system.

[0092] in, Indicates data center Energy efficiency coefficient; and These represent the server's peak power and idle power, respectively. Indicates data center The total number of servers.

[0093] S4. Construct a heterogeneous data center collaboration model based on demand response parameters to realize the collaboration of multiple data centers and integrated energy systems within an industrial park; the heterogeneous data center collaboration model includes a data center system model, an integrated energy system optimization model, and a cost function; the data center system model includes a data center energy consumption model, a workload scheduling model, and a service level agreement model; the integrated energy system optimization model includes a combined heat and power equipment operation characteristic model and an integrated energy system power balance model.

[0094] The data center energy consumption model is as follows:

[0095]

[0096] in, Indicates data center During the period Energy consumption; It is a data center During the period The number of active servers; Indicates data center The server during the time period The amount of workload processed.

[0097] The workload scheduling model includes both batch and interactive workloads; therefore, the workload scheduling model is as follows:

[0098]

[0099] In the formula, and They respectively represent data centers During the period Processing batch workloads and interactive workloads; Indicates data center During the period The first processing Batch processing workload per user; Indicates data center During the period The first processing Interactive workloads;

[0100] The service level agreement model is as follows:

[0101]

[0102] In the formula, and They represent data centers During the period Batch processing and the first The number of servers for interactive workloads.

[0103] Since the integrated energy system within the industrial park includes combined heat and power (CHP) equipment, the integrated energy system optimization model includes both the CHP equipment operation characteristic model and the integrated energy system power balance model, as follows:

[0104] The operating characteristic model of a combined heat and power (CHP) unit is as follows:

[0105]

[0106] In the formula, and These represent the time periods of the combined heat and power (CHP) equipment. The electrical and thermal power output to the outside; , , , , These represent the energy conversion efficiency of the cogeneration equipment, the calorific value of natural gas, the heating coefficient, the power generation efficiency of the cogeneration, and the natural gas input, respectively. and The minimum and maximum gas volumes input to the combined heat and power (CHP) equipment; This indicates the maximum ramp power of the combined heat and power (CHP) equipment.

[0107] The integrated energy system power balance model is as follows:

[0108]

[0109] In the formula, Indicates time period Waste heat generated by data centers; For consumers within the park during certain time periods The heat demand; To meet the electricity demand of consumers within the park during time period t; This indicates the output of distributed photovoltaic new energy within the industrial park; Indicates the upper limit of distributed photovoltaic new energy output; This indicates the electrical power supplied by the grid.

[0110] The cost functions for data center operators and integrated energy system operators are:

[0111]

[0112]

[0113] In the formula, and These represent the time periods for data center operators and integrated energy system operators, respectively. Operating costs; and These are the unit price of natural gas and the price of electricity; It is the incentive price given by the power grid; It is the price for consumers to receive heating.

[0114] S5. Using the heterogeneous data center collaboration model as a constraint, optimize the operating cost function of data center operators and integrated energy system operators based on the Shapley algorithm, and output a demand response control strategy that maximizes the benefits for different data center operators and integrated energy operators. Step S5 specifically includes the following steps:

[0115] S51, Initialization , Order number This refers to an integrated energy operator; when an integrated energy operator collaborates with a data center operator, it aggregates... When integrated energy operators do not cooperate with data center operators, ;in, , This represents the collection of all potential collaborators, including data center operators and integrated energy operators. yes A subset of;

[0116] S52. Analyze whether the following conditions are met;

[0117]

[0118]

[0119] In the formula, This indicates the serial number of the data center operator or integrated energy operator participating in the collaboration; Indicates the first The optimal operating cost for each participant is also the Shapley value to be determined. The cost can be obtained using the cost function model of the heterogeneous data center collaborative model. Indicates the first Participants and sets The optimal total operating cost for all participants; Represents a set The optimal total operating cost for all participants; This represents the operating costs without collaboration. It is an empty set.

[0120] S53. Calculate the Shapley value of the collaborator using the following formula:

[0121]

[0122] In the formula, Indicates participants The Shapley value; and Represents a set and Number of participants

[0123] S54, Based on Shapley value Output demand response control strategy, which includes , , , , .in, and These represent the time periods of the combined heat and power (CHP) equipment. The electrical and thermal power output to the outside; Indicates time period Waste heat generated by data centers; This indicates the output of distributed photovoltaic new energy within the industrial park; Indicates the amount of natural gas input; This indicates the electrical power supplied by the grid.

[0124] This embodiment enables collaboration among different data center operators, integrating computing resources within the same park, allowing workloads to be flexibly transferred between servers of different data center operators, thereby achieving the goal of regulating power demand. At the same time, it enables data center operators and integrated energy providers to collaborate and maximize the benefits of both through integrated energy strategy optimization.

[0125] Those skilled in the art will understand that all or part of the steps in the methods of the above embodiments can be implemented by a program instructing related hardware. Therefore, this application can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Moreover, this application can take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0126] The above embodiments provide a detailed description of the present invention. Specific examples have been used to illustrate the principles and implementation methods of the present invention. The descriptions of the above embodiments are only for the purpose of helping to understand the method and core ideas of the present invention. At the same time, for those skilled in the art, there will be changes in the specific implementation methods and application scope based on the ideas of the present invention. Therefore, the content of this specification should not be construed as a limitation of the present invention.

Claims

1. A method for coordinated control of data center chiller unit demand response based on Shapley algorithm, characterized in that, The method The method comprises the following process: Obtaining user parameters of each data center in the industrial park, and constructing a feature parameter matrix of data center users according to the user parameters; Based on the feature parameter matrix of the data center users, interactive workloads of the data centers participating in cooperation in the industrial park are equally spaced clustered according to processing time delay; from a set of data centers obtain demand response parameters for each data center from the set of data centers; Based on the demand response parameters, a heterogeneous data center collaboration model for realizing cooperation of multiple data centers and integrated energy systems in the industrial park is constructed, and the heterogeneous data center collaboration model comprises a data center system model, an integrated energy system optimization model, and a cost function; The data center system model comprises a data center energy consumption model, a workload scheduling model, and a service level agreement model, namely: The data center energy consumption model is: wherein, represents the energy consumption of a data center at a time period ; represents the energy efficiency coefficient of a data center ; and respectively represent the peak power and idle power of a server; is the time interval of the data center scheduling the workload; is the processing speed of the workload; is the number of active servers of a data center at a time period ; represents the amount of workload processed by the servers of a data center at a time period ; The workload scheduling model is: In the formula, and They respectively represent data centers During the period Processing batch workloads and interactive workloads; Indicates the first The first data center Individual user batch processing workload; Indicates data center During the period The first processing Batch processing workload per user; Indicates data center During the period The first processing Interactive workloads; and They represent the first The first data center The processing deadline and submission time for each user's batch workload; This is the time interval coefficient; This indicates the number of collaborative data centers within the industrial park; This indicates the total number of batch processing workloads in the data center; The service level agreement model is: wherein and denote the data centers The total number of servers of the data centers The processing batch and the first The number of servers for the interactive workloads; denote the data centers The total number of servers of the data centers The minimum processing latency for the interactive workloads of all participating data centers within the industrial park for the time period The processing latency coefficient of the data centers for the interactive workloads for the time period The processing latency coefficient of the data centers for the interactive workloads for the time period The processing latency coefficient of the data centers for the interactive workloads for the time period The heterogeneous data center collaboration model is taken as a constraint condition, and the operation cost function of the data center operator and the integrated energy system operator is optimized based on the Shapley algorithm, and a demand response control strategy for realizing maximum benefits of different data center operators and integrated energy operators is output. 2.The data center chiller plant demand response cooperative control method of claim 1, wherein, The user parameters include , , , , , , , , , and ; wherein, represents the set of all data centers within the industrial park; represents the set of data centers within the industrial park that participate in collaboration, i.e., that can transfer interactive workloads to each other; represents the number of data centers within the industrial park that collaborate; represents the number of data centers; represents the total number of interactive workloads of a data center; represents the total number of batch workloads of a data center; represents the workload amount of the th interactive workload of the th data center during the time period ; represents the maximum processing delay of the th interactive workload of the th data center during the time period ; represents the workload amount of the th batch workload of the th user of a data center; and represent the processing deadline and the submission time of the th batch workload of the th user of a data center, respectively. 3.The data center chiller plant demand response cooperative control method of claim 2, wherein, a matrix of characteristic parameters of the data center users and are defined as follows: wherein is a feature parameter of the th interactive workload of the th data center at time period ; is a feature parameter of the th user batch workload of the th data center.

4. The method of claim 3, wherein, The interactive workloads of the data centers participating in cooperation in the industrial park are equally spaced clustered according to processing time delay, and the specific process comprises the following steps: S21, introducing auxiliary variables and dividing the interactive workload processing latency into a number of time slots, i.e. wherein is the time period the minimum processing latency of all the data centers participating in the collaboration within the industrial park for the interactive workload, is the time period the maximum processing latency of all the interactive workloads, is the time interval coefficient; denotes the time period the processing latency coefficient of the data center for the interactive workload; S22, introduce variables and let variable be an auxiliary variable, denote the amount of interactive workloads of class to be computed in the time period ; S23. traversing each data center within the industrial park that participates in the collaboration characteristic parameter matrix of each interactive workload of the data center , according to the relationship between the maximum processing delay and , the following formula is used to update the first interactive workload amount to be calculated in the time period , that is: In the formula, represents the lower limit of the processing latency for interactive workloads of the first class; represents the lower limit of the processing latency for interactive workloads of the first class; S24, judging whether the following condition is established, if yes, entering step S26, otherwise entering step S25; S25, determining a period the first interactive workload i.e.: S26, let , let and determine whether equal to is, end, otherwise return to step S22.

5. The method of claim 4, wherein, The demand response parameters include , , , , , ; wherein, represents the total number of servers of the data center .

6. The data center chiller plant demand response coordinated control method of claim 1, wherein, The integrated energy system optimization model comprises a combined heat and power equipment operation characteristic model and an integrated energy system power balance model, namely: The combined heat and power equipment operation characteristic model is: wherein, and respectively represent the electric power and the thermal power exported by the cogeneration plant to the outside during the time interval ; , , , , respectively represent the energy conversion efficiency of the cogeneration plant, the natural gas calorific value, the heating coefficient, the cogeneration electric efficiency and the natural gas input quantity; and are the minimum and maximum gas volume input to the cogeneration plant; represents the maximum ramp power of the electric power of the cogeneration plant; The integrated energy system power balance model is: wherein denotes the time period the waste heat power generated by the data center; the thermal demand of the consumers in the industrial park during the time period ; the electric demand of the consumers in the industrial park during the time period ; denotes the distributed photovoltaic new energy output in the industrial park; denotes the upper limit of the distributed photovoltaic new energy output; denotes the electric power under the power supply of the power grid; The cost function of the data center operator and the integrated energy system operator is: wherein, and respectively represent the operation cost of the data center operator and the integrated energy system operator in each time period ; and respectively are the unit price of natural gas and electricity price; is the incentive price given by the power grid; is the price of consumer heating.

7. The method of claim 2, wherein, The heterogeneous data center collaboration model is taken as a constraint condition, and the operation cost function of the data center operator and the integrated energy system operator is optimized based on the Shapley algorithm, and a demand response control strategy for realizing maximum benefits of different data center operators and integrated energy operators is output, and the specific process comprises the following steps: S51, initialization , ; let sequence number denote the set of integrated energy operators, when the integrated energy operators cooperate with the data center operators, the set ; when the integrated energy operators do not cooperate with the data center operators, ; wherein, , denote the set of all potential collaborators, including data center operators and integrated energy operators, is a subset of ; S52, analyzing whether the following condition is met; In the formula, represents the serial number of the data center operator or integrated energy operator participating in cooperation; represents the optimal operation cost of the first participant, which is also the Shapley value to be solved, and the cost adopts the cost function model of the heterogeneous data center cooperation model; represents the optimal operation cost of the first participant, which is also the Shapley value to be solved, and the cost adopts the cost function model of the heterogeneous data center cooperation model; represents the optimal total operation cost of the first participant and all participants in the set represents the optimal total operation cost of the first participant and all participants in the set represents the optimal total operation cost of the first participant and all participants in the set represents the optimal total operation cost of all participants in the set represents the optimal total operation cost of all participants in the set represents the operation cost without cooperation, is an empty set; S53, calculating the Shapley value of the collaborator, and the calculation formula is: In the formula, Indicates participants The Shapley value; and Represents a set and Number of participants S54, based on the Shapley value outputting a demand response control strategy, the demand response control strategy comprising , , , , ; wherein, and respectively represent the electric power and the heat power outputted by the combined heat and power plant in the time period ; represents the waste heat power generated by the data center in the time period ; represents the distributed photovoltaic new energy output in the industrial park; represents the natural gas input; represents the electric power under the power grid power supply.

8. A computer-readable storage medium, characterized in that, It stores computer executable instructions, and the computer executable instructions are executed by the processor to realize the data center unit demand response collaboration control method in any one of claims 1 to 7.

9. A computer device, comprising: It comprises a processor and a memory, and the memory is used to store a computer program, and the computer program is executed by the processor to realize the data center unit demand response collaboration control method in any one of claims 1 to 7.

Images