Economical low-carbon operation optimization method for data center in power distribution system

Abstract

The invention discloses an economic low-carbon operation optimization method for a data center in a power distribution system, and the method comprises the steps: building a power distribution system operation optimization model, solving the power distribution system operation optimization model, and obtaining a power flow result of the power distribution system and the marginal electricity price of a power distribution node; solving node carbon intensity of the power distribution system based on a power flow result of the power distribution system and a carbon emission flow theory; the power distribution node marginal electricity price of the power distribution system and the node carbon intensity of the power distribution system are sent to the data center, and the data center establishes a data center economic low-carbon operation optimization model with the purpose of minimizing the comprehensive cost, solves and obtains an electricity purchasing and selling result between the data center and the power distribution system, and sends the electricity purchasing and selling result to the power distribution system. Based on the marginal electricity price and the carbon intensity of the power distribution node, space-time scheduling is carried out on the workload, and the comprehensive operation cost of the data center is reduced. And moreover, the power distribution system operator and the data center operator only need to interact the result of the boundary variable, so that the data privacy can be effectively protected.

Description

Technical Field

[0001] The present invention belongs to the technical field of low-carbon economic operation scheduling of data centers, and in particular relates to an economical low-carbon operation optimization method for data centers in a power distribution system. Background Art

[0002] The goal of optimizing the economic operation of the distribution system is to improve the efficiency of power grid operation and reduce system costs by reasonably allocating resources and optimizing the operation mode. In this process, the node marginal electricity price, as a key indicator, reflects the temporal and spatial characteristics of the electricity price of the distribution system. The node marginal electricity price can not only accurately depict the marginal operating costs of different nodes, but also guide distributed energy and loads to reasonably participate in market transactions and improve resource utilization efficiency. At the same time, it provides a scientific basis for the optimal dispatch of regional energy, which helps to improve the economy, flexibility and sustainability of the power grid.

[0003] As a tool to quantify carbon emissions from loads, carbon emission flow can directly reflect the impact of energy consumption on the environment and has important application value. Carbon emission flow can evaluate the carbon emission level of loads at different times and locations, and provide data support for identifying high emission loads and optimizing energy consumption structure. In addition, through the analysis of carbon emission flow, users can be guided to give priority to the use of low-carbon or renewable energy to reduce carbon emissions, and provide a basis for the formulation of a pricing mechanism based on carbon emissions, thus promoting the realization of low-carbonization goals.

[0004] In the power distribution network, data centers have scheduling flexibility in time and space due to their workloads, and can reduce energy costs and carbon emission costs by optimizing scheduling. When formulating the operation plan of the data center, it is necessary to fully consider the interaction with the power distribution system. In addition, since data center operators and power distribution system operators are different entities, data privacy protection is also an issue that must be addressed.

[0005] In summary, there is an urgent need to propose an optimization method for economical and low-carbon operation of data centers in power distribution systems. Summary of the invention

[0006] In order to solve the above technical problems, the present invention proposes an optimization method for economical and low-carbon operation of a data center in a power distribution system to solve the problems existing in the above-mentioned prior art.

[0007] To achieve the above object, the present invention provides a method for optimizing economic and low-carbon operation of a data center in a power distribution system, which is characterized by comprising the following steps:

[0008] With the goal of minimizing the energy economic cost of the distribution system, the DIST-flow power flow constraint is adopted to establish the distribution system operation optimization model;

[0009] Solving the distribution system operation optimization model to obtain the power flow results of the distribution system and the marginal electricity price of the distribution nodes;

[0010] Based on the power flow results of the distribution system and the carbon emission flow theory, the carbon intensity of the distribution system nodes is solved;

[0011] The marginal electricity price of the distribution nodes of the distribution system and the carbon intensity of the distribution system nodes are sent to the data center. The data center aims to minimize the comprehensive cost and establish an economic low-carbon operation optimization model for the data center.

[0012] Solving the economic and low-carbon operation optimization model of the data center, obtaining the results of electricity purchase and sale between the data center and the power distribution system, and sending them to the power distribution system;

[0013] The power distribution system and the data center iteratively solve the corresponding optimization models in turn until the interactive power of the two converges, completing the economic and low-carbon operation optimization of the data center.

[0014] Optionally, the objective function of the power distribution system operation optimization model is as shown in the following formula:

[0015]

[0016] In the formula, c g,t It represents the wholesale market electricity price in the upper power grid; P represents the electricity price in the distribution system; g,t Indicates the power purchased and sold by the distribution system in the wholesale market; represents the electricity purchased from distributed generators; Indicates the power purchased from the data center.

[0017] Optionally, a DIST-flow power flow constraint is as follows:

[0018]

[0019]

[0020] Where P ij,t and Q ij,t Respectively represent the active and reactive power on branch ij; and They represent the active and reactive power injected into node j respectively; l ij,t represents the current on branch ij; u j,t represents the voltage amplitude of node j; r ij and x ij They represent the resistance and reactance of line ij respectively; and They represent the connection relationship between the power distribution system node j and the distributed generator and the data center respectively; and denote the active and reactive loads of node j respectively; and Respectively represent the upper and lower limits of the voltage amplitude at node j; λ j,t represents the dual variable of the corresponding constraint; P jk,t and Q jk,t denote the active and reactive power on line jk respectively, is a binary parameter, indicating the connection status between node j and the upper power grid, P g,t Represents the interactive active power with the upper grid, P dg,t Represents the output active power of distributed generators, P d,t Indicates the active load power of the data center, Represents the output reactive power of distributed generators, Q g,t Represents the interactive reactive power with the upper grid, Q d,t Indicates the reactive load power of the data center.

[0021] Optionally, the carbon intensity calculation formula of energy on the distribution system node i and line ij is as follows:

[0022]

[0023] In the formula, ρ ij and ρ i are the carbon intensity of line ij and node i respectively; is the energy flowing from node i to line ij; is the energy loss generated on line ij; is the energy flowing from line ij to node j; and They are and The corresponding carbon emission rate.

[0024] Optionally, the process of solving the carbon intensity of the distribution system nodes also includes: calculating the carbon emission flow in the distribution system based on the law of conservation of carbon emissions, the principle of proportional allocation, and the principle of energy merging.

[0025] Optionally, the objective function of the data center economic low-carbon operation optimization model is as shown in the following formula:

[0026]

[0027] In the formula, and Respectively represent the power purchase and power sales of the data center; λ d,t and Respectively represent the purchase price and sale price of electricity of the data center; ce is the carbon emission cost coefficient; Represents the carbon intensity of a data center node in the power distribution system.

[0028] Optionally, the constraints of the data center operation optimization model include: workload operation constraints, server startup constraints, operation constraints of energy storage in UPS, and data center power constraints.

[0029] Optionally, in the process of iteratively solving the corresponding optimization models of the power distribution system and the data center in sequence, an iterative solution algorithm based on the dichotomy method is adopted.

[0030] The present invention also provides a computer device, comprising a memory, a processor and a computer program stored in the memory, wherein the processor executes the computer program to implement the steps of the method.

[0031] The present invention also provides a computer-readable storage medium on which a computer program is stored. When the computer program is executed by a processor, the steps of the method are implemented.

[0032] Compared with the prior art, the present invention has the following advantages and technical effects:

[0033] The present invention proposes an economic and low-carbon operation optimization method for a data center in a power distribution system. First, an operation optimization model of a power distribution system is established, and the power flow results of the power distribution system and the marginal electricity price of the power distribution nodes are obtained by solving the problems and sending them to the data center. According to the power flow results of the power distribution system and based on the carbon emission flow theory, the carbon intensity of the power distribution system nodes is calculated and sent to the data center. With the goal of minimizing the comprehensive cost, the data center establishes and solves the economic and low-carbon operation optimization model of the data center, obtains the results of power purchase and sale, and sends them to the power distribution system.

[0034] The present invention performs spatiotemporal scheduling of workloads based on marginal electricity prices and carbon intensity of distribution nodes, thereby reducing the comprehensive operating costs of data centers. Moreover, the distribution system operator and the data center operator only need to interact with the results of boundary variables, which can effectively protect data privacy. BRIEF DESCRIPTION OF THE DRAWINGS

[0035] The drawings constituting a part of the present application are used to provide a further understanding of the present application. The illustrative embodiments and descriptions of the present application are used to explain the present application and do not constitute an improper limitation on the present application. In the drawings:

[0036] Figure 1 A topological structure diagram of a data center and an IEEE33-node power distribution system according to an embodiment of the present invention;

[0037] Figure 2 This is a framework diagram of the interaction between the power distribution system and the data center according to an embodiment of the present invention;

[0038] Figure 3 Schematic diagram of the convergence process of iterative solution based on the bisection method in an embodiment of the present invention, wherein (a) is a schematic diagram of the convergence process of iterative solution based on the bisection method in data center 2, and (b) is a schematic diagram of the convergence process of iterative solution based on the bisection method in data center 3;

[0039] Figure 4 This is a schematic diagram of marginal electricity price results of distribution nodes according to an embodiment of the present invention;

[0040] Figure 5 A schematic diagram of the carbon intensity of nodes in a power distribution system according to an embodiment of the present invention;

[0041] Figure 6 A schematic diagram of spatial scheduling results of a delay-sensitive workload according to an embodiment of the present invention;

[0042] Figure 7 It is a schematic diagram of time scheduling results of a delay-tolerant workload according to an embodiment of the present invention. DETAILED DESCRIPTION

[0043] It should be noted that, in the absence of conflict, the embodiments and features in the embodiments of the present application can be combined with each other. The present application will be described in detail below with reference to the accompanying drawings and in combination with the embodiments.

[0044] It should be noted that the steps shown in the flowcharts of the accompanying drawings can be executed in a computer system such as a set of computer executable instructions, and that, although a logical order is shown in the flowcharts, in some cases, the steps shown or described can be executed in an order different from that shown here.

[0045] Embodiment 1

[0046] This embodiment provides an economical and low-carbon operation optimization method for a data center in a power distribution system. In this method, the power distribution system optimizes its operation with the goal of minimizing the energy economic cost, and obtains the marginal electricity price of the distribution node as the electricity price of the load in the distribution system. According to the power flow results of the distribution system and based on the carbon emission flow theory, the carbon intensity of the distribution system nodes is calculated to quantify the carbon emissions of the load in the distribution system. At the same time, based on the received marginal electricity price of the distribution node and the carbon intensity of the node, the data center formulates an operation plan, performs spatiotemporal scheduling of the workload, and sends the results of power purchase and sale to the distribution system. In order to solve the oscillation problem that may occur between the distribution system and the data center during the iterative solution process, an iterative solution method based on the dichotomy method is proposed.

[0047] To achieve the above purpose, if Figure 1As shown, this embodiment adopts a power distribution system of three data centers and IEEE33 nodes, and constructs an interactive framework between the power distribution system and the data center, such as Figure 2 As shown in the figure. In this framework, the power distribution system and the data center formulate their operation plans in turn and interact with the results of boundary variables. The specific steps include:

[0048] This embodiment aims to minimize the energy economic cost of the distribution system, adopts DIST-flow power flow constraints, and establishes a distribution system operation optimization model;

[0049] Solving the distribution system operation optimization model to obtain the power flow results of the distribution system and the marginal electricity price of the distribution nodes;

[0050] Based on the power flow results of the distribution system and the carbon emission flow theory, the carbon intensity of the distribution system nodes is solved;

[0051] The marginal electricity price of the distribution nodes of the distribution system and the carbon intensity of the distribution system nodes are sent to the data center. The data center aims to minimize the comprehensive cost and establish an economic low-carbon operation optimization model for the data center.

[0052] Solving the economic and low-carbon operation optimization model of the data center, obtaining the results of electricity purchase and sale between the data center and the power distribution system, and sending them to the power distribution system;

[0053] The distribution system and the data center iteratively solve the corresponding optimization model in turn, perform time-space scheduling on the workload, solve their own economic and low-carbon operation plan, obtain the results of power purchase and sale between the distribution system, and send them to the distribution system. The distribution system and the data center iteratively formulate their operation plans in turn until the interactive power converges.

[0054] As a specific implementation method, with the goal of minimizing the energy economic cost of the distribution system, the DIST-flow flow constraint is adopted to establish a distribution system operation optimization model, and the flow results of the distribution system and the marginal electricity price of the distribution node are obtained by solving the process including:

[0055] First, the distribution network aims to minimize the energy economic cost and establishes an objective function, which includes the cost of purchasing and selling electricity from the upper power grid, the cost of purchasing electricity from local distributed generators, and the cost of purchasing electricity from data centers. The objective function of the distribution system operation optimization model is shown in the following formula:

[0056]

[0057] In the formula, c g,t It represents the wholesale market electricity price in the upper power grid; P represents the electricity price in the distribution system; g,t The power purchased and sold by the distribution system in the wholesale market; Electricity purchased from distributed generators; Indicates the power purchased from the data center.

[0058] The Dist-flow model is used to describe the power flow constraints of the distribution system, as shown below:

[0059]

[0060] Where P ij,t and Q ij,t Respectively represent the active and reactive power on branch ij; and They represent the active and reactive power injected into node j respectively; l ij,t represents the current on branch ij; u j,t represents the voltage amplitude of node j; r ij and x ij They represent the resistance and reactance of line ij respectively; and are binary parameters, representing the connection relationship between the distribution system node j and the distributed generators and data centers respectively; and denote the active and reactive loads of node j respectively; and Respectively represent the upper and lower limits of the voltage amplitude at node j; λ j,t represents the dual variable of the corresponding constraint; P jk,t and Q jk,t denote the active and reactive power on line jk respectively, is a binary parameter, indicating the connection status between node j and the upper power grid, P g,t Represents the interactive active power with the upper grid, P dg,t Represents the output active power of distributed generators, P d,t Indicates the active load power of the data center, Represents the output reactive power of distributed generators, Q g,t Represents the interactive reactive power with the upper grid, Q d,t Indicates the reactive load power of the data center.

[0061] The distribution system operation optimization model established above is a second-order cone programming. By solving the distribution system operation optimization model above, the system flow results and the dual variable λ are obtained. j,t The result is sent to the data center as the marginal electricity price of the distribution node.

[0062] As a specific implementation method, according to the power flow results of the distribution system and based on the carbon emission flow theory, a distribution system node carbon intensity calculation model is established. The calculation process of the distribution system node carbon intensity includes:

[0063] The carbon emission rate (denoted by R, unit tonCO2 / h) represents the carbon emission rate corresponding to the energy flow (denoted by E, unit MW). The energy flowing from node i to line ij is The energy loss on line ij is The energy flowing from line ij to node j is and The corresponding carbon emission rates are and Then on line ij, the following equation needs to be satisfied:

[0064]

[0065] Carbon intensity (denoted as ρ, unit tonCO2 / MWh) represents the carbon emissions corresponding to the unit energy. Obviously, the relationship between carbon emission rate and carbon intensity is: R = ρE. Therefore, the carbon intensity of energy on node i and line ij is calculated using the following formula:

[0066]

[0067] In the formula, ρ ij and ρ i are the carbon intensity of line ij and node i respectively.

[0068] The calculation of carbon emission flows in the distribution system is based on the following three principles:

[0069] (1) The law of conservation of carbon emissions: This states that the total carbon inflow to each node in the energy network is equal to the total carbon outflow, as shown in the following formula:

[0070]

[0071] In the formula, R s and R d They represent the energy production E at node j respectively. s and demand E d The corresponding carbon emission rate is Represents the carbon emission rate flowing into line jk.

[0072] (2) Proportional distribution principle: The distribution between carbon outflow is proportional to the energy outflow, as shown in the following formula:

[0073]

[0074] In the formula, represents the energy flowing into line jk.

[0075] (3) Energy merging principle: When energy from different branches or producers is injected into a node, the carbon emissions corresponding to the energy will be merged. Therefore, the carbon emission intensity of node j is calculated using the following formula:

[0076]

[0077] In the formula, X=1 means that the carbon emissions corresponding to energy transmission losses are allocated to energy consumers.

[0078] After calculating the carbon intensity of nodes in the distribution system, it is sent to the data center.

[0079] As a specific implementation method, based on the marginal electricity price and carbon intensity of the distribution nodes of the distribution system, the data center operator establishes an economic and low-carbon operation optimization model for the data center with the goal of minimizing its comprehensive cost, solves the operation plan of the data center, and the workload scheduling strategy, determines the power purchase and sale between it and the distribution system, and sends it to the distribution system operator.

[0080] The objective function of the data center economic low-carbon operation optimization model is as follows:

[0081]

[0082] In the formula, and Respectively represent the power purchase and power sales of the data center; λ d,t and Respectively represent the purchase price and selling price of electricity of the data center; c e is the carbon emission cost coefficient; Represents the carbon intensity of data center nodes in the power distribution system.

[0083] The data center operation optimization model includes the following constraints:

[0084] (1) Workload operation constraints:

[0085]

[0086]

[0087] In the formula, and They represent the flexible latency-sensitive workload and the fixed latency-sensitive workload of the front-end server, respectively; and denote the upper and lower bounds of flexible latency-sensitive workloads assigned to data centers, respectively; represents the number of qth delay-tolerant workloads in the data center; and They represent the start time and end time of the qth delay-tolerant workload processing period in the data center, respectively; represents the latency-sensitive workloads scheduled to data center d; represents the delay-tolerant workload processed by data center d; represents the total number of workloads processed, Indicates that the operation scheduling time period is considered. represents the processable period of the qth delay-tolerant workload in data center d.

[0088] (2) Server startup constraints:

[0089]

[0090] In the formula, Represents the number of servers in the data center; N d,q represents the number of latency-tolerant workloads; Indicates the number of servers turned on to handle latency-sensitive workloads; represents the number of servers turned on to process latency-tolerant workloads; S d,t Indicates the total number of servers enabled; μ d Indicates the processing rate of the server; t D Represents the processing latency bounds for flexible latency-sensitive workloads; represents the transmission delay for latency-sensitive workloads, Represents flexible latency-sensitive workloads in data centers, Represents fixed latency-sensitive workloads in a data center.

[0091] (3) Operation constraints of energy storage in UPS:

[0092]

[0093] In the formula, E d,t Indicates the stored electrical energy; and Respectively represent the charge and discharge efficiency; and Respectively represent the charge and discharge power; and Respectively represent the lower and upper limits of stored electrical energy; Indicates energy storage capacity; E d,0 Indicates the initial stored electrical energy; and They represent the upper limits of the charging and discharging power of energy storage, and Respectively represent the lower limit of energy storage.

[0094] (4) Data center power constraints:

[0095]

[0096] In the formula, Indicates the power of IT equipment, and Represent the output power of photovoltaic and wind turbine respectively; Represents other power of the data center; P d,t Represents the interaction power between the data center and the power distribution system; P peak,d and P idle,d They represent the peak power and idle power of IT equipment respectively; pue represents the PUE value of the data.

[0097] As a specific implementation method, in the iterative solution process between the distribution system operator and the data center operator, an iterative solution algorithm based on the dichotomy method is proposed to solve the oscillation problem that may occur in the iterative process and ensure the convergence of the iterative solution process. The proposed dichotomy iterative solution algorithm is specifically shown in Table 1:

[0098] Table 1

[0099]

[0100] The results of the iterative solution process based on the binary search method are as follows Figure 3 The marginal price of distribution nodes is shown in Figure 4 The results of the carbon intensity of the distribution system nodes are shown in Figure 5 The spatial scheduling results of latency-sensitive workloads in data centers are shown in Figure 6 The time scheduling results of the data center delay-tolerant workload are shown in Figure 7 The operating cost results of the data center are shown in Table 2.

[0101] Table 2

[0102]

[0103] Embodiment 2

[0104] This embodiment further provides a computer device, including a memory, a processor, and a computer program stored in the memory, wherein the processor executes the computer program to implement the steps of the method.

[0105] Embodiment 3

[0106] This embodiment also provides a computer-readable storage medium on which a computer program is stored. When the computer program is executed by a processor, the steps of the method are implemented.

[0107] The above are only preferred specific implementations of the present application, but the protection scope of the present application is not limited thereto. Any changes or substitutions that can be easily thought of by a person skilled in the art within the technical scope disclosed in the present application should be included in the protection scope of the present application. Therefore, the protection scope of the present application should be based on the protection scope of the claims.

Claims

1. A method for optimizing economic and low-carbon operation of a data center in a power distribution system, characterized in that: The following steps are involved: With the goal of minimizing the energy economic cost of the distribution system, the DIST-flow power flow constraint is adopted to establish the distribution system operation optimization model; Solving the distribution system operation optimization model to obtain the power flow results of the distribution system and the marginal electricity price of the distribution nodes; Based on the power flow results of the distribution system and the carbon emission flow theory, the carbon intensity of the distribution system nodes is solved; The marginal electricity price of the distribution nodes of the distribution system and the carbon intensity of the distribution system nodes are sent to the data center. The data center aims to minimize the comprehensive cost and establish an economic low-carbon operation optimization model for the data center. Solving the economic and low-carbon operation optimization model of the data center, obtaining the results of electricity purchase and sale between the data center and the power distribution system, and sending them to the power distribution system; The power distribution system and the data center iteratively solve the corresponding optimization models in turn until the interactive power of the two converges, completing the economic and low-carbon operation optimization of the data center.

2. The method according to claim 1, characterized in that The objective function of the distribution system operation optimization model is as follows: In the formula, c g,t It represents the wholesale market electricity price in the upper power grid; P represents the electricity price in the distribution system; g,t Indicates the power purchased and sold by the distribution system in the wholesale market; represents the electricity purchased from distributed generators; Indicates the power purchased from the data center.

3. The method according to claim 1, characterized in that The DIST-flow power flow constraint is as follows: Where P ij,t and Q ij,t Respectively represent the active and reactive power on branch ij; and They represent the active and reactive power injected into node j respectively; l ij,t represents the current on branch ij; u j,t represents the voltage amplitude of node j; r ij and x ij They represent the resistance and reactance of line ij respectively; and They represent the connection relationship between the power distribution system node j and the distributed generator and the data center respectively; and denote the active and reactive loads of node j respectively; and They represent the upper and lower limits of the voltage amplitude at node j respectively; λ j,t represents the dual variable of the corresponding constraint; P jk,t and Q jk,t denote the active and reactive power on line jk respectively, is a binary parameter, indicating the connection status between node j and the upper power grid, P g,t Represents the interactive active power with the upper grid, P dg,t Represents the output active power of distributed generators, P d,t Indicates the active load power of the data center, Represents the output reactive power of distributed generators, Q g,t Represents the interactive reactive power with the upper grid, Q d,t Indicates the reactive load power of the data center.

4. The method according to claim 1, characterized in that: The carbon intensity calculation formula of energy on node i and line ij of the distribution system is as follows: In the formula, ρ ij and ρ i are the carbon intensity of line ij and node i respectively; is the energy flowing from node i to line ij; is the energy loss generated on line ij; is the energy flowing from line ij to node j; and They are and The corresponding carbon emission rate.

5. The method according to claim 4, characterized in that The process of solving the carbon intensity of the distribution system nodes also includes: calculating the carbon emission flow in the distribution system based on the law of conservation of carbon emissions, the principle of proportional allocation, and the principle of energy merging.

6. The method according to claim 1, characterized in that The objective function of the data center economic low-carbon operation optimization model is as follows: In the formula, and Respectively represent the power purchase and power sales of the data center; λ d,t and Respectively represent the purchase price and sale price of electricity of the data center; c e is the carbon emission cost coefficient; Represents the carbon intensity of a data center node in the power distribution system.

7. The method according to claim 6, characterized in that The constraints of the data center operation optimization model include: workload operation constraints, server startup constraints, operation constraints of energy storage in UPS, and data center power constraints.

8. The method according to claim 1, characterized in that In the process of iteratively solving the corresponding optimization models of the power distribution system and the data center in turn, an iterative solution algorithm based on the dichotomy method is adopted.

9. A computer device comprising a memory, a processor and a computer program stored in the memory, characterized in that: The processor executes the computer program to implement the steps of the method according to any one of claims 1 to 8.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 8 are implemented.

Images