Data center energy system collaborative optimization method and system considering task flexibility

Abstract

The application discloses a data center energy system collaborative optimization method and system considering task flexibility, and relates to the field of power and computing power collaborative optimization. The method obtains computing task data including real-time tasks and delayable tasks, constructs a computing task model and calculates a workload, establishes a delayable task queue to generate data center demand power; a comprehensive energy system model is constructed to solve the operating state with the minimum operating cost as the target; the energy side coupling coefficient is constructed according to the operating state and fed back to the computing task model, the delayable task scheduling time sequence is adjusted, and the demand power is updated; the interaction iteration is repeated until convergence, and the scheduling scheme and the operation strategy are output. Through double-layer collaborative optimization, the application realizes the coordinated matching of computing task scheduling and energy system operation, reduces task backlog and operating cost, improves photovoltaic consumption and equipment utilization efficiency, and enhances the economy and flexibility of the data center energy system.

Description

Technical Field

[0001] This invention relates to the field of power computing power collaborative optimization, and more specifically to a data center energy system collaborative optimization method and system that takes into account task flexibility. Background Technology

[0002] The statements in this section are merely background information related to the present invention and do not necessarily constitute prior art.

[0003] As a crucial infrastructure for cloud computing, big data, and intelligent computing, data centers are experiencing a continuous increase in computing tasks due to the rapid development of artificial intelligence technology, leading to a sustained rise in power consumption and cooling demands. Integrated energy systems can coordinate various energy resources, including power grids, photovoltaics, energy storage, and cooling, to provide power and cooling support for data centers, which is of great significance for reducing operating costs and improving energy efficiency.

[0004] Most existing optimization methods for integrated energy systems in data centers optimize the operation of energy equipment based on the data center load demand. Although they consider the coupling relationship between computing power, electricity and heat, they usually treat the data center load as a fixed input and focus on single-layer optimization on the energy side. They lack full utilization of the flexibility of computing task scheduling and have not established a two-way collaborative mechanism between computing task scheduling and integrated energy system operation. It is difficult to dynamically adjust the execution sequence of deferred tasks according to the energy system operation status, thus limiting the further improvement of the data center load regulation capability and the economic efficiency of integrated energy system operation. Summary of the Invention

[0005] To address the shortcomings of existing technologies, the present invention aims to provide a collaborative optimization method and system for data center energy systems that considers task flexibility. This invention achieves coordinated matching between computing task scheduling and integrated energy system operation through two-layer interactive collaboration, thereby improving the economic efficiency and energy utilization efficiency of the data center's integrated energy system.

[0006] To achieve the above objectives, the present invention adopts the following technical solution: Firstly, a collaborative optimization method for data center energy systems that considers task flexibility is provided, including: Acquire data center computing task data, which includes real-time tasks and deferred tasks; A computing task model is constructed based on the computing task data, and the workload of each task is calculated; a deferred task queue is constructed, and the deferred workload to be scheduled for execution is calculated; the total workload is determined based on the real-time task workload and the deferred workload to be scheduled for execution, and the data center power demand is generated based on the total workload. Based on the power demand of the data center, a comprehensive energy system model is constructed by combining load relationships, equipment models and energy balance constraints. With the goal of minimizing the operating cost of the comprehensive energy system, the operating status of the comprehensive energy system is obtained by solving the problem. An energy-side coupling coefficient is constructed based on the operating status of the integrated energy system and fed back to the computing task model. The scheduling sequence of deferred tasks is adjusted based on the energy-side coupling coefficient, and the power demand of the data center is updated. Repeatedly execute the process of generating data center power demand, calculating the operating status of the integrated energy system, and feeding back the energy-side coupling coefficient until the preset optimization conditions are met, and output the computing task scheduling scheme and the integrated energy system operation strategy.

[0007] Furthermore, the step of constructing a computing task model based on the computing task data and calculating the workload of each task specifically includes: constructing a computing task model based on the TPC-H benchmark test, and determining the query complexity coefficient of each query by comparing the running time of each query in the benchmark environment with the average running time. Based on the query complexity coefficient and the database size, each query is normalized to obtain the normalized workload of each query under different database sizes. Based on the real-time tasks and deferred tasks arriving within the time period, and according to the query type and database size corresponding to each task, the workload of real-time tasks and the workload of deferred tasks are calculated by summing them up.

[0008] Furthermore, the construction of a comprehensive energy system model based on the data center's power requirements, combined with load relationships, energy equipment models, and electrical / cooling energy balance constraints, specifically includes: The cooling load requirement and recoverable heat of the data center are determined based on the power demand of the data center. Based on the recoverable heat, an absorption chiller model is constructed, along with an air conditioning cooling output model and an energy storage system state-of-charge model. Based on the power demand of the data center, the photovoltaic power generation and the state of charge of the energy storage system, an electrical load balance constraint is constructed, and based on the cooling load demand of the data center, the cooling output of the absorption chiller and the cooling output of the air conditioner, a cooling load balance constraint is constructed. A comprehensive energy system model is constructed based on the data center's cooling load demand, recyclable heat, absorption chiller model, air conditioning cooling output model, energy storage system state of charge model, and electrical load balance constraints and cooling load balance constraints.

[0009] Furthermore, the goal of minimizing the operating cost of the integrated energy system and solving for the operating state of the integrated energy system specifically involves: minimizing the sum of the backlog cost of deferred tasks, grid power purchase cost, energy storage operating cost, absorption chiller operating cost, and curtailment cost of solar power; solving the integrated energy system model to obtain the optimal output of each energy device and the system state variables.

[0010] Furthermore, an energy-side coupling coefficient is constructed based on the operating status of the integrated energy system, specifically as follows: ; in, For energy-side coupling coefficient, This is a function mapping relationship. Power supplied to the power grid For photovoltaic output power, Indicates the state of charge of the energy storage system. The electrical power consumed by the air conditioner.

[0011] Furthermore, the repeated execution of the data center power demand generation, calculation of the integrated energy system operating status, and energy-side coupling coefficient feedback process until the preset optimization conditions are met specifically includes: Based on the current power demand of the data center, the integrated energy system model is solved to obtain the corresponding candidate operating states and objective function values ​​of the integrated energy system. The energy-side coupling coefficient is calculated based on the candidate operating states of the integrated energy system and fed back to the computing task model to adjust the scheduling sequence of deferred tasks, thereby obtaining the updated data center power demand. The system determines whether the preset convergence threshold is met based on the change in the objective function between two consecutive interactive iterations. If it is met, the system outputs the computational task scheduling scheme and integrated energy system operation strategy corresponding to the current iteration. If it is not met, the system continues interactive iteration based on the updated data center power demand.

[0012] Secondly, a collaborative optimization system for data center energy systems that considers task flexibility is provided, including: Data acquisition module, upper-layer computing task scheduling module, lower-layer integrated energy optimization module; The data acquisition module is used to acquire data center computing task data, which includes real-time tasks and deferred tasks. The upper-layer computing task scheduling module is used to construct a computing task model based on the computing task data, calculate the workload of each task; construct a queue of deferred tasks, calculate the deferred workload to be scheduled for execution; determine the total workload based on the real-time task workload and the deferred workload to be scheduled for execution, and generate the data center power requirement based on the total workload; receive the energy-side coupling coefficient fed back by the lower-layer integrated energy optimization module, adjust the scheduling sequence of deferred tasks based on the energy-side coupling coefficient, update the data center power requirement, and output a computing task scheduling scheme. The lower-level integrated energy optimization module receives the power demand of the data center, constructs an integrated energy system model based on the data center's power demand, combined with load relationships, equipment models, and energy balance constraints, and solves for the integrated energy system's operating status with the goal of minimizing the integrated energy system's operating cost. It then constructs an energy-side coupling coefficient based on the integrated energy system's operating status and feeds this coefficient back to the upper-level computing task scheduling module. Finally, it re-solves for the integrated energy system's operating status and energy-side coupling coefficient based on the updated data center power demand, until preset optimization conditions are met, and outputs the integrated energy system's operating strategy.

[0013] Thirdly, an electronic device is also provided, comprising: Memory is used to store computer-readable instructions in a non-transitory manner. Processor, for executing the computer-readable instructions, When the computer-readable instructions are executed by the processor, they perform the method described in the first aspect above.

[0014] Fourthly, a computer-readable storage medium is provided having a program stored thereon that, when executed by a processor, implements the method described in the first aspect above.

[0015] Fifthly, a computer program product is provided, employing the following technical solution: A computer program product includes software code, wherein a program in the software code performs the steps of the method described in the first aspect of the present invention.

[0016] The above technical solution has the following advantages or beneficial effects: This invention synergistically considers data center computing task scheduling and integrated energy system operation optimization. By distinguishing between real-time and deferred tasks, it constructs a deferred task queue and dynamically generates the data center's power demand based on task workload, fully exploring the load adjustment potential of deferred tasks over time. Simultaneously, it constructs an energy-side coupling coefficient based on the operating status of each device in the integrated energy system and feeds it back to the computing task model, guiding deferred tasks to shift to periods with better energy conditions. Through the two-layer interaction and iterative optimization of computing task scheduling and energy system operation strategies, it achieves coordinated matching between computing load and energy supply, reduces task queue backlog and integrated energy system operating costs, improves photovoltaic energy absorption, energy storage and cooling equipment utilization efficiency, and enhances the economy, flexibility, and energy utilization efficiency of the data center energy system. Attached Figure Description

[0017] The accompanying drawings, which form part of this invention, are used to provide a further understanding of the invention. The illustrative embodiments of the invention and their descriptions are used to explain the invention and do not constitute an improper limitation of the invention.

[0018] Figure 1 This is a flowchart of a data center energy system collaborative optimization method considering task flexibility in a specific embodiment of the present invention; Figure 2 This is a diagram of the integrated energy system structure for a data center. Detailed Implementation

[0019] It should be noted that the following detailed descriptions are exemplary and intended to provide further illustration of the invention. Unless otherwise specified, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention pertains.

[0020] It should be noted that the terminology used herein is for the purpose of describing particular embodiments only and is not intended to limit the exemplary embodiments of the invention. The terms "comprising" and "having," and any variations thereof, are intended to cover non-exclusive inclusion, for example, a process, method, system, product, or apparatus that comprises a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or units not explicitly listed or inherent to such processes, methods, products, or apparatus.

[0021] In this embodiment of the invention, "and / or" is merely a description of the relationship between related objects, indicating that three relationships can exist. For example, A and / or B can represent: A existing alone, A and B existing simultaneously, and B existing alone. Furthermore, in the description of this invention, "multiple" refers to two or more.

[0022] Furthermore, to facilitate a clear description of the technical solutions of the embodiments of the present invention, the terms "first" and "second" are used in the embodiments of the present invention to distinguish identical or similar items with essentially the same function and effect. Those skilled in the art will understand that the terms "first" and "second" do not limit the quantity or execution order, and the terms "first" and "second" are not necessarily different.

[0023] Where there is no conflict, the embodiments and features in the embodiments of the present invention can be combined with each other.

[0024] All data acquisition in this embodiment is carried out in compliance with laws and regulations and with user consent, and the data is used legally.

[0025] Example 1 like Figure 1 As shown, this embodiment provides a collaborative optimization method for data center energy systems that considers task flexibility, including: S101: Obtain data center computing task data, which includes real-time tasks and deferred tasks; S102: Construct a computing task model based on the computing task data, calculate the workload of each task; construct a deferred task queue, calculate the deferred workload to be scheduled for execution; determine the total workload based on the real-time task workload and the deferred workload to be scheduled for execution, and generate the data center power requirement based on the total workload. S103: Based on the power demand of the data center, and combined with load relationships, equipment models and energy balance constraints, construct a comprehensive energy system model, and solve for the operating status of the comprehensive energy system with the goal of minimizing the operating cost of the comprehensive energy system. S104: Construct an energy-side coupling coefficient based on the operating status of the integrated energy system and feed it back to the computing task model. Adjust the scheduling sequence of deferred tasks based on the energy-side coupling coefficient and update the power demand of the data center. S105: Repeatedly execute the data center demand power generation, calculation of integrated energy system operation status and energy-side coupling coefficient feedback process until the preset optimization conditions are met, and output the computing task scheduling scheme and integrated energy system operation strategy.

[0026] like Figure 2 As shown, the integrated energy system for the data center in this embodiment includes a power grid, a photovoltaic power generation system, energy storage equipment, air conditioning, a data center, an absorption chiller, etc.

[0027] In step S101, data center computing task data is acquired. This data can be categorized into real-time tasks and deferred tasks. Real-time tasks must be processed within the current scheduling cycle to ensure the required service response; deferred tasks can be processed within an acceptable timeframe and are managed through a task queue. Therefore, deferred tasks can be considered as flexible computing loads oriented towards energy scheduling.

[0028] Step S102 is as follows: S102.1: Construct a computing task model based on the computing task data, and calculate the workload of each task. The workload of each task includes real-time task workload and deferred task workload.

[0029] Data center energy consumption varies with computing tasks. However, traditional methods typically process computing tasks as they arrive, without considering the impact of computing load scheduling on power requirements. Therefore, this embodiment constructs a computing task model based on the TPC-H benchmark and calculates the workload of each task.

[0030] TPC-H contains 22 standard database queries, denoted as... ,in Different queries have different runtimes and computational complexities, thus they can be used to describe heterogeneous database workloads in data centers.

[0031] set up Indicates the first Four TPC-H database sizes are considered in this embodiment: 1GB, 3GB, 5GB, and 10GB. Indicates the first The runtime of a TPC-H query in the baseline environment. This represents the average runtime of all queries. To describe the relative complexity between different queries, a query complexity coefficient is defined as: (1); Among them, the larger ones This indicates that the query has a longer execution time, which also means that the query has higher computational complexity. Based on this coefficient, the... The first type of database size The normalized workload of a query is represented as: (2); in, This represents the workload of a single task. In this way, computational tasks are no longer considered as identical tasks, and their workload depends on both the database size and the query type.

[0032] set up and These represent the real-time task workload that arrives within time period t and must be executed in the current time period, and the deferred task workload with scheduling flexibility, respectively. The real-time task workload is calculated as follows: (3); The workload of newly arrived deferred tasks within time period t is calculated as follows: (4); S102.2: Construct a queue of deferred tasks and calculate the deferred workloads to be scheduled for execution.

[0033] To describe the backlog and scheduling process of deferred tasks, a deferred task queue is constructed. , representing the backlog of deferred workloads that have not yet been processed in the queue at the beginning of time period t.

[0034] The state of the deferred queue changes dynamically with the arrival and processing of new deferred workloads within time interval t. Its dynamic equation can be expressed as: (5); in, This represents the initial backlog queue actually executed within time period t. Delayable task workloads that are called out for execution must satisfy the following constraints: (6); The length of the delayable queue is also constrained by the maximum queue capacity: (7); in, Maximum queue capacity is defined as the maximum number of scheduling periods during which deferred tasks are allowed to be delayed. Maximum processable deferred task workload within a single scheduling period The product of is: .

[0035] In addition, to ensure that there is no historical backlog at the beginning of the scheduling period and that all deferred tasks are completed at the end of the scheduling period, initial queue and terminal queue constraints are set: (8); in, This indicates the duration of the entire task cycle.

[0036] S102.3: Determine the total workload based on the real-time task workload and the scheduled delayed workload, and generate the data center power requirement based on the total workload.

[0037] Time period t The total workload of internal processing consists of real-time task workloads and scheduled, deferred task workloads: (9); Accordingly, the power demand of the data center is modeled as a linear function of the total workload: (10); in, For the basic power consumption of data centers, To address the power factor corresponding to workload, this model demonstrates that task scheduling directly alters the power requirements of a data center. Unlike methods that do not consider data center task flexibility, this embodiment dynamically generates the data center's required power based on task arrival and deferred task scheduling.

[0038] Step S103 is as follows: S103.1: Calculate the data center's cooling load requirements and recyclable heat based on the data center's power requirements.

[0039] Data center power demand is closely related to the heat it generates. During data center operation, a significant portion of the electrical energy consumed by computing devices is converted into heat, thus increasing cooling load demand. Therefore, cooling load demand and recyclable heat are modeled as linear functions of data center power demand: (11);

[0040] (12); in, To meet cooling load requirements, For recoverable heat, This is the cooling load factor. The coefficient of recyclable heat.

[0041] S103.2: Construct an absorption chiller model based on recoverable heat.

[0042] Absorption chillers are used to convert recyclable heat generated in data centers into cooling output. Therefore, the cooling output of an absorption chiller is expressed as: (13); in, This represents the coefficient of performance (COP) of the absorption chiller. Its operating constraints are: (14); in, It is an absorption chiller that can absorb the maximum amount of heat.

[0043] S103.3: Construct an air conditioning cooling output model.

[0044] In this embodiment, the air conditioner specifically adopts a precision air conditioner, also known as a computer room air conditioner or a constant temperature and humidity air conditioner. It is an air conditioning device designed specifically for electronic equipment environments such as data centers and communication equipment rooms that have strict requirements for temperature, humidity, and cleanliness.

[0045] Air conditioning provides cooling to data centers by consuming electrical energy. The relationship between its cooling output and input electrical power is expressed as: (15); in, For air conditioning cooling output, The electrical power consumed by the air conditioner The air conditioning performance coefficient (CPC) is subject to the following constraints on cooling capacity: (16); in, This is the maximum cooling capacity that the air conditioner can output.

[0046] S103.4: Construct the state equation of the energy storage system.

[0047] Energy storage systems can be charged by photovoltaic power or grid power and can discharge to power data center loads. The state-of-charge equation for an energy storage system is expressed as: (17); in, This indicates the state of charge of the energy storage system during time period t; The change in state of charge during time period t: (18); in, and These are the charging power and discharging power of the energy storage system, respectively. For energy storage capacity, and These are the equivalent charging efficiency and equivalent discharging efficiency, respectively. The charging and discharging power constraints of the energy storage system are as follows: (19); (20); in, and These are the maximum charging and discharging power of the energy storage system, respectively.

[0048] To avoid simultaneous charging and discharging of the energy storage system, the following constraints are introduced: (twenty one); The state of charge of an energy storage system is limited by the following ranges: (twenty two); S103.5: Establish electrical load balance constraints based on the power demand of the data center, the power generation of photovoltaic power, and the state of charge of the energy storage system; establish cooling load balance constraints based on the cooling load demand of the data center, the cooling output of the absorption chiller, and the cooling output of the air conditioner.

[0049] S103.5.1: Construct electrical load balance constraints.

[0050] The electrical load of a data center's integrated energy system mainly includes the data center's power demand and air conditioning power consumption, which are supplied jointly by the power grid, photovoltaic power generation, and energy storage systems. The electrical load balance relationship is represented as follows: (twenty three); in, Power supplied to the power grid This refers to the output power of photovoltaics.

[0051] S103.5.2: Constructing cooling load balance constraints.

[0052] The cooling load of the data center's integrated energy system is supplied by both air conditioning and absorption chillers. To avoid unnecessary overcooling, the cooling supply must match the cooling load demand. (twenty four); S103.6: Integrate the load relationship between the data center's power demand, cooling load demand, and recyclable heat, the absorption chiller model, the air conditioning cooling output model, and the energy storage system's state of charge model, as well as the electrical load balance constraints and cooling load balance constraints, to construct a comprehensive energy system model with the data center's power demand as the load input, the equipment's power output as the decision variable, and the equipment operation constraints and energy balance constraints as the constraints.

[0053] S103.7: The objective function for minimizing the operating cost of the integrated energy system is expressed as follows: (25); in, These are the costs of delayed queue backlog, grid purchase costs, energy storage operation costs, absorption chiller operation costs, and curtailment costs. These cost items are defined as follows: (26); (27); (28); (29); (30); in, Indicates the time domain length for scroll optimization. Cost of delayed queue backlog per unit; For the backlog of tasks; Power exchange for the power grid; To absorb heat for the refrigeration unit; This refers to the power of solar power curtailment.

[0054] Using equation (25) as the objective function, the integrated energy system model is solved to obtain the operating state of the integrated energy system. The operating state is the set of optimal output of each energy device and system state variables obtained after solving the integrated energy system model under the given conditions of data center power demand, photovoltaic output prediction, electricity price, initial state of charge of energy storage and equipment operating parameters.

[0055] Step S104 is as follows: S104.1: Construct the energy-side coupling coefficient based on the operating status of the integrated energy system.

[0056] To coordinate computational task scheduling with the operation of the integrated energy system, an energy-side coupling coefficient is constructed. This coefficient characterizes the energy friendliness of different time periods, and its value is related to the operating status of the integrated energy system. A higher coupling coefficient indicates that the current time period is more suitable for handling deferred workloads.

[0057] The energy-side coupling coefficient can be expressed as a function of the operating state of the integrated energy system: (31); in, , This represents a function mapping relationship.

[0058] S104.2: Adjust the scheduling sequence of deferred tasks based on the energy-side coupling coefficient and update the data center's power requirements.

[0059] Energy-side coupling coefficient and data center power requirements This allows for coupling. Real-time workloads are processed immediately, while deferred workloads can be transferred between different time periods.

[0060] Therefore, within the current optimized time domain, the upper-level computing task scheduling model uses the total workload of deferred tasks actually executed in each time period. As an optimization variable, the optimization objective for the currently scheduled deferred task workload is to reduce the cumulative queue backlog of deferred workloads: (32); in, This indicates the time domain length for scroll optimization. When... When the load is large, it prompts the upper-level computing task scheduling model to increase the processing load of historical backlog tasks in the current time period, so as to realize the transfer of deferred tasks to time periods with better energy conditions.

[0061] S105 specifically refers to: Repeat steps S102-S104 until the preset optimization conditions are met, and output the computational task scheduling scheme and the integrated energy system operation strategy. The specific process is as follows: At the current scheduling time t, based on the current computational task data, the status of the deferred task queue, and the operating status of the integrated energy system, the data center power requirement is initialized, and the number of interactive iterations is set to zero. In the k-th interactive iteration, the integrated energy system model is solved based on the current data center power requirement, obtaining the candidate operating status and objective function value of the integrated energy system corresponding to the k-th iteration; the energy-side coupling coefficient is calculated based on the candidate operating status, and the energy-side coupling coefficient is fed back to the computational task model to adjust the scheduling sequence of deferred tasks, obtaining the data center power requirement and candidate operating strategy for the (k+1)-th iteration.

[0062] Calculate the change in the objective function between the k-th iteration and the (k-1)-th iteration. When the change in the objective function is no greater than a preset convergence threshold, the interactive optimization process at the current scheduling moment is considered converged, i.e.: (33); in, Let be the change in the objective function during the k-th iteration. Let this be the change in the objective function during the (k-1)th iteration. This is a minimum value, which can be specifically set by those skilled in the art based on the composition of the corresponding integrated energy system. In this embodiment... Take 10 -3 This means that when the relative rate of change of the objective function value between two adjacent iterations does not exceed 0.1%, the two-level interactive optimization process at the current scheduling moment is considered to have converged.

[0063] The candidate operating strategy obtained in the current iteration is determined as the integrated energy system operating strategy at the current scheduling time; otherwise, the number of interactive iterations is increased, and the integrated energy system optimization and computation task scheduling update continues.

[0064] Upon entering the next scheduling moment, the updated computational task data, the status of the deferred task queue, and the operating status of the integrated energy system are obtained, and the above interactive optimization process is re-executed.

[0065] Example 2 This embodiment provides a data center energy system collaborative optimization system that considers task flexibility, including: Data acquisition module, upper-layer computing task scheduling module, lower-layer integrated energy optimization module; The data acquisition module is used to acquire data center computing task data, which includes real-time tasks and deferred tasks. The upper-layer computing task scheduling module is used to construct a computing task model based on the computing task data, calculate the workload of each task; construct a queue of deferred tasks, calculate the deferred workload to be scheduled for execution; determine the total workload based on the real-time task workload and the deferred workload to be scheduled for execution, and generate the data center power requirement based on the total workload; receive the energy-side coupling coefficient fed back by the lower-layer integrated energy optimization module, adjust the scheduling sequence of deferred tasks based on the energy-side coupling coefficient, update the data center power requirement, and output a computing task scheduling scheme. The lower-level integrated energy optimization module receives the power demand of the data center, constructs an integrated energy system model based on the data center's power demand, combined with load relationships, equipment models, and energy balance constraints, and solves for the integrated energy system's operating status with the goal of minimizing the integrated energy system's operating cost. It then constructs an energy-side coupling coefficient based on the integrated energy system's operating status and feeds this coefficient back to the upper-level computing task scheduling module. Finally, it re-solves for the integrated energy system's operating status and energy-side coupling coefficient based on the updated data center power demand, until preset optimization conditions are met, and outputs the integrated energy system's operating strategy.

[0066] This invention establishes a computing task model encompassing real-time and deferred tasks based on the TPC-H benchmark in the upper-layer computing task scheduling module. It describes the backlog and flexible scheduling process of deferred tasks through a deferred task queue, aiming to reduce deferred task queue backlog and generate a data center computing task scheduling strategy. In the lower-layer integrated energy optimization module, given the data center computing tasks, the integrated energy system formulates an energy equipment operation optimization strategy with the operational economy of the integrated energy system as the optimization objective. Simultaneously, it constructs an energy-side coupling coefficient based on the system operating status and feeds it back to the upper-layer computing task scheduling module, guiding deferred tasks to shift to time periods with better energy conditions. Through this two-layer interactive collaboration, the invention ultimately achieves coordinated matching between computing task scheduling and integrated energy system operation, improving the operational economy and energy utilization efficiency of the data center's integrated energy system.

[0067] The descriptions of each embodiment in the above embodiments have different focuses. For parts not described in detail in a certain embodiment, please refer to the relevant descriptions in other embodiments.

[0068] The proposed system can be implemented in other ways. For example, the system embodiments described above are merely illustrative, and the division of modules described above is only a logical functional division. In actual implementation, there may be other division methods. For example, multiple modules may be combined or integrated into another system, or some features may be ignored or not executed.

[0069] Example 3 This embodiment also provides an electronic device, including: one or more processors, one or more memories, and one or more computer programs; wherein, the processor is connected to the memory, and the one or more computer programs are stored in the memory. When the electronic device is running, the processor executes the one or more computer programs stored in the memory to cause the electronic device to perform the method described in Embodiment 1.

[0070] It should be understood that in this embodiment, the processor can be a central processing unit (CPU), or it can be other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. The general-purpose processor can be a microprocessor or any conventional processor, etc.

[0071] Memory may include read-only memory and random access memory, and provides instructions and data to the processor. A portion of memory may also include non-volatile random access memory. For example, memory may also store information about the device type.

[0072] In the implementation process, each step of the above method can be completed by the integrated logic circuits in the processor hardware or by software instructions.

[0073] The method in Embodiment 1 can be directly implemented by a hardware processor, or implemented by a combination of hardware and software modules within the processor. The software modules can reside in readily available storage media in the art, such as random access memory, flash memory, read-only memory, programmable read-only memory, electrically erasable programmable memory, or registers. This storage medium is located in memory; the processor reads information from the memory and, in conjunction with its hardware, completes the steps of the above method. To avoid repetition, a detailed description is not provided here.

[0074] Those skilled in the art will recognize that the units and algorithm steps described in connection with the various examples of this embodiment can be implemented in electronic hardware or a combination of computer software and electronic hardware. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this invention.

[0075] Example 4 Embodiment 4 of the present invention provides a computer-readable storage medium.

[0076] A computer-readable storage medium having a program stored thereon that, when executed by a processor, implements the steps of the method as described in Embodiment 1 of the present invention.

[0077] The detailed steps are the same as those provided in Example 1, and will not be repeated here.

[0078] Example 5 Embodiment 5 of the present invention provides a computer program product.

[0079] A computer program product includes software code, wherein the program in the software code performs the steps described in Embodiment 1 of the present invention.

[0080] The above description is merely a preferred embodiment of the present invention and is not intended to limit the invention. Various modifications and variations can be made to the present invention by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.

Claims

1. A collaborative optimization method for data center energy systems considering task flexibility, characterized in that, include: Acquire data center computing task data, which includes real-time tasks and deferred tasks; A computing task model is constructed based on the computing task data, and the workload of each task is calculated; a deferred task queue is constructed, and the deferred workload to be scheduled for execution is calculated; the total workload is determined based on the real-time task workload and the deferred workload to be scheduled for execution, and the data center power demand is generated based on the total workload. Based on the power demand of the data center, a comprehensive energy system model is constructed by combining load relationships, equipment models and energy balance constraints. With the goal of minimizing the operating cost of the comprehensive energy system, the operating status of the comprehensive energy system is obtained by solving the problem. An energy-side coupling coefficient is constructed based on the operating status of the integrated energy system and fed back to the computing task model. The scheduling sequence of deferred tasks is adjusted based on the energy-side coupling coefficient, and the power demand of the data center is updated. Repeatedly execute the process of generating data center power demand, calculating the operating status of the integrated energy system, and feeding back the energy-side coupling coefficient until the preset optimization conditions are met, and output the computing task scheduling scheme and the integrated energy system operation strategy.

2. The data center energy system collaborative optimization method considering task flexibility as described in claim 1, characterized in that, The step of constructing a computing task model based on the computing task data and calculating the workload of each task specifically includes: constructing a computing task model based on the TPC-H benchmark test, and determining the query complexity coefficient of each query by comparing the running time of each query in the benchmark environment with the average running time. Based on the query complexity coefficient and the database size, each query is normalized to obtain the normalized workload of each query under different database sizes. Based on the real-time tasks and deferred tasks arriving within the time period, and according to the query type and database size corresponding to each task, the workload of real-time tasks and the workload of deferred tasks are calculated by summing them up.

3. The data center energy system collaborative optimization method considering task flexibility as described in claim 1, characterized in that, The construction of a comprehensive energy system model based on the power demand of the data center, combined with load relationships, energy equipment models, and electrical and cooling energy balance constraints, specifically includes: The cooling load requirement and recoverable heat of the data center are determined based on the power demand of the data center. Based on the recoverable heat, an absorption chiller model is constructed, along with an air conditioning cooling output model and an energy storage system state-of-charge model. Based on the power demand of the data center, the photovoltaic power generation and the state of charge of the energy storage system, an electrical load balance constraint is constructed, and based on the cooling load demand of the data center, the cooling output of the absorption chiller and the cooling output of the air conditioner, a cooling load balance constraint is constructed. A comprehensive energy system model is constructed based on the data center's cooling load demand, recyclable heat, absorption chiller model, air conditioning cooling output model, energy storage system state of charge model, and electrical load balance constraints and cooling load balance constraints.

4. The data center energy system collaborative optimization method considering task flexibility as described in claim 1, characterized in that, The goal of minimizing the operating cost of the integrated energy system is to solve for the operating state of the integrated energy system. Specifically, the integrated energy system model is solved with the goal of minimizing the sum of the queue backlog cost of deferred tasks, grid power purchase cost, energy storage operation cost, absorption chiller operation cost, and curtailment cost, so as to obtain the optimal output of each energy device and the system state variables.

5. The data center energy system collaborative optimization method considering task flexibility as described in claim 1, characterized in that, The energy-side coupling coefficient is constructed based on the operating status of the integrated energy system, specifically as follows: ； in, For energy-side coupling coefficient, This is a function mapping relationship. Power supplied to the power grid For photovoltaic output power, Indicates the state of charge of the energy storage system. The electrical power consumed by the air conditioner.

6. The data center energy system collaborative optimization method considering task flexibility as described in claim 1, characterized in that, The process of repeatedly executing data center power demand generation, calculating the integrated energy system operating status, and feedback the energy-side coupling coefficient until the preset optimization conditions are met specifically includes: Based on the current power demand of the data center, the integrated energy system model is solved to obtain the corresponding candidate operating states and objective function values ​​of the integrated energy system. The energy-side coupling coefficient is calculated based on the candidate operating states of the integrated energy system and fed back to the computing task model to adjust the scheduling sequence of deferred tasks, thereby obtaining the updated data center power demand. The system determines whether the preset convergence threshold is met based on the change in the objective function between two consecutive interactive iterations. If it is met, the system outputs the computational task scheduling scheme and integrated energy system operation strategy corresponding to the current iteration. If it is not met, the system continues interactive iteration based on the updated data center power demand.

7. A collaborative optimization system for data center energy systems that considers task flexibility, characterized in that, include: Data acquisition module, upper-layer computing task scheduling module, lower-layer integrated energy optimization module; The data acquisition module is used to acquire data center computing task data, which includes real-time tasks and deferred tasks. The upper-layer computing task scheduling module is used to construct a computing task model based on the computing task data, calculate the workload of each task; construct a queue of deferred tasks, calculate the deferred workload to be scheduled for execution; determine the total workload based on the real-time task workload and the deferred workload to be scheduled for execution, and generate the data center power requirement based on the total workload; receive the energy-side coupling coefficient fed back by the lower-layer integrated energy optimization module, adjust the scheduling sequence of deferred tasks based on the energy-side coupling coefficient, update the data center power requirement, and output a computing task scheduling scheme. The lower-level integrated energy optimization module receives the power demand of the data center, constructs an integrated energy system model based on the data center's power demand, combined with load relationships, equipment models, and energy balance constraints, and solves for the integrated energy system's operating status with the goal of minimizing the integrated energy system's operating cost. It then constructs an energy-side coupling coefficient based on the integrated energy system's operating status and feeds this coefficient back to the upper-level computing task scheduling module. Finally, it re-solves for the integrated energy system's operating status and energy-side coupling coefficient based on the updated data center power demand, until preset optimization conditions are met, and outputs the integrated energy system's operating strategy.

8. An electronic device, characterized in that, It includes a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor, when executing the program, implements the steps of the data center energy system co-optimization method considering task flexibility as described in any one of claims 1-6.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that, When executed by the processor, the program implements the steps in the collaborative optimization method for data center energy systems that takes into account task flexibility as described in any one of claims 1-6.

10. A computer program product, comprising software code, characterized in that, The program in the software code performs the steps in the data center energy system collaborative optimization method considering task flexibility as described in any one of claims 1-6.

Images