Power distribution network operation regulation and control method and device based on data center, terminal and medium

Abstract

The invention discloses a power distribution network operation regulation and control method and device based on a data center, a terminal and a medium. According to the scheme provided by the invention, a joint optimization model is established, energy storage charging and discharging power and data center load distribution are used as associated decision variables, and power grid physical characteristics and task processing requirements are considered in an optimization target at the same time; and the synergistic interaction of the two flexible resources is realized. Through the technical scheme, the operation strategies of the battery energy storage system and the data center can be dynamically coordinated, so that the power regulation capability of the BES and the space-time flexibility of the data center are complementary, deep space-time coordinated regulation and control in a daily operation scene of a power distribution network between the BES and the data center are realized, the space-time distribution of the power is optimized, and the network loss is reduced; and dual goals of power balance and network loss reduction are realized.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of power distribution networks, and particularly relates to a power distribution network operation regulation method and device based on a data center, a terminal and a medium. BACKGROUND

[0002] With the development of smart grids, the penetration rate of renewable energy is continuously increasing, but its uncertainty brings challenges to the supply and demand balance of the power grid. In order to improve the flexibility of the power grid, the current power grid gradually increases the use of battery energy storage systems (BES) and other devices. At the same time, in recent years, the intelligent load dynamically adjusted by the algorithm of the data center can actively participate in the optimization of the power grid and play a similar optimization and regulation role to the BES in the power grid system.

[0003] However, the prior art does not realize the collaborative regulation of the BES and the data center, and when the battery energy storage system (BES) or the data center flexibility regulation is used alone, it is difficult to simultaneously achieve the dual goals of peak clipping and valley filling and reducing network loss in the daily operation of the power grid. SUMMARY

[0004] The present application provides a power distribution network operation regulation method and device based on a data center, a terminal and a medium, which are used to realize the collaborative optimization and regulation of the data center and the battery energy storage system, and improve the efficiency and reliability of the power grid.

[0005] To achieve the above-mentioned purpose, the first aspect of the present application provides a power distribution network operation regulation method based on a data center, comprising:

[0006] According to the topological structure information of the power distribution network, the nodes contained in the power distribution network and the node static parameters of each node are determined, wherein the nodes include: load nodes, renewable energy nodes, data center nodes and BES nodes;

[0007] Real-time operation data of each node is collected, and based on the node static parameters and the real-time operation data, an operation regulation optimization model is constructed, wherein the operation regulation optimization model includes: a cost minimization objective function, a data center load constraint sub-model, a BES charging and discharging constraint sub-model and a data center space-time flexibility constraint sub-model;

[0008] The operation regulation optimization model is solved to obtain the optimal operation regulation scheme of the power distribution network according to the solving result.

[0009] Preferably, the cost minimization objective function is specifically:

[0010]

[0011]

[0012]

[0013]

[0014]

[0015] wherein, is the cost of electricity purchased from the main grid; is the cost of active power loss of the distribution grid; is the cost of operation loss of the battery energy storage system (BES); is the comprehensive cost of data center task scheduling, is the time-of-use electricity price of time period t, is the active power purchased from the main grid; E is the set of branches of the distribution grid, is the resistance of branch ij, , is the active and reactive power of the branch, is the voltage amplitude of the node, is the unit energy loss cost, , are the charging and discharging efficiencies of the BES, respectively, , is the optimization variable for load fluctuation smoothing; is the IT device power consumption of data center d in time period t, is the task delay amount, , are the energy consumption and delay weight coefficients.

[0016] Preferably, the data center power consumption constraint sub-model is specifically:

[0017]

[0018]

[0019] wherein, is the data center power consumption, is the data center computing device power, is the power usage effectiveness of the data center, is the CPU power consumption bias term coefficient, is the CPU computing power bias term coefficient, is the server power consumption coefficient, is the number of servers running in the data center.

[0020] Preferably, the BES charging and discharging constraint sub-model is specifically:

[0021]

[0022]

[0023]

[0024]

[0025]

[0026]

[0027] In the formula, is an index related to the BES number, is a time index, is the SOC variable of the BES, is the BES efficiency, and are the charge and discharge power of the BES, respectively, is the capacity of the BES, and are the upper and lower limits of the SOC of the BES, respectively, and are the upper limits of the active charge and discharge of the BES, respectively, is a binary variable indicating the charge and discharge of the BES, is the reactive power related to the BES, and are the upper and lower limits of the reactive power of the BES, is the minimum SOC value specified after the end of the scheduling period.

[0028] Preferably, the data center space-time flexibility constraint sub-model is specifically:

[0029]

[0030]

[0031]

[0032]

[0033]

[0034]

[0035]

[0036]

[0037]

[0038]

[0039]

[0040]

[0041]

[0042] wherein, is the index of the data center, is the index of time, is the index of the front-end server, is the total number of system front-end servers, is the index of the workload type, is the amount of workload assigned, is the total number of all other data centers in the system except for data center d, is the maximum number of delay periods allowed for the workload, is the total length of the scheduling period, is the amount of workload migrated from other data centers to the data center d, is the amount of workload assigned to the data center d by the front-end server and other data centers, is the amount of workload received by the data center and the front-end server, is the total amount of user workload requests received by the front-end server, is the real-time workload amount of the data center d at time t; is the amount of workload processed by the data center d, is the amount of workload destroyed in the data center d, is the amount of workload stored in the data center d, is the maximum workload storage capacity, is the amount of process workload that must be completed, is the initial inventory amount expected by the data center at the beginning of the scheduling period, and are binary variables associated with the processed workload and the destroyed workload, respectively.

[0043] Preferably, the operation regulation optimization model further comprises a robust optimization sub-model for optimizing the uncertain variables of the renewable energy relay and the load node.

[0044] Preferably, the operation regulation optimization model further comprises, before being solved:

[0045] The electricity purchase cost is converted into a linear form through a piecewise linear processing method.

[0046] The second aspect of the present application provides a data center-based power distribution network operation regulation device, comprising:

[0047] A node information determination unit is configured to determine nodes contained in the power distribution network and node static parameters of each node according to topology structure information of the power distribution network, wherein the nodes include load nodes, renewable energy nodes, data center nodes and BES nodes.

[0048] A node operation data acquisition unit is configured to collect real-time operation data of each node, and construct an operation regulation optimization model based on the node static parameters and the real-time operation data, wherein the operation regulation optimization model includes a cost minimization objective function, a data center load constraint sub-model, a BES charging and discharging constraint sub-model and a data center space-time flexibility constraint sub-model.

[0049] A regulation scheme optimization unit is configured to solve the operation regulation optimization model, and obtain an optimal operation regulation scheme of the power distribution network according to a solution result.

[0050] The third aspect of the present application provides a data center-based power distribution network operation regulation terminal, comprising a memory and a processor.

[0051] The memory is configured to store program code corresponding to the data center-based power distribution network operation regulation method provided in the first aspect of the present application.

[0052] The processor is configured to read and execute the program code to implement the data center-based power distribution network operation regulation method.

[0053] The fourth aspect of the present application provides a computer readable storage medium, characterized in that the computer readable storage medium has program code saved therein, the program code being configured to be read and executed by a processor to implement the data center-based power distribution network operation regulation method provided in the first aspect of the present application.

[0054] As can be seen from the above technical solutions, the present application has the following advantages:

[0055] The scheme provided in the application realizes the synergistic effect of the two flexible resources by establishing a joint optimization model, taking the storage charging and discharging power and the data center load distribution as correlated decision variables, and simultaneously considering the power grid physical characteristics and task processing requirements in the optimization objective. Through the above technical scheme, the application can dynamically coordinate the operation strategies of the battery energy storage system and the data center, so that the power regulation capability of the BES and the space-time flexibility of the data center form a complement, realize the deep space-time collaborative regulation between the BES and the data center in the daily operation scene of the distribution network, optimize the power space-time distribution, reduce the network loss, and realize the dual goals of power balance and network loss reduction. BRIEF DESCRIPTION OF DRAWINGS

[0056] In order to more clearly illustrate the technical solutions in the embodiments of the application or the prior art, the following will briefly introduce the drawings needed to be used in the embodiments or the prior art description. Obviously, the drawings in the following description are only some embodiments of the application, and for those skilled in the art, other drawings can also be obtained without creative labor.

[0057] Figure 1 The flowchart of an embodiment of a power distribution network operation regulation method based on a data center provided by the application.

[0058] Figure 2 The architecture diagram of an embodiment of a power distribution network operation regulation device based on a data center provided by the application.

[0059] Figure 3 The architecture diagram of an embodiment of a power distribution network operation regulation terminal based on a data center provided by the application. DETAILED DESCRIPTION

[0060] In the prior art, with the increasing penetration of renewable energy in the power grid, the power generation fluctuation of renewable energy makes it difficult to balance the supply and demand of the power grid. The traditional method adjusts the power by the battery energy storage system or uses the data center load transfer to smooth the fluctuation, but the battery energy storage system cannot optimize the calculation task distribution, and the data center scheduling ignores the physical characteristics of the power grid, so it is difficult to consider both peak clipping and network loss reduction when the two are applied separately. For example, during the noon photovoltaic output peak period, if only the battery energy storage is relied on to absorb excess power, the renewable energy may be abandoned due to the limitation of the storage capacity, and if the data center only allocates the load according to the fixed strategy, it cannot dynamically respond to the real-time state of the power grid.

[0061] To solve the above problems, the inventors find that the charge-discharge characteristics of the battery energy storage system and the space-time flexibility of the data center are complementary. The battery energy storage system can quickly adjust the charge-discharge power, but is limited by the capacity and efficiency; the data center can flexibly adjust the load distribution through task migration, delay processing and the like, but lacks deep coupling with the physical model of the power grid. Thus, the core idea is to jointly include both in a unified optimization framework, to build a comprehensive model including the topology of the power grid, device parameters, real-time operation data, to jointly optimize the charge-discharge plan and the load distribution scheme, to achieve the dual goals of power balance and network loss reduction.

[0062] Therefore, the embodiments of the present application provide a power distribution network operation regulation method and device based on a data center, a terminal and a medium, which are used to realize the collaborative optimization and regulation of the data center and the battery energy storage system, and improve the efficiency and reliability of the power grid.

[0063] In order to make the application purposes, features and advantages of the present application more obvious and easy to understand, the technical solutions in the embodiments of the present application will be described clearly and completely below with reference to the drawings in the embodiments of the present application. Obviously, the embodiments described below are only a part of the embodiments of the present application, rather than all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative labor fall within the scope of the present application.

[0064] First, a detailed description of an embodiment of the power distribution network operation regulation method based on a data center provided by the present application is as follows:

[0065] Referring to Figure 1 The power distribution network operation regulation method based on a data center provided by the present embodiment comprises the following steps:

[0066] Step 101, determining the nodes contained in the power distribution network and the node static parameters of each node according to the topology structure information of the power distribution network;

[0067] The nodes include: load nodes, renewable energy nodes, data center nodes and BES nodes;

[0068] Step 102, collecting real-time operation data of each node, and constructing an operation regulation optimization model based on the node static parameters and the real-time operation data;

[0069] The operation regulation optimization model includes: a cost minimization objective function, a data center load constraint sub-model, a BES charge-discharge constraint sub-model and a data center space-time flexibility constraint sub-model;

[0070] Step 103, solving the operation regulation optimization model to obtain an optimal operation regulation scheme of the power distribution network according to the solution.

[0071] It should be noted that the node static parameter refers to a fixed parameter describing the basic attribute of the node, which can be specifically implemented by using load type, rated capacity, geographic location information, and is used to establish the topological relationship of the power grid and the operation boundary of the equipment. The real-time operation data includes renewable energy output, load demand, battery energy storage state of charge and data center task queue, which is used to reflect the dynamic change of the system. The cost minimization objective function is formed by integrating the power purchase cost, network loss cost, energy storage loss and task scheduling cost to form the economic optimization orientation. The data center load constraint is quantitatively calculated by establishing a correlation model between server power consumption and task processing capacity to reflect the influence of load on power grid power. The battery energy storage charging and discharging constraint ensures the safe operation of the energy storage system by defining the upper and lower limits of the state of charge, the charging and discharging power limit and the efficiency parameter. The space-time flexibility constraint is represented by introducing task migration, delay processing and storage capacity limit to represent the adjustable range of data center load in time and space dimensions.

[0072] Specifically, the method first analyzes the distribution network topology structure, identifies four types of nodes and extracts static parameters. After collecting the real-time operation data of each node, the static parameters and dynamic data are input into the optimization model. The model takes the minimum total cost as the target, and couples the data center task allocation and the battery energy storage charging and discharging behavior through the constraint condition. For example, during the period of excess renewable energy, the model can simultaneously increase the energy storage charging power and adjust the data center task processing time to absorb the excess power; during the load peak period, the model coordinates the energy storage discharging and data center load migration to reduce the main grid power purchase demand. In the solving process, the mixed integer linear programming algorithm is used to process discrete variables and continuous variables, and finally the optimization scheme including the energy storage charging and discharging instruction and the data center task scheduling strategy is output.

[0073] The method establishes a joint optimization model to take the energy storage charging and discharging power and the data center load allocation as related decision variables, and simultaneously considers the physical characteristics of the power grid and the task processing demand in the optimization target, so as to realize the synergistic effect of the two flexible resources. Through the above technical scheme, the application can dynamically coordinate the operation strategy of the battery energy storage system and the data center, optimize the power space-time distribution, and reduce the network loss under the scenario of renewable energy fluctuation and load demand change. For example, during the photovoltaic output sudden drop, the model can simultaneously reduce the data center non-emergency task processing amount and start the energy storage discharging to avoid the network loss increase caused by the sudden increase of the main grid power purchase power. At the same time, through the joint optimization of task migration and energy storage charging and discharging plan, the power grid's ability to absorb distributed energy is improved, and the renewable energy waste phenomenon is reduced.

[0074] On the basis of the above basic embodiment, further, the cost minimization objective function mentioned in the above basic embodiment has the expression:

[0075]

[0076]

[0077]

[0078]

[0079]

[0080] wherein, is the cost of purchasing power from the main grid by the substation; is the cost of active power loss in the distribution grid; is the cost of operation loss of the battery energy storage system (BES); is the comprehensive cost of data center task scheduling, including energy consumption and delay penalty, which is optimized by flexible scheduling of tasks in time (delayed execution) and space (migration to low-loss nodes), is the time-of-use electricity price of time period t, is the active power purchased from the main grid; E is the set of branches of the distribution grid, is the resistance of branch ij, , is the active and reactive power of the branch, is the node voltage amplitude, and the grid loss is directly determined by the power flow distribution, which is affected by the BES reactive power support and the spatial distribution of data center load, is the unit energy loss cost, , are the charging and discharging efficiencies of the BES, respectively, , is the optimization variable for load fluctuation smoothing; is the IT device power consumption of data center d in time period t, is the task delay amount, , are the energy consumption and delay weight coefficients.

[0081] The time-of-use electricity price refers to the dynamic change parameter of the power grid electricity price in different time periods, which can be specifically realized by using real-time electricity price data or a time-of-use electricity price model, and is used to reflect the power supply and demand relationship in different time periods, thereby optimizing the main grid electricity purchase cost. The active power loss cost refers to the energy loss cost of the distribution network in the transmission process due to resistance, which can be specifically calculated by the product of branch resistance and current square, and is used to quantify the influence of network loss on economy. The charging and discharging efficiency refers to the efficiency parameter of energy conversion of the battery energy storage system in the charging and discharging process, which can be specifically represented by the ratio of charging and discharging power to energy loss, and is used to constrain the energy loss cost of the BES operation. The task delay amount refers to the delay time of the data center in processing the computing task due to the scheduling strategy, which can be specifically obtained through a task queue model or real-time monitoring data, and is used to balance the relationship between computing load allocation and delay penalty.

[0082] Specifically, the cost minimization objective function integrates the main grid electricity purchase cost, the distribution network active loss cost, the BES operation loss cost, and the comprehensive cost of data center task scheduling to construct a multi-dimensional economy optimization model. The main grid electricity purchase cost is dynamically adjusted based on the time-of-use electricity price, and the electricity purchase amount is increased in the electricity price valley period to reduce the total cost; the distribution network active loss cost is calculated by branch resistance, current and voltage parameters, and is used to optimize the power distribution in the network topology; the BES operation loss cost combines the charging and discharging efficiency and the power variable to constrain the energy conversion efficiency of the energy storage system; and the comprehensive cost of data center task scheduling is the product of IT equipment power consumption, task delay amount and weight coefficient, to realize the trade-off between computing load allocation and delay penalty. By solving the objective function, the main grid electricity purchase strategy, the distribution network power distribution, the BES charging and discharging plan, and the data center load allocation can be dynamically adjusted, thereby realizing the minimization of the overall operation cost.

[0083] The scheme introduces parameters such as time-of-use electricity price, branch resistance, dynamic charging and discharging efficiency, and task delay amount to construct a more accurate multi-objective cost model, which can comprehensively reflect the economic influence factors in the operation of the distribution network. In addition, by including the energy consumption and delay weight coefficient into the objective function, the collaborative optimization of data center load scheduling and grid physical characteristics is realized, breaking through the limitations of the traditional method of separating the two. Through the above technical scheme, the comprehensive cost of the distribution network operation can be effectively reduced, which is specifically manifested as: reducing the main grid electricity purchase amount and preferentially calling the BES stored energy in the electricity price peak period to reduce the electricity purchase cost; reducing the line loss by optimizing the distribution network power distribution to reduce the active power loss cost; dynamically adjusting the charging and discharging strategy in combination with the BES charging and discharging efficiency to reduce the energy conversion loss; and dynamically allocating the computing load according to the task delay sensitivity of the data center to minimize the energy consumption cost on the premise of ensuring the quality of service. Finally, the collaborative optimization of multi-dimensional economy indicators is realized, and the operation efficiency and economic benefit of the distribution network are improved.

[0084] Further, regarding the data center power consumption constraint sub-model mentioned in the above basic embodiment, specifically includes the following expressions:

[0085]

[0086]

[0087] In the formula, is the data center power consumption, is the data center computing device power, is the power usage efficiency of the data center, is the CPU power consumption bias term coefficient, is the CPU computing power bias term coefficient, is the server power consumption coefficient, is the number of servers running in the data center.

[0088] Among them, the power usage efficiency refers to the ratio of the overall energy consumption of the data center to the IT device energy consumption, which can be realized by real-time monitoring or preset empirical value, and the energy efficiency level of auxiliary facilities such as refrigeration and power supply is reflected by dynamically adjusting the parameter, which provides basis for optimizing the overall power consumption.

[0089] Further, regarding the BES charging and discharging constraint sub-model mentioned in the above basic embodiment, the following constraint conditions are included:

[0090]

[0091]

[0092]

[0093]

[0094]

[0095]

[0096] In the formula, is the index related to the BES number, is the time index, is the SOC variable of the BES, is the BES efficiency, and are the charging and discharging power of the BES, is the capacity of the BES, and are the lower limit of the SOC of the BES, and are the upper limit of the active charging and discharging of the BES, is a binary variable indicating the BES charging and discharging state, is the reactive power related to the BES, and are the upper and lower limits of the BES reactive power, is the minimum SOC value after the end of the dispatching period.

[0097] wherein the state of charge variable is used to track the remaining capacity of the battery energy storage system in real time, which can be realized by integrating the charging and discharging power, and the upper and lower limits can prevent overcharging or overdischarging. The charging and discharging efficiency is used to represent the loss in the energy conversion process, which can be realized by the ratio of charging and discharging power to the change of state of charge, and the actual charging and discharging amount can be accurately calculated through the efficiency parameter. The charging and discharging binary variable is used to represent the mutual exclusivity of charging and discharging state, which can be realized by logical constraints, for example, the charging and discharging state cannot be activated at the same time in the same period. The reactive power constraint is used to control the reactive power regulation capability of the battery energy storage system, which can be determined by the inverter capacity or the grid reactive power demand, to ensure that it participates in grid regulation within a reasonable range. The minimum state of charge after the end of the dispatching period is used to ensure the standby capability of the energy storage system, which can be set by a pre-set threshold or grid dispatching demand, to avoid affecting the subsequent dispatching flexibility due to excessive discharge.

[0098] Further, regarding the data center space-time flexibility constraint sub-model mentioned in the above basic embodiment, the following constraint conditions are included:

[0099]

[0100]

[0101]

[0102]

[0103]

[0104]

[0105]

[0106]

[0107]

[0108]

[0109]

[0110]

[0111]

[0112] wherein, is the data center number index, is the time index, is the front-end server number index, is the total number of system front-end servers, is the workload type index, is the assigned workload amount, is the total number of all other data centers in the system except data center d, is the maximum number of delay periods allowed for the workload, is the total length of the scheduling period, is the amount of workload migrated from other data centers to the data center d, is the amount of workload assigned to data center d by the front-end servers and other data centers, is the amount of workload received by the data center and the front-end servers, is the total amount of user workload requests received by the front-end servers, is the real-time workload amount of data center d at time t; is the amount of workload processed by data center d, is the amount of workload destroyed in data center d, is the amount of workload stored in data center d, is the maximum workload storage capacity, is the amount of process workload that must be completed, is the initial inventory amount expected by the data center at the beginning of the scheduling period, and are binary variables associated with processed workload and destroyed workload, respectively.

[0113] The workload amount allocated to each data center by the front-end server and other data centers refers to the amount of tasks allocated to the current data center through front-end server scheduling or other data center task migration, which can be implemented by using a distributed task scheduling algorithm for dynamically adjusting cross-regional workload allocation; the workload amount processed by each data center refers to the amount of workload actually completed in the calculation task in the period, which can be implemented by using a priority queue management mechanism for balancing real-time processing capacity and task backlog; the destroyed workload amount refers to the amount of tasks abandoned for processing due to delay overrun or insufficient resources, which can be implemented by using a timeout discard strategy combined with a quality of service threshold for avoiding invalid resource occupation; the stored workload amount refers to the amount of tasks temporarily stored in the data center for subsequent processing, which can be implemented by using a cache pool capacity dynamic allocation mechanism for realizing time-space translation of task processing; the maximum workload storage capacity refers to the upper limit of tasks that can be temporarily stored in the data center, which can be dynamically adjusted by using storage hardware resource configuration or virtualization technology for preventing system crash caused by storage overflow; the process workload amount that must be completed refers to critical tasks that must be processed before the end of the scheduling period, which can be implemented by using task label classification and deadline constraints for guaranteeing core business continuity; the binary variable refers to a logical identifier for controlling the processing and destruction state of the workload, which can be implemented by using a state machine model combined with a conditional judgment statement for ensuring the mutual exclusivity and completeness of the operation logic.

[0114] Specifically, the time-space flexibility of the data center includes time flexibility and space flexibility. The time flexibility is achieved by cross-period transfer of workload, and the space flexibility is achieved by preferentially allocating load to data centers close to the upstream network to reduce line loss. During the operation of the power distribution network, the data center dynamically receives workload according to the front-end server scheduling instruction and the task migration request of the adjacent data center. The received workload and the tasks temporarily stored in the last period form a task pool to be processed. The processing module determines the actual processing amount in the current period according to the real-time power supply state and the availability of computing resources. The unprocessed tasks are part of the destruction judgment link, and if the preset delay tolerance is exceeded, the destruction operation is triggered. The remaining tasks are transferred to the storage queue. The storage module continuously monitors the cache capacity and triggers an early warning and adjusts the task receiving strategy when approaching the upper limit. At the end of the period, the system forces all tasks marked as must be completed to ensure that critical business is not affected by time-space scheduling. The processing and destruction operations are mutually exclusive controlled by binary variables to avoid logical conflicts.

[0115] By the technical solution, the application can effectively coordinate the distribution of the data center work load in time and space, reduce real-time power consumption by temporarily storing non-urgent tasks during the peak period of the power grid load, and process backlog tasks during the period when the renewable energy output is abundant. This dynamic adjustment mechanism makes the power demand curve of the data center better match the power supply capacity of the distribution network, not only alleviating the power imbalance problem caused by rigid load in the traditional method, but also optimizing the overall network loss distribution through cross-regional task scheduling, providing a basic support condition for the coordinated regulation of BES and data center.

[0116] The technical solution of the application is based on the operation characteristics of the BES of the distribution network and the data center, and an operation regulation optimization model containing a cost minimization objective function, a data center load constraint sub-model, a BES charging and discharging constraint sub-model, and a data center time and space flexibility constraint sub-model is constructed through unified modeling. Then, through joint solving, the power regulation capability of the BES and the time and space flexibility of the data center are complementary, deep time and space coordinated regulation between the BES and the data center is realized, the power purchase cost is reduced, the network loss is reduced, and the overall economy and reliability of the system are improved.

[0117] Among them, the time dimension coordination refers to that during the low price valley period (such as night), the BES charges, and at the same time the data center can increase the non-real-time task load, realizing "double charging"; during the peak period, the BES discharges, and the data center reduces the load or delays the task, jointly reducing , and realizing peak shaving. The space dimension coordination refers to that the optimization model can preferentially allocate the computing task to the data center close to the power supply or the upstream node of the power grid (such as the node ij voltage is high and the line is short), thereby reducing and . The BES can also be deployed at the voltage weak point to provide reactive power support, improve , and indirectly support the stable operation of the data center.

[0118] Further, the embodiment also proposes to add a robust optimization sub-model in the operation regulation optimization model, which is used to optimize the uncertain variables of the renewable energy generation and the load node.

[0119] Among them, the robust optimization sub-model refers to using the robust optimization method to process the uncertainty of the renewable energy output fluctuation and the load demand change. Specifically, the multi-scenario robust optimization or interval robust optimization method can be used to realize it, and the fluctuation range set of the uncertain variable is constructed to make the optimization decision in the worst case. The uncertain variables include the photovoltaic power prediction deviation, the wind power random fluctuation, and the real-time power demand deviation of the load node. The fluctuation boundary of these variables is determined through historical data statistics or prediction error analysis.

[0120] Robust optimization adopts the box-type uncertainty set to deal with the fluctuation of uncertain variables in renewable energy and load, and the calculation formula is shown in the following formula:

[0121]

[0122]

[0123]

[0124] wherein, and are the average value of the uncertain variable of the load in the robust optimization and the disturbance of the uncertain variable of the load in the robust optimization, respectively; and are the upper limit of the uncertain variable and the lower limit of the active power, wherein the uncertain variables include the photovoltaic power prediction deviation, the wind power random fluctuation and the real-time power demand deviation of the load node, and the fluctuation boundary of these variables is determined through historical data statistics or prediction error analysis.

[0125] On the basis of the time dimension coordination and the space dimension coordination of the above-mentioned sub-models, the application further optimizes the deep coordination of the BES and the data center through the robust coordination under uncertainty by introducing the robust optimization logic, the robust coordination of the embodiment optimizes the minimum cost under the worst scenario by adopting the box-type uncertainty set to describe the renewable energy output and the load demand, dynamically adjusts the task allocation and the BES charging and discharging strategy, improves the robustness of the scheduling strategy of the BES and the data center, improves the anti-interference ability of the distribution network, and ensures that the power grid can still maintain stable operation under the most unfavorable scenario.

[0126] Further, the embodiment further proposes that before solving the operation regulation and optimization model, the purchase power cost is converted into a linear form through a segmented linear processing mode, the purchase power cost and the active power loss of the distribution network under the time-of-use electricity price or the tiered electricity price are converted into a linear expression, and after linearization processing, the whole optimization problem is converted into a mixed integer linear programming model, so as to improve the solving efficiency.

[0127] It should be noted that, in order to ensure the solvability of the model, the embodiment performs segmented linearization processing on the nonlinear term. For example, for the purchase power cost under the time-of-use / tiered electricity price, it is divided into k intervals in the power dimension, a linear function is used for approximation in each interval, and a 0-1 variable is introduced to control the activation order of the interval. For the active power loss of the distribution network, the method of combining the direct current flow assumption with the segmented linearization of the power square term is adopted to convert it into a linear expression. After linearization processing, the whole optimization problem can be converted into a mixed integer linear programming (MILP), which is solved efficiently by using a commercial solver, and supports online rolling optimization.

[0128] The segmented linear processing mode refers to converting a nonlinear function into a linear expression through a segmented linear approximation method, and can be implemented by setting multiple segmentation points, for example, dividing the power interval corresponding to the electricity price ladder into multiple linear segments, and keeping the electricity price constant in each interval. This processing mode can convert the originally nonlinear electricity purchase cost function into a linear form, thereby reducing the model complexity. The mixed integer linear programming model refers to a linear optimization problem containing continuous variables and integer variables, and can be efficiently solved by using a commercial solver. This model introduces binary variables to describe the charging and discharging state, task scheduling logic and other discrete decisions, while retaining the linear constraint structure, which can significantly shorten the calculation time while ensuring the accuracy of the solution.

[0129] Specifically, in the time-of-use electricity price scenario, the electricity purchase cost and the main grid electricity purchase power can present a nonlinear relationship, for example, in the step electricity price mechanism, different power intervals correspond to different electricity prices. Through segmented linearization processing, the electricity price curve is divided into multiple linear intervals, and the electricity price and power in each interval have a fixed proportional relationship. At the same time, the quadratic term in the active power loss of the distribution network can be converted into a linear expression by a linear approximation method. After the above processing, the nonlinear terms in the original optimization model are eliminated, and the objective function and the constraint conditions are all converted into linear forms. Combined with the linearized sub-models of the data center load constraints, BES charging and discharging constraints, etc., the entire problem is reconstructed into a mixed integer linear programming model, ensuring that the function is solvable, and then an optimization solver is used to realize fast solving.

[0130] The above is a detailed description of an embodiment of the data center-based distribution network operation regulation method provided by the present application. The following is a detailed description of related embodiments of a data center-based distribution network operation regulation device, a terminal, and a storage medium provided by the present application.

[0131] Please refer to Figure 2 The data center-based distribution network operation regulation device provided by the present embodiment comprises:

[0132] The node information determination unit 201 is configured to determine the nodes contained in the distribution network and the node static parameters of each node according to the topological structure information of the distribution network, wherein the nodes include load nodes, renewable energy nodes, data center nodes, and BES nodes.

[0133] The node operation data acquisition unit 202 is configured to collect real-time operation data of each node, and construct an operation regulation optimization model based on the node static parameters and the real-time operation data, wherein the operation regulation optimization model comprises a cost minimization objective function, a data center load constraint sub-model, a BES charging and discharging constraint sub-model, and a data center space-time flexibility constraint sub-model.

[0134] The regulation scheme optimization unit 203 is configured to solve the operation regulation optimization model, and obtain an optimal operation regulation scheme of the power distribution network according to a solution result.

[0135] Specifically, the node information determination unit first parses the power grid topology file, identifies the load nodes, renewable energy generation nodes, data center nodes and energy storage nodes, and extracts the static parameters of each node, such as rated power, capacity limit and the like. The node operation data acquisition unit collects voltage, current and power data in real time through sensors deployed in the power grid, and constructs an optimization model including a cost function and multiple constraints in combination with the static parameters. The regulation scheme optimization unit calls a mathematical programming algorithm to solve the model, and outputs a data center load distribution scheme and a battery energy storage charging and discharging plan, and finally forms an executable scheduling instruction set.

[0136] Through the above technical solutions, the application realizes joint scheduling of data center computing resources and energy storage systems, effectively reduces the power grid peak-valley difference and network loss by dynamically adjusting the work load distribution and energy storage charging and discharging timing under the premise of meeting the power grid power balance constraint. The device converts the time and space flexibility resources into adjustable power grid support capacity through the model-driven optimization decision mechanism, and improves the power grid operation stability in the renewable energy high penetration rate scenario.

[0137] Further, as shown in Figure 3 The power distribution network operation regulation terminal based on a data center provided by the embodiment includes a memory 33 and a processor 31, and the memory 33 and the processor 31 are connected through a communication bus 34.

[0138] The memory refers to a semiconductor device or a magnetic disk medium with a data storage function, which can be specifically implemented by a solid state disk or a flash memory chip, and is used to persistently save program codes including power distribution network topology analysis, real-time data acquisition, optimization model construction and solving logic. The processor refers to an integrated circuit with arithmetic logic operation capability, which can be specifically implemented by a multi-core central processing unit or a graphics processing unit, and is used to analyze program codes and perform model solving, instruction generation and other computationally intensive tasks.

[0139] Specifically, the terminal takes the power distribution network node static parameters and real-time operation data as inputs through the pre-stored program codes in the memory, constructs a mixed integer linear programming model including a cost objective function, data center load constraints, battery energy storage charging and discharging constraints and time and space flexibility constraints. The processor calls a linear solver to perform iterative calculation on the model, generates an optimal solution including a data center work load distribution scheme and a battery energy storage charging and discharging plan, and converts the calculation result into an executable scheduling instruction set. In the running process, the real-time operation data is periodically collected and updated to the model parameters, so as to ensure that the regulation scheme dynamically adapts to the power grid state change.

[0140] The terminal fuses the data center task delay cost and the battery energy storage charge and discharge efficiency parameters in the program code through the built-in joint optimization algorithm, so that the regulation scheme meets the calculation resource scheduling and the power grid physical constraint, and completes the power distribution network regulation task considering peak load shifting and network loss reduction.

[0141] Further, the computer readable storage medium provided by the embodiment has program codes saved therein, the program codes are read and executed by the processor to implement the power distribution network operation regulation method based on the data center, the method comprises determining nodes and node static parameters according to the topological structure information of the power distribution network, collecting real-time operation data to construct an operation regulation optimization model comprising a cost minimization objective function, a data center load constraint sub-model, a BES charge and discharge constraint sub-model and a data center space-time flexibility constraint sub-model, solving the model to obtain an optimal operation regulation scheme and issuing the optimal operation regulation scheme to an execution unit.

[0142] The computer readable storage medium refers to a physical carrier for saving program codes, which can be implemented by a solid state disk, a mechanical hard disk or a flash disk, and functions to provide persistent storage support for the program codes. The program codes refer to an executable instruction set comprising optimization model construction, solving logic and instruction generation functions, which can be implemented by a mixed integer linear programming algorithm and constraint condition coding, and functions to convert the operation regulation method into steps executable by the processor. The processor refers to a calculation unit for executing the program codes, which can be implemented by a central processing unit or a graphics processing unit, and functions to complete model solving and instruction generation by reading the program codes in the storage medium.

[0143] Those skilled in the art can clearly understand that, for the convenience and brevity of description, the specific working process of the terminal, device and unit described above can refer to the corresponding process in the foregoing method embodiments, which will not be described here.

[0144] In the several embodiments provided by the present application, it should be understood that the disclosed terminal, device and method can be implemented by other ways. For example, the device embodiments described above are only schematic, and for the convenience of description, the division of the units is only a logical function division, and there can be another division way in actual implementation, for example, multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. In addition, the displayed or discussed mutual couplings or direct couplings or communication connections between the units can be indirect couplings or communication connections through some interfaces, devices or units, and can be electrical, mechanical or other forms.

[0145] The terms “first,” “second,” “third,” “fourth,” etc. (if present) in the specification and accompanying drawings of this application are used to distinguish similar objects and are not necessarily used to describe a particular order or sequence. It should be understood that such data can be interchanged where appropriate so that embodiments of the application described herein can be implemented, for example, in orders other than those illustrated or described herein. Furthermore, the terms “comprising” and “having,” and any variations thereof, are intended to cover a 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.

[0146] It should be understood that in this application, "at least one (item)" means one or more, and "more than" means two or more. "And / or" is used to describe the relationship between related objects, indicating that three relationships can exist. For example, "A and / or B" can represent three cases: only A exists, only B exists, and both A and B exist simultaneously, where A and B can be singular or plural. The character " / " generally indicates that the preceding and following related objects are in an "or" relationship. "At least one (item) of the following" or similar expressions refer to any combination of these items, including any combination of single or plural items. For example, at least one (item) of a, b, or c can represent: a, b, c, "a and b", "a and c", "b and c", or "a and b and c", where a, b, and c can be single or multiple.

[0147] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.

[0148] Furthermore, the functional units in the various embodiments of the present invention can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or as a software functional unit.

[0149] The integrated unit, if implemented in the form of a software function unit and sold or used as an independent product, can be stored in a computer readable storage medium. Based on such understanding, the technical solutions of the present application essentially or say the part that contributes to the prior art or the whole or part of the technical solutions can be embodied in the form of a software product. The computer software product is stored in a storage medium and includes a number of instructions for causing a computer device (which can be a personal computer, a server, or a network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of the present application. The aforementioned storage medium includes: a U disk, a mobile hard disk, a read-only memory (ROM, Read-Only Memory), a random access memory (RAM, Random Access Memory), a magnetic disk or an optical disk, and various media that can store program codes.

[0150] The above-described embodiments are only used to illustrate the technical solutions of the present application, rather than limit them; although the present application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that they can still modify the technical solutions recorded in the foregoing embodiments, or make equivalent replacements for some technical features; and these modifications or replacements do not make the corresponding technical solutions deviate from the spirit and scope of the technical solutions of the embodiments of the present application.

Claims

1. A data center-based power distribution network operation control method, characterized in that, include: Based on the topology information of the distribution network, the nodes included in the distribution network and the static parameters of each node are determined, wherein the nodes include: load nodes, renewable energy nodes, data center nodes and BES nodes; Real-time operating data of each node is collected. Based on the static parameters of the nodes and the real-time operating data, an operation control and optimization model is constructed. The operation control and optimization model includes: a cost minimization objective function, a data center load constraint sub-model, a BES charging and discharging constraint sub-model, and a data center spatiotemporal flexibility constraint sub-model. By solving the operation and control optimization model, the optimal operation and control scheme of the distribution network can be obtained based on the solution results.

2. The data center-based power distribution network operation control method according to claim 1, characterized in that, The cost minimization objective function is specifically as follows: In the formula, The cost of purchasing electricity from the main grid; The cost of active power loss in the distribution network; The operating loss cost of the battery energy storage system; The overall cost of data center task scheduling. The time-of-use electricity price for period t. E represents the active power purchased from the main grid; E represents the set of distribution network branches. Let be the resistance of branch ij. , For the active and reactive power of the branch circuit, The node voltage amplitude, Cost per unit of energy loss , These represent the charge / discharge efficiency of BES. , These are optimization variables used to smooth out load fluctuations; The power consumption of IT equipment in data center d during time period t. This refers to the task delay. , The weighting coefficients are for energy consumption and delay.

3. The data center-based power distribution network operation control method according to claim 1, characterized in that, The specific sub-model for data center power consumption constraints is as follows: In the formula, It is the power consumption of the data center. Power consumption for data center computing equipment. For data center power efficiency, This is the coefficient for the CPU power consumption bias term. Calculate the power bias term coefficient for the CPU. This is the server power consumption factor. The number of servers running in the data center.

4. The data center-based power distribution network operation control method according to claim 1, characterized in that, The BES charge / discharge constraint sub-model is specifically as follows: In the formula, It is an index related to the BES number. It is a time index. It is a SOC variable of BES. It's about BES efficiency. and These are the charging and discharging power of BES. This is the capacity of BES. and These are the upper limits of BES's SOC. and These are the upper limits of BES active power charging and discharging. As a binary variable indicating the charging and discharging of BES, It refers to reactive power related to BES. and These are the upper and lower limits of BES reactive power. The minimum SOC value specified at the end of the scheduling cycle.

5. The data center-based power distribution network operation control method according to claim 1, characterized in that, The specific sub-model for data center spatiotemporal flexibility constraints is as follows: In the formula, It is a data center number index. It is a time index. For the front-end server's number index, This represents the total number of front-end servers in the system. For workload type indexes, The amount of workload allocated. This represents the total number of all data centers in the system except for data center d. The maximum number of latency periods allowed for the workload. The total duration of the scheduling cycle. This represents the workload of data center d being migrated from other data centers to this data center. This refers to the workload allocated to data center d by the front-end servers and other data centers. This refers to the workload received by the data center and front-end servers. This represents the total number of user workload requests received by the front-end server. Let be the real-time workload of data center d at time t; This represents the workload already processed by data center d. This refers to the amount of workload that has been destroyed in data center d. This refers to the workload stored in data center d. This is the maximum workload storage capacity. This refers to the workload of processes that must be completed. This represents the initial inventory level that the data center expects at the start of the scheduling cycle. and These are binary variables related to the workload being processed and binary variables associated with the destroyed workload, respectively.

6. The data center-based power distribution network operation control method according to claim 1, characterized in that, The operation and control optimization model also includes a robust optimization sub-model, which is used to optimize the uncertainties of renewable energy relays and load nodes.

7. The data center-based power distribution network operation control method according to claim 6, characterized in that, Before solving the aforementioned operational control optimization model, the following steps are also included: The electricity purchase cost is converted into a linear form through a piecewise linear processing method.

8. A data center-based power distribution network operation control device, characterized in that, include: The node information determination unit is used to determine the nodes included in the distribution network and the static parameters of each node based on the topology information of the distribution network. The nodes include: load nodes, renewable energy nodes, data center nodes and BES nodes. The node operation data acquisition unit is used to collect real-time operation data of each node, and construct an operation control optimization model based on the node static parameters and the real-time operation data. The operation control optimization model includes: a cost minimization objective function, a data center load constraint sub-model, a BES charging and discharging constraint sub-model, and a data center spatiotemporal flexibility constraint sub-model. The control scheme optimization unit is used to solve the operation control optimization model to obtain the optimal operation control scheme of the distribution network based on the solution results.

9. A data center-based power distribution network operation and control terminal, characterized in that, include: Memory and processor; The memory is used to store program code, which corresponds to the data center-based power distribution network operation and control method as described in any one of claims 1 to 7. The processor is used to read and execute the program code to implement the data center-based power distribution network operation and control method.

10. A computer-readable storage medium, characterized in that, The computer-readable storage medium contains program code that is read and executed by a processor to implement the data center-based power distribution network operation and control method as described in any one of claims 1 to 7.

Images