Search

CN-122019190-A - Data center micro-grid calculation-electricity-heat cooperative optimization scheduling method and system

CN122019190ACN 122019190 ACN122019190 ACN 122019190ACN-122019190-A

Abstract

The invention provides a data center micro-grid calculation-electricity-heat cooperative optimization scheduling method and system, and relates to the technical field of data center energy scheduling. The method comprises the steps of dividing loads into undelayable workloads and delayable workloads according to operation data of a data center micro-grid, constructing environment quality indexes, and designing a dynamic load scheduling mechanism according to the environment quality indexes, wherein the dynamic load scheduling mechanism comprises the steps of formulating scheduling rules and designing scheduling reward and punishment rules for the delayable workloads, constructing a multi-objective collaborative optimization scheduling model representing a calculation-electricity-heat strong coupling relation based on the dynamic load scheduling mechanism, and solving the model to obtain a workload and energy scheduling strategy. The invention constructs an environment-aware work load scheduling mechanism, can dynamically adjust the distribution of the deferrable work load according to the environmental quality index, realizes the matching of calculation power and the energy running state of the micro-grid, greatly improves the green electricity consumption and reduces the energy consumption cost.

Inventors

  • ZHOU KAILE
  • CHU YIBO
  • YANG ZIWEI
  • LU XINHUI

Assignees

  • 合肥工业大学

Dates

Publication Date
20260512
Application Date
20260410

Claims (10)

  1. 1. The utility model provides a data center microgrid calculation-electricity-heat cooperative optimization scheduling method which is characterized by comprising the following steps: acquiring operation data of a micro-grid of a data center; Dividing the load into a non-deferrable work load and a deferrable work load according to the operation data, constructing an environment quality index, and designing a dynamic load scheduling mechanism according to the environment quality index, wherein the dynamic load scheduling mechanism comprises a scheduling rule formulated for the deferrable work load and a scheduling reward and punishment rule designed for the deferrable work load; based on a dynamic load scheduling mechanism, constructing a multi-objective collaborative optimization scheduling model representing a calculation-electric-thermal strong coupling relation; And solving the multi-target collaborative optimization scheduling model to obtain a workload and energy scheduling strategy.
  2. 2. The data center micro-grid calculation-electricity-heat cooperative optimization scheduling method according to claim 1, wherein the environmental quality index is set by comprehensively considering electricity price and availability of renewable energy sources, and specifically comprises the following steps: Wherein, the The electricity price at the time t is indicated, Representing the output of the renewable energy source, The highest electricity price is indicated by the utility model, Indicating the highest output of renewable energy sources, And Coefficients respectively representing the environmental quality index, A value representing an environmental quality indicator.
  3. 3. The data center microgrid calculation-electric-thermal collaborative optimization scheduling method according to claim 1, wherein the scheduling rules comprise: releasing the deferrable workload from the set of untreated deferrable workloads requires that the following two conditions be met: Wherein, the The environmental quality index is indicated to be the same, Representing an environmental quality release threshold value, Represent the first The arrival time of the workload of the group task group, Indicating the maximum delay time of the workload, The task group index is represented as a function of the task group index, Represent the first The number of workloads of the group task group, Expressed in time A time-delay in the work load, Indicating the amount of delay work to be performed eventually, Representing the set of deferrable task groups that are actually performed at time t.
  4. 4. The data center microgrid calculation-electric-thermal collaborative optimization scheduling method according to claim 1, wherein the scheduling reward and punishment rules comprise: Wherein, the Indicating a net delay prize associated with the deferrable workload at time t; representing for a task group If it transfers the task from the delay queue to the execution state at time t; representing for a task group If the workload delay is not timely processed at time t, punishment is received; a set of deferrable task groups that represent scheduled processing that has been released at time t; Representing a set of deferrable task groups that remain in a waiting state, i.e., unprocessed, at time t; Representing a task group index; Represent the first The number of workloads of the group task groups; is a device for measuring load slave Delay to Then, an evaluation function of the comprehensive improvement degree in the aspects of electricity price and renewable energy utilization is provided; representing a base delay penalty rate; Representing penalty coefficients associated with the environment; representing a maximum power generation price over the entire period of time; Indicating electricity price at time t; Representing the output of the renewable energy source, Representing the highest renewable energy output.
  5. 5. The data center micro-grid calculation-electric-thermal collaborative optimization scheduling method according to any one of claims 1-4, wherein the objective function of the multi-objective collaborative optimization scheduling model comprises: Wherein, the Representing the operating cost; A net delay reward associated with the deferrable workload; representing a renewable energy utilization index; Total thermal safety risk; 、 、 、 Representing the weight parameters; Representing a minimized objective function.
  6. 6. The data center micro-grid calculation-electric-thermal collaborative optimization scheduling method according to any one of claims 1-4, wherein the solving the multi-objective collaborative optimization scheduling model comprises: and solving the multi-target collaborative optimization scheduling model by adopting a large model-assisted deep reinforcement learning model.
  7. 7. The data center micro-grid calculation-electric-thermal collaborative optimization scheduling method according to any one of claims 1-4, wherein the data center micro-grid calculation-electric-thermal collaborative optimization scheduling method further comprises: the work load and the energy scheduling strategy are issued to each system device of the data center micro-grid; and performing three-level verification of basic constraint, core safety and optimization targets, and performing automatic fine adjustment and emergency adjustment closed-loop treatment on the abnormality found by the verification.
  8. 8. A data center microgrid calculation-electric-thermal collaborative optimization scheduling system, comprising: the data acquisition module is used for acquiring the operation data of the data center micro-grid; The system comprises a work load scheduling mechanism construction module, a dynamic load scheduling mechanism, a control module and a control module, wherein the work load scheduling mechanism construction module is used for dividing a load into an undelayable work load and a deferrable work load according to operation data, constructing an environment quality index and designing the dynamic load scheduling mechanism according to the environment quality index; the multi-objective optimization model construction module is used for constructing a multi-objective collaborative optimization scheduling model representing a calculation-electric-thermal strong coupling relation based on a dynamic load scheduling mechanism; and the model solving module is used for solving the multi-target collaborative optimization scheduling model to obtain a workload and energy scheduling strategy.
  9. 9. A computer-readable storage medium storing a computer program for data center micro grid co-operation-electro-thermal co-operation optimization scheduling, wherein the computer program causes a computer to execute the data center micro grid co-operation-optimization scheduling method according to any one of claims 1 to 7.
  10. 10. An electronic device, comprising: One or more processors, a memory, and one or more programs, wherein the one or more programs are stored in the memory and configured to be executed by the one or more processors, the programs comprising instructions for performing the data center microgrid co-optimal scheduling method of any of claims 1-7.

Description

Data center micro-grid calculation-electricity-heat cooperative optimization scheduling method and system Technical Field The invention relates to the technical field of data center energy scheduling, in particular to a data center micro-grid calculation-electricity-heat cooperative optimization scheduling method and system. Background As an important infrastructure supporting digital economy operation, the workload of data centers continues to increase, driving their electricity demand to continuously climb, directly resulting in increased operating costs. Meanwhile, the data center also faces multiple challenges such as high power supply reliability requirement, high energy consumption intensity, and strict carbon emission constraint in the operation process. In order to effectively solve the problems, the data center micro-grid integrating the renewable energy sources, the energy storage system and the energy conversion equipment is gradually popularized and applied. The micro-grid can obviously improve the power supply guarantee capability of the data center by configuring a local power supply and flexibly adjusting resources, reduce the external electricity purchasing cost, and reduce carbon emission in the operation process to a certain extent, so that the micro-grid becomes an important support for green and efficient operation of the data center. The related research is mainly carried out from two dimensions of modeling optimization and workload scheduling of the micro-grid of the data center in the prior art. In the aspect of micro-grid modeling optimization, the method is mainly divided into two types, namely, one type of modeling is used for describing the structural description of a stress system and the multi-energy coupling relation, the data center is used for modeling an energy hub so as to support electric heating cooperative operation, and the other type of modeling is used for describing the coupling characteristic between an electricity utilization process and a heat generation process. In terms of workload scheduling, most of the prior art models workload scheduling and micro-grid energy management respectively, part of methods try to migrate the deferrable workload in different spaces or different time periods so as to match the local renewable energy output and electricity price level, and meanwhile, deep reinforcement learning technology has been widely applied to data center operation optimization for solving the scheduling problems under the conditions of multi-energy coupling, renewable energy output uncertainty and random workload. Furthermore, the existing optimization techniques mostly aim at minimizing energy consumption or minimizing operating costs. However, prior art solutions lack the ability to dynamically schedule data center workloads, and it is difficult to dynamically adjust the distribution of deferrable workloads according to environmental conditions. Disclosure of Invention (One) solving the technical problems Aiming at the defects of the prior art, the invention provides a data center micro-grid calculation-electricity-heat cooperative optimization scheduling method and a system, which solve the technical problem that the prior art scheme lacks the capability of dynamic scheduling of the workload of a data center. (II) technical scheme In order to achieve the above purpose, the invention is realized by the following technical scheme: In a first aspect, the invention provides a data center micro-grid calculation-electricity-heat collaborative optimization scheduling method, which comprises the following steps: acquiring operation data of a micro-grid of a data center; Dividing the load into a non-deferrable work load and a deferrable work load according to the operation data, constructing an environment quality index, and designing a dynamic load scheduling mechanism according to the environment quality index, wherein the dynamic load scheduling mechanism comprises a scheduling rule formulated for the deferrable work load and a scheduling reward and punishment rule designed for the deferrable work load; based on a dynamic load scheduling mechanism, constructing a multi-objective collaborative optimization scheduling model representing a calculation-electric-thermal strong coupling relation; And solving the multi-target collaborative optimization scheduling model to obtain a workload and energy scheduling strategy. Preferably, the environmental quality index is set by comprehensively considering electricity price and availability of renewable energy sources, and specifically includes: Wherein, the The electricity price at the time t is indicated,Representing the output of the renewable energy source,The highest electricity price is indicated by the utility model,Indicating the highest output of renewable energy sources,AndCoefficients respectively representing the environmental quality index,A value representing an environmental quality indicator. Preferably, the s