Power distribution network and data center collaborative planning method based on security domain analysis

Abstract

The invention relates to the field of power distribution network planning, and discloses a power distribution network and data center collaborative planning method based on security domain analysis, and the method comprises the following steps: S1, constructing a power distribution network-data center system security domain model containing a distributed power supply; s2, based on the constructed power distribution network-data center system security domain model, constructing a power distribution network-data center system double-layer collaborative optimization model; s3, solving the double-layer collaborative optimization model by using an algorithm solver to obtain a collaborative planning result of the power distribution network and the data center; in the step S2, the power distribution network-data center system security domain model represents a set of all working points meeting normal operation N-0 constraint and N-1 security constraint of the system. By constructing the power distribution network-data center security domain model and proposing the joint quantitative index, the continuous representation of the system security margin is realized, the anti-disturbance capability and the operation stability of the system are improved, and the local overload risk is avoided.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of power distribution network planning, in particular to a power distribution network and data center collaborative planning method based on safety domain analysis. BACKGROUND

[0002] With the development of digital economy, the scale and energy consumption of data centers continue to grow. Data center load presents the technical characteristics of high power density and strong dynamic fluctuation. Large-scale access of such loads to traditional power distribution networks will cause a series of technical challenges. On the one hand, power fluctuations of data center load may cause node voltage of power distribution network to exceed the specified limit or line load to exceed the thermal stability limit, affecting power quality and equipment safety. On the other hand, power supply interruption on the power distribution network side will directly affect the normal operation of the data center, which may cause data loss or service interruption, resulting in economic losses. Therefore, improving the operation safety of the power distribution network and data center interactive system through collaborative planning method is an important technical issue in building a new type of power distribution system.

[0003] However, it still has some problems that cannot be overcome. First of all, the current research generally relies on the traditional N-1 criterion when checking safety. However, in the presence of large-scale uncertain distributed power and high dynamic data center load, this criterion is less efficient in handling uncertainty, and it fails to quantitatively analyze the safety margin of the system, making it difficult to meet the planning flexibility requirements of the new power distribution system. At the same time, existing researches mostly focus on the optimization of power supply side and energy storage system, and the safety analysis framework for load side, especially for data centers with high density and high reliability, has not been perfected. In addition, although safety domain analysis methods have been applied to stability analysis in the system operation stage, their application in the collaborative planning of power distribution network and data center is still insufficient, and they fail to effectively embed safety boundaries and margins into the planning model, thus limiting the comprehensive performance of the planning scheme. SUMMARY

[0004] In view of the deficiencies of the prior art, the present application provides a power distribution network and data center collaborative planning method based on safety domain analysis, which solves the problem that the prior art lacks effective quantitative evaluation means for system safety margin when planning the collaborative planning of power distribution network with distributed power and data center, thus making it difficult to develop a planning scheme that takes into account both economic efficiency and operation safety.

[0005] To achieve the above purpose, the present application realizes the following technical scheme: a power distribution network and data center collaborative planning method based on safety domain analysis, comprising the following steps:

[0006] S1, constructing a safety domain model of power distribution network-data center system with distributed power;

[0007] S2, constructing a power distribution network-data center system double-layer collaborative optimization model based on the constructed power distribution network-data center system security domain model;

[0008] S3, solving the double-layer collaborative optimization model by using an algorithm solver to obtain a collaborative planning result of the power distribution network and the data center;

[0009] In the S2 step, the power distribution network-data center system security domain model represents a set of all working points of the system satisfying normal operation N-0 constraints and N-1 safety constraints, and quantifies the practical safety boundary of the system;

[0010] In the S3 step, the power distribution network-data center system double-layer collaborative optimization model includes an upper planning layer and a lower operation layer;

[0011] The upper planning layer optimizes a planning scheme with the minimum life cycle cost, the maximum system maximum energy supply capability index, and the minimum full-dimensional safety margin standard deviation index as targets;

[0012] The lower operation layer solves an optimal dynamic adjustment strategy with the minimum operation cost as a target, and solves the system maximum energy supply capability index.

[0013] Preferably, in the S1 step, the construction of the power distribution network-data center system security domain model including distributed power sources includes:

[0014] A working point is defined as a vector of load power compositions of all non-balance nodes in normal operation of the power distribution network, and the load power is limited within a certain range;

[0015] Normal operation N-0 constraints are established, including but not limited to power flow constraints and data center constraints;

[0016] N-1 safety constraints are established, wherein the distributed power output remains unchanged after N-1 failure, and the data center load can be reduced and transferred, and the network bandwidth and task transfer time limit are considered;

[0017] The power distribution network-data center system security domain model is constructed according to the working point, the normal operation N-0 constraint, and the N-1 safety constraint.

[0018] Preferably, the power flow constraint in the normal operation N-0 constraint includes a balance equation for calculating line and main transformer power according to downstream node load power and distributed power output, and the specific formula content is as follows:

[0019]

[0020] In the formula, P i B / P iTF are the power of line / main transformer i, i∈N; are the sets of nodes downstream of line / main transformer i respectively; P j DG is the output of all DGs connected to node j, in MW.

[0021] Preferably, the data center constraints in the normal operation N-0 constraints include:

[0022] A constraint on the total power consumption of the data center, where the total power consumption is determined by the power consumption of the servers and the power consumption of the cooling equipment, and a maximum value of the total power consumption is defined;

[0023] Server energy consumption constraints, where the server energy consumption is related to the number of servers in the power-on state and the amount of data tasks, and limits the number of servers in the power-on state and the CPU utilization of a single server;

[0024] A cooling system power consumption constraint is defined, wherein the cooling system power consumption is related to the cooling power of the data center and defines a maximum value for the cooling system power consumption.

[0025] Preferably, other constraints in the normal operation N-0 constraint include:

[0026] Line capacity constraints;

[0027] Main transformer capacity constraints;

[0028] Distributed power generation output constraints.

[0029] Preferably, the N-1 security constraints include:

[0030] The distributed generation keeps its instantaneous output unchanged after N-1 failure;

[0031] After an N-1 failure, the data center can reduce some of its load and transfer some of it to other data centers, but it is also limited by the network bandwidth capacity and task transfer delay between data centers;

[0032] N-1 line capacity constraint and N-1 main transformer capacity constraint, wherein the operating point at the N-1 fault moment is converted into the operating point at the normal operation moment through a mapping relationship.

[0033] Preferably, the quantification of the practical safety boundary of the system includes:

[0034] Define a safety upper boundary that reflects the combined energy supply capabilities of energy supply equipment, energy storage equipment, and distributed power sources, as well as the dynamic adjustment of data center load reduction and migration;

[0035] Define a lower safety boundary that reflects the minimum load requirements of the power supply equipment and takes into account the task migration characteristics of the data center;

[0036] define a safety distance, which is the shortest distance from the current working point of the system to each safety boundary;

[0037] describe the safety performance of the system by a full-dimensional safety margin vector, which is composed of the minimum distance from the working point of each time period to each safety upper boundary;

[0038] describe the balance of the system safety margin in the spatial dimension by the full-dimensional safety margin standard deviation index.

[0039] Preferably, the upper planning layer aims to minimize the life cycle cost;

[0040] The life cycle cost includes:

[0041] investment cost converted by the annualization factor;

[0042] operational cost converted by the annualization factor.

[0043] Preferably, the operational cost includes:

[0044] device operation and maintenance cost;

[0045] demand response cost;

[0046] purchasing power cost from the upper grid;

[0047] carbon emission cost;

[0048] network loss cost;

[0049] wind and light curtailment cost.

[0050] Preferably, the lower operation layer aims to minimize the operational cost, and the constraint conditions include:

[0051] safety domain constraint, i.e. ensuring that each element of the full-dimensional safety margin vector is greater than zero;

[0052] distributed power output constraint;

[0053] distribution network node voltage constraint and line current constraint;

[0054] source node injection power constraint;

[0055] grid flow constraint;

[0056] purchasing power capacity constraint;

[0057] demand response constraint, including the relationship between data center load reduction and reduction factor, and the upper and lower limits of the reduction factor;

[0058] Energy storage constraints, including charging / discharging state, charging / discharging power, energy range and sustainability constraints.

[0059] The application provides a power distribution network and data center collaborative planning method based on security domain analysis.

[0060] The application has the following beneficial effects:

[0061] 1、The application builds a power distribution network-data center system security domain model containing distributed power supply, and proposes a joint quantitative index system based on the standard deviation of the full-dimensional security margin vector and the maximum energy supply capacity of the system, which achieves the effect of continuously quantifying and representing the system security margin in space, solves the technical problems of discrete security constraints and difficulty in evaluating the global security boundary in traditional planning methods, significantly improves the system's anti-disturbance ability and operational stability, and avoids local overload risk.

[0062] 2、The application builds a double-layer collaborative optimization model of the power distribution network and data center system, and optimizes the three targets of the full-life cycle cost, the maximum energy supply capacity of the system and the standard deviation of the full-dimensional security margin in the upper layer planning, which achieves the effect of considering economy and safety in planning and decision-making, solves the technical problem that a planning scheme that simply targets economy cannot guarantee the long-term operational safety of the system, and thus realizes a significant reduction in operational cost by precisely improving safety on the premise of a small increase in total investment cost, and obtains a balance between long-term economic and safety benefits. BRIEF DESCRIPTION OF DRAWINGS

[0063] Figure 1 It is a schematic diagram of the DN-DC system of the application;

[0064] Figure 2 It is a schematic diagram of the relationship of the double-layer optimization model of the application;

[0065] Figure 3 It is a schematic diagram of the planning model solution process of the application. DETAILED DESCRIPTION

[0066] The technical solutions in the embodiments of the application will be described clearly and completely below with reference to the drawings in the specification of the application. Obviously, the described embodiments are only some of the embodiments of the application, not all. Based on the embodiments in the application, all other embodiments obtained by those skilled in the art without creative labor are within the scope of protection of the application.

[0067] Please refer to the drawings in the specification of the application Figure 1 - the drawings in the specification of the application Figure 3 The embodiments of the application provide a power distribution network and data center collaborative planning method based on security domain analysis, which includes the following steps:

[0068] S1, constructing a power distribution network-data center system security domain model containing distributed power sources;

[0069] S2, based on the constructed power distribution network-data center system security domain model, constructing a power distribution network-data center system double-layer collaborative optimization model;

[0070] S3, using an algorithm solver to solve the double-layer collaborative optimization model to obtain collaborative planning results of the power distribution network and the data center;

[0071] In the S2 step, the power distribution network-data center system security domain model represents a set of all working points of the system satisfying normal operation N-0 constraints and N-1 safety constraints, and quantifies the practical safety boundary of the system;

[0072] In the S3 step, the power distribution network-data center system double-layer collaborative optimization model includes an upper planning layer and a lower operation layer;

[0073] The upper planning layer optimizes the planning scheme with the minimum life cycle cost, maximizes the system maximum power supply capacity index, and minimizes the full-dimensional safety margin standard deviation index as the target;

[0074] The lower operation layer solves the optimal dynamic adjustment strategy with the minimum operation cost as the target, and solves the system maximum power supply capacity index.

[0075] In the above S1 step, the specific technical scheme contents are as follows:

[0076] In this embodiment, the construction of the power distribution network-data center system security domain model containing distributed power sources is one of the core steps, which aims to quantify the operation boundary and safety margin of the system. The model is based on comprehensive consideration of system operation state, normal operation constraints (N-0 constraints) and single-point fault safety constraints (N-1 safety constraints).

[0077] First, the working point is defined as the vector composed of the load power of all non-balance nodes when the power distribution network is normally operated. When the system has n non-balance nodes, the working point W can be represented as:

[0078]

[0079] Where i represents the node number, its value range is 1≤i≤n, and i∈N; n is the dimension of the working point; P i Load is the load power connected to node i. In addition, the load power P i Load is the load power connected to node i. In addition, the load power P iLoad is limited in a certain range, i.e.,

[0080]

[0081] wherein: the lower limit / upper limit.

[0082] Secondly, the normal operation N-0 constraint is established to ensure that the system can safely operate under normal load conditions. The N-0 constraint includes but is not limited to power flow constraints and data center constraints. In terms of power flow constraints, considering that the 10kV urban distribution network studied has a short power supply radius, mainly powered by load and distributed power consumption, the distribution of active power flow becomes the key. Therefore, the power flow calculation is simplified as a power balance equation. The power P i B and the power P i TF of the main transformer i can be calculated by the load power and distributed power output of its downstream nodes. The specific formula is:

[0083]

[0084] wherein, P i B and P i TF represent the power of line i and main transformer i, i∈N; and represent the set of downstream nodes of line i and main transformer i; P j DG represents the output of all distributed power connected to node j, in MW.

[0085] In terms of data center constraints, since the power consumption of servers and refrigeration equipment accounts for a high proportion of the total power consumption of the data center, the total power consumption of the data center can be represented as:

[0086]

[0087] and the maximum value of the total power consumption of the data center is limited:

[0088]

[0089] wherein, represents the total active power consumption; and represent the active power of servers and refrigeration equipment in the data center, respectively; λ DC is the energy efficiency coefficient of the data center; P DC，max ​is the maximum data processing load of the data center, which is determined by the server parameters of the data center. The subscript i represents the node, s represents the scenario, and t represents the time. Further, the calculation method of the server energy consumption is:

[0090]

[0091] wherein the number of servers in the on state and the CPU utilization of a single server are limited: respectively represent the idle power and peak power of a single server when no task is processed; ψ i is the server processing rate; is the number of servers in the on state at time t; W i,s,t is the data task amount of the data center; is the number of server configurations; φ is the server standby coefficient; S max is the upper limit of the CPU utilization of a single server. The operating characteristics of the refrigeration system power consumption can be represented as:

[0092]

[0093] The maximum value of the refrigeration system power consumption is limited:

[0094] wherein, is the air conditioner electric power; is the data center refrigeration power; λ Air is the energy efficiency coefficient of the refrigeration equipment; is the refrigeration equipment capacity configured by the data center. In addition to the above power flow constraints and data center constraints, normal N-0 operation also includes other constraints, such as line capacity constraints, main transformer capacity constraints, and distributed power output constraints.

[0095] The line capacity constraint is represented as:

[0096] The main transformer capacity constraint is represented as:

[0097] The distributed power output constraint is represented as:

[0098] wherein, and are the capacities of the line and the main transformer, respectively; B and T are the sets of all lines and main transformers; is the distributed power output; is the predicted value of the distributed power; is the prediction error.

[0099] ​​Thirdly, an N-1 safety constraint is established, requiring the system to maintain power supply even in the event of a single point failure. The N-1 safety constraint takes into account the following situation: the distributed generation maintains its instantaneous output unchanged after an N-1 failure to ensure the available capacity of the network. Its expression is:

[0100]

[0101] Where, is the output of DG after the system N-1 failure; P t DG Contribute to the normal operation of DG. After the N-1 failure, the data center can reduce some of the load and transfer part of it to other data centers to ensure the continuity of power supply. Its load adjustment is expressed as:

[0102]

[0103] Among them, α i is the ratio of data center load reduction of node i, 0≤α i ≤1;β j,i is the proportion of task load transferred from data center j to i, and ∑ j≠i β j,i ≤α i , The load transfer is limited by the load reduction ratio. To ensure that the data center is i is the ratio of data center load reduction of node i, 0≤α i ≤1;β j,i is the proportion of task load transferred from data center j to i, and ∑ j≠i β j,i ≤α i , Load transfer is limited by the load reduction ratio. To ensure that the network bandwidth between data centers does not exceed the capacity and latency limits of each data center during load transfer, the network bandwidth and task transfer time are limited as follows:

[0104]

[0105]

[0106] Where B j,i and T j,i are the network bandwidth capacity and task transfer delay from data center j to data center i, respectively. In addition, the N-1 safety constraint also includes the N-1 line capacity constraint and the N-1 main transformer capacity constraint.

[0107] Since the variable studied in this method is the working point W at the N-0 fault moment, and the model needs to take into account the N-1 fault moment W t+1safety constraints, so W t+1 is transformed into W. This transformation is described by a mapping h, where h: W→ W t+1 Unlike the case of W=W t+1 in traditional N-1 security analysis, it is not an identity mapping when considering data center demand response, and the specific relationship is:

[0108]

[0109] where E is the identity matrix; and a is a vector of load shedding proportion of each interruptible load. After element d experiences an N-1 fault, the distribution network will perform network topology reconfiguration to restore power supply in the non-fault area, and the power balance equation will also change accordingly.

[0110] The N-1 line capacity constraint is:

[0111]

[0112] where ψ d is the fault scenario of element d; is the power of other loads except the data center after element d experiences an N-1 fault; is the power of line Bi after element d experiences an N-1 fault; is the energy storage power configured for the data center; is the set of downstream nodes of Bi. The N-1 main transformer capacity constraint is:

[0113]

[0114] where ψ is the power of main transformer Ti after element d experiences an N-1 fault; is the set of downstream nodes of Ti. Let the fault set be If the above N-1 line capacity constraint and N-1 main transformer capacity constraint are both satisfied for a working point W under , then W satisfies the N-1 security criterion.

[0115] Finally, the security domain model Ω DN-DC of the distribution network-data center system is constructed as the set of all working points in the state space that satisfy the N-0 security and N-1 security criteria. This model can be represented as:

[0116] Finally, the security domain model Ω DN-DC of the distribution network-data center system is constructed as the set of all working points in the state space that satisfy the N-0 security and N-1 security criteria. This model can be represented as:

[0117]

[0118] At the same time, the following conditions are met:

[0119]

[0120] For At the same time, the following conditions are met:

[0121]

[0122] In the formula, Θ is a bounded set composed of all reasonable range working points, that is, the state space. This safety domain model effectively characterizes the safe operation range of the system under normal and fault conditions, providing a basis for subsequent collaborative optimization.

[0123] On this basis, the embodiment further quantifies the practical safety boundary of the system, and defines the maximum energy supply capacity index of the system.

[0124] Among them, the practical safety boundary of the power distribution network-data center system represents all critical operating points of the system under the premise of meeting the N-1 safety check. The boundary is determined by the system structure and key device parameters, and is composed of a group of hyperplanes. When a fault occurs, the data center can carry out demand response, and the energy storage device can be used as a temporary energy supply resource to ensure that the system does not lose load as much as possible. The general expression of the practical safety boundary is as follows:

[0125]

[0126] In the formula, B i represents the i-th energy supply line; is the safe upper boundary of the feeder, reflecting the maximum energy supply capacity of the energy supply line; is the safe lower boundary of the feeder, indicating the minimum load that the feeder can maintain after load reduction, reflecting the minimum load requirement of the line, which is mainly affected by the minimum load requirement of the device and the load transfer characteristics of the data center;C U is the capacity upper limit of the feeder;C L is the minimum load of the feeder;μ i , μ j is a fixed coefficient for adjusting the weight of load transfer or reduction;P i is the load size of each outlet pipeline.

[0127] For the power distribution network-data center system, the safe upper boundary reflects the comprehensive energy supply capacity of the energy supply device, the energy storage device and the distributed power supply, as well as the dynamic adjustment of the data center load reduction and migration, which can be expressed as:

[0128]

[0129] In the formula, C Uthe capacity upper limit of the energy supply device; the ellipsis (...) indicates other constraints that can exist. The lower safety boundary reflects the minimum load requirement of the energy supply device, i.e., to ensure that the device can operate normally, and in combination with the characteristics of data center task migration, is expressed as:

[0130]

[0131] where C L is the minimum load limit.

[0132] The full-dimensional observation of the power distribution network-data center system adopts an indirect observation method based on a safety distance. The safety distance is defined as the shortest distance from the current working point of the system to each safety boundary. When the working point is located within the safety domain, this distance is positive; when the working point is located outside the safety domain, the distance is negative, and can be expressed as:

[0133]

[0134] The safety performance of the system is described using a full-dimensional safety margin (FDSM) vector for the working point of the system in 24 time periods of a typical day, expressed as:

[0135]

[0136] where D FDSM is a vector composed of the minimum values of the distances from the working point of each time period to each safety upper boundary, and is defined as the FDSM; is the minimum value of the distance from the working point of each time period to the safety upper boundary , and has a unit of MW; represents the distance from the working point of the t time period to , and has a unit of MW.

[0137] The random fluctuation characteristics of the load can cause the working point of the system to approach or cross any safety upper boundary. To overcome this limitation, it is necessary to ensure that the numerical distribution of each element in the FDSM vector is balanced. The standard deviation index σ FDSM of the FDSM is used in the present embodiment to describe the balance of the safety margin of the system in the spatial dimension, and the closer the value is to 0, the more balanced the safety margin. The FDSM standard deviation constructed based on the above formula is as follows:

[0138]

[0139] where d SDPLU,av represents the average value of each element of the FDSM of the system, and has a unit of MW.

[0140] In addition, the total supply capability index of the system, denoted as TSC, represents the maximum value of the supply of the system under the N-1 criterion. The calculation can be expressed as the following optimization problem:

[0141]

[0142] wherein ∑P i is the total load of the export supply, in MW; P min is the upper limit of the system operating point; P max is the lower limit of the system operating point; P max is the state space of the system operating point; P min respectively represent the upper limit / lower limit constraints of the system operating point, and the unit of each element in the vector is MNW; h(P) = 0 represents the normal operation constraint of the system; w(P) ≤ 0 represents the N-1 safety constraint. The solution of the optimization problem determines the maximum load that the system can provide under the premise of ensuring safety.

[0143] In the above S2 step, the specific technical solution contents are as follows:

[0144] In the embodiment, the distribution network-data center system double-layer collaborative optimization model aims to collaboratively optimize the planning and operation of the distribution network and the data center to achieve the balance between economy and safety. The model includes an upper planning layer and a lower operation layer.

[0145] In the embodiment, based on the constructed distribution network-data center system safety domain model, a distribution network-data center system double-layer collaborative optimization model is established. The model includes an upper planning layer and a lower operation layer. The configuration parameters are passed from the upper planning layer to the lower operation layer, and the optimization results of the operation layer are fed back to the upper planning layer to calculate the objective function.

[0146] The upper planning layer is responsible for optimization planning from a macro perspective, and makes decisions on the configuration capacity and installation location of the data center and the distributed power supply with the multiple objectives of minimum life cycle cost, maximum system supply capability, and minimum standard deviation of full-dimensional safety margin (FDSM).

[0147] The primary objective of the model is to minimize the life cycle cost of the system, which is composed of the annualized investment cost and the annual operation cost.

[0148] F1 = min(C I +C O );

[0149] wherein F1 is the economic objective function; C I is the annualized investment cost; C O is the annual operation cost.

[0150] The annualized investment cost C IThe investment of all new or reconstruction equipment such as lines, main transformers, distributed power and data centers is covered.

[0151] C I = C I,Line + C I,TF + C I,DG + C I,DC ;

[0152] In the formula, C I,Line is the annualized investment cost of lines; C I,TF is the annualized investment cost of main transformer expansion; C I,DG is the annualized investment cost of distributed power; and C I,DC is the annualized investment cost of data center.

[0153] Each investment cost is converted to each year through the equipment annualization factor, and the specific calculation is as follows:

[0154]

[0155] In the formula, b is the equipment annualization factor; Ω Line is the set of candidate installation nodes of the system line; Φ Line is the set of candidate investment types of the line; is the unit length investment cost of the line of type m; L ij is the length of the line ij; is the 0-1 decision variable of whether the line ij invests in type m.

[0156] C I,TF = bc TF P TF-add ;

[0157] In the formula, b is the equipment annualization factor; c TF is the configuration cost of the unit capacity transformer; P TF-add is the expansion capacity of the transformer.

[0158]

[0159] In the formula, b is the equipment annualization factor; Ω PV is the set of candidate installation nodes of photovoltaic; c PV is the configuration cost of a single photovoltaic; is the installation number of photovoltaic; Ω WT is the set of candidate installation nodes of wind turbine; c WT is the configuration cost of a single wind turbine; is the installation number of wind turbine.

[0160]

[0161] Where b is the annualization factor of equipment; Ω DC is the set of candidate installation nodes of data center; c Ser is the configuration cost of a single server; is the installation number of servers; c Air is the configuration cost of a single refrigeration equipment; is the installation number of refrigeration equipment; c ESS is the configuration cost of a single energy storage; is the installation number of energy storage.

[0162] Where the calculation method of the annualization factor b of equipment is as follows:

[0163]

[0164] Where, is the discount rate; y is the service life of equipment. The annual operation cost C O Then the comprehensive consideration is given to the equipment operation and maintenance, demand response, electricity purchase from the upper grid, carbon emission, network loss and wind and light abandonment.

[0165] C O = C O,Eq + C O,DR + C O,Pur + C O,C + C O,Loss + C O,Waste ;

[0166] Where, C O,Eq is the equipment operation and maintenance cost; C O,DR is the demand response cost; C O,Pur is the electricity purchase cost; C O,C is the carbon emission cost; C O,Loss is the network loss cost; C O,Waste is the wind and light abandonment cost.

[0167] The calculation of the operation cost is as follows:

[0168] The equipment operation and maintenance cost C O,Eq is composed of transformer operation and maintenance cost, DG operation and maintenance cost and DC operation and maintenance cost.

[0169]

[0170]

[0171] Where, Λ TF / Λ PV / Λ WT / Λ DC are the sets of nodes where transformers / PVs / WTs / DCs are installed; f O,PV / fO,WT respectively, the operation and maintenance cost of PV / WT per unit of electricity; respectively, the active power of PV / WT injection node i in period t; O,TF O,Ser O,Air O,ESS respectively, the operation and maintenance cost of transformer, data center server, data center refrigeration equipment, and data center energy storage. Δt is the duration of period t, all of which are 1h.

[0172] DR cost C O,DR The DR cost is reflected by reducing the load node and its corresponding compensation cost.

[0173]

[0174] In the formula, Λ DC,cut is the set of reducible load nodes;f O,DC,cut is the load reduction compensation per unit of electricity; ΔP i , s, t is the active power size of the reduced load after DR.

[0175] Power purchase cost C O,Pur The current power purchase cost considers time-of-use electricity price, as shown in the following calculation formula:

[0176]

[0177] In the formula, D L (1, :) is the set of end nodes of the branch with source node 1 as the first end node; is the time-of-use electricity price in period t; P 1j,s,t is the active power of the line connected to source node 1 flowing into the current power grid.

[0178] Carbon emission cost C O,C Then, it is calculated by the following formula:

[0179]

[0180] In the formula, is the system carbon emission fee corresponding to period t, e TPG is the external grid carbon emission coefficient, is the grid power purchase power.

[0181] Regarding network loss cost C O,Loss , although it has been included in the upper grid power purchase cost, in order to better optimize the network loss, the penalty cost of network loss is used as the representation.

[0182]

[0183] In the formula, P​​​s t Loss,Net is the network loss size for time period t; R i j is the resistance size of line ij; I i j,s,t 2 is the current size of line ij for time period t, Ω DL is the DC branch set.

[0184] abandoned wind and light cost C O,Waste is the abandoned wind and light cost caused by the inability to fully utilize wind power and photovoltaic power generation, so the difference between the predicted power output and the actual demand is calculated as follows.

[0185]

[0186] In the formula, f O,WT,cut / f O,PV,aut are the abandoned wind / light cost per unit of electricity; are the sum of the predicted active power output of the installed PV / WT at node i for time period t.

[0187] In addition to economy, the second goal of the model is to maximize the system energy supply capacity index T TSC to improve the reliability of the system in extreme cases.

[0188] F2 = max(T TSC );

[0189] In the formula, F2 is the reliability objective function; T TSC is the maximum energy supply capacity of the system under the N-1 criterion.

[0190] Finally, to ensure the balance and safety of system operation, the third goal aims to minimize the standard deviation of the full-dimensional safety margin (FDSM), thereby balancing the safety margin of each line.

[0191]

[0192] In the formula, F3 is the balance objective function; d i is the number of days of the i-th typical day.

[0193] To ensure the feasibility of the planning scheme, the upper model also needs to meet a series of constraint conditions. First, the investment and installed capacity of various types of equipment cannot exceed its upper limit, which includes the expansion capacity of transformers, the capacity of data centers and their internal equipment, and the number of installed distributed power sources.

[0194] 0 ≤ P TF-add ≤ P TF,max ;

[0195] where P TF-add is the transformer expansion capacity; P TF,max is the upper limit of the transformer installation capacity.

[0196]

[0197] where P is the data center capacity installed at node i; is the upper limit of the data center capacity at node i.

[0198]

[0199] where P is the number of servers installed at node i; is the upper limit of the number of servers.

[0200]

[0201] where P is the number of refrigeration equipment installed at node i; is the upper limit of the number of refrigeration equipment.

[0202]

[0203] where P is the number of energy storage installed at node i; is the upper limit of the number of energy storage.

[0204]

[0205] where P is the number of photovoltaic installed at node i; is the upper limit of the number of photovoltaic; Ω PV,max is the upper limit of the number of photovoltaic; Ω PV is the set of photovoltaic candidate installation nodes. is the number of wind turbines installed at node i; is the upper limit of the number of wind turbines; Ω WT is the set of wind turbine candidate installation nodes.

[0206] For the investment and construction of lines, each candidate location can only choose one type of line for construction at most:

[0207]

[0208] where P is the 0-1 decision variable of whether to invest in line ij of type m; Ω Line is the set of system line candidate installation nodes.

[0209] The lower layer receives the planning scheme from the upper layer and establishes a steady-state optimization model. Its core task is to minimize the total operation cost of the system by scheduling the active control devices in the system, to solve the optimal dynamic adjustment strategy, and to feed back the calculated operation cost and maximum power supply capacity of the system to the upper layer.

[0210] To achieve this goal and ensure the safe and stable operation of the system under various scenarios, the lower layer model must follow a series of strict constraint conditions. The first is the safety domain constraint, i.e. the system still needs to maintain safe operation under N-1 failure, which is manifested as the distance of each period working point to the safety boundary must be greater than zero.

[0211]

[0212] In the formula, is the minimum distance of each period working point to the safety upper boundary. The actual output of the distributed power supply cannot exceed its predicted maximum output at the current time.

[0213]

[0214] In the formula, and are the actual active power output of photovoltaic and wind turbine at t period; P i ,s,t PV,max and are the predicted maximum active power output of photovoltaic and wind turbine at t period; Λ PV and Λ WT are the node set where photovoltaic and wind turbine are installed.

[0215] The operation of the distribution network itself needs to meet a series of physical constraints. The voltage of each node must be maintained within the safe range, and the line current cannot exceed its thermal stability limit.

[0216]

[0217] In the formula, is the voltage square of node i at t period; U i ,min 2 and are the lower and upper limits of the node voltage square; Ω DN is the set of all distribution network nodes; is the current square of line ij at t period; is the safety current square of line ij; Ω DL is the set of DC branches.

[0218] The exchange power of the source node with the upper grid should also be within the specified limit.

[0219]

[0220] P 1,j,s,t is the active power of the line connected source node 1 flowing into the current level power grid in period t; and are the lower limit and upper limit of the active power injected by the source node into the power grid, respectively; is the set of end nodes of the branch with the source node 1 as the head node.

[0221] In addition, the power of the entire network must be balanced, i.e. the injected power is equal to the consumed power.

[0222]

[0223] P is the power purchased by the power grid; P i ,s,t PV is the active power of the photovoltaic; P i ,s,t WT is the active power of the wind turbine; P i ,s,t BSS,down is the discharging power of the energy storage; P i ,s,t ESS,up is the charging power of the energy storage; P i ,s,t Load is the original load power; ΔP i ,s,t L0ad is the load power reduced by demand response. The amount of electricity purchased from the upper-level power grid is limited by the total capacity of the transformer.

[0224] 0≤P t Trans ≤(P TF-add +P TF-0 );

[0225] P t Trans is the power purchased at time t; P TF-add is the expansion capacity of the transformer; P TF-0 is the original capacity of the transformer.

[0226] As a flexible load, the data center can participate in demand response. The amount of load reduction is determined by the reduction coefficient, which can be adjusted within a certain range.

[0227]

[0228] P is the reduced load active power; v i,s,t is the data center load reduction coefficient; is the active power consumption of the data center; Λ DC,cuta set of data center nodes that can participate in demand response; v min and v max are the lower and upper limits of the load reduction factor, respectively.

[0229] The operation of the energy storage system involves multiple constraints, including the mutual exclusivity of the charge and discharge states, the upper limit of the charge and discharge power, the dynamic change of the energy storage capacity, the energy conservation of the operation cycle, and the upper and lower limits of the state of charge, etc.

[0230]

[0231] wherein, and are 0-1 state variables of the energy storage charging and discharging, respectively; and are the energy storage charging and discharging power, respectively; is the upper limit of the energy storage charging and discharging power; is the energy storage capacity at time t; η up and η down are the charging and discharging efficiencies, respectively; Δt is the time period length; and are the lower and upper limits of the energy storage capacity, respectively.

[0232] In the above S3 step, the specific technical solution contents are as follows:

[0233] This step aims to solve the power distribution network-data center system double-layer collaborative optimization model to obtain the collaborative planning results of the power distribution network and the data center. The solving process is as follows: the double-layer collaborative optimization model is composed of the upper planning layer and the lower operation layer, and the solving process is an iterative optimization process. The upper planning layer adopts the improved normalization method plane constraint method for multi-objective optimization, and the lower operation optimization and the solution of the maximum energy supply capacity of the system are completed by calling the commercial solver Gurobi.

[0234] When the specific solving process is started, the upper planning layer first initializes a population, which is composed of multiple individuals, each individual representing a selected planning scheme. Each planning scheme includes a set of determined decision variables, specifically the configuration capacity and installation location of the data center and the distributed power, covering the line investment decision variable, the main transformer expansion capacity, the number of installed photovoltaic, the number of installed wind turbine, the number of installed server, the number of installed refrigeration equipment, and the number of installed energy storage.

[0235] Subsequently, based on the constructed Utopia hyperplane, the upper planning layer converts the original multi-objective problem into a set of single-objective optimization sub-problems with additional constraints. Each sub-problem aims to minimize the life-cycle cost Fl, but with additional constraints that limit the solution space on a specific normal line from the Utopia hyperplane. By adjusting the position of the normal line, the entire Pareto frontier can be systematically explored.

[0236] For each single-objective sub-problem, the solution process is as follows:

[0237] The upper optimizer adjusts a set of planning decision variables, including line investment decisions, main transformer expansion capacity, and the number of data center and distributed power configurations. For each given set of planning decision variables, it is passed as a fixed parameter to the lower operation layer. After receiving the parameters, the lower operation layer uses the Gurobi solver to perform two calculations: one is to minimize the annual operation cost under all operation constraints, and the other is to maximize the system maximum power supply capacity T TSC and calculate the specific value of this indicator.

[0238] After the lower layer calculation is completed, the obtained annual operation cost C O and the system maximum power supply capacity indicator T TSC are returned to the upper planning layer. In the upper layer, combined with the known annualized investment cost, the three objective function values corresponding to this planning scheme are calculated: the life-cycle cost Fl, the system maximum power supply capacity F2, and the standard deviation of the full-dimensional safety margin (FDSM) F3.

[0239] Fl = min (C I + C O );

[0240] In the formula, C I is the annualized investment cost; C O is the annual operation cost.

[0241] F2 = max (T TSC );

[0242] In the formula, T TSC is the maximum power supply capacity of the system under the N-1 criterion.

[0243]

[0244] In the formula, is the variance of the system full-dimensional safety margin (FDSM); d i is the number of the i-th typical day.

[0245] The upper planning layer obtains a series of effective solutions uniformly distributed on the Pareto frontier by solving these decomposed single-objective sub-problems with different additional constraints. The set of these solutions directly constitutes the final Pareto frontier solution set, each solution in the solution set being a Pareto optimal planning scheme, representing a specific trade-off relationship between the three objectives of economy, power supply capacity and safety.

[0246] Finally, the final collaborative planning result is selected from the output Pareto frontier solution set. A balanced decision-making method is adopted, specifically: firstly, the three objective function values of all schemes in the solution set are normalized, then the Euclidean distance of each scheme to the ideal point is calculated, and the scheme with the minimum distance is selected as the final collaborative planning result with balanced target performance.

[0247] Although the embodiments of the present application have been shown and described, it is to be understood that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the present application, the scope of which is defined by the appended claims and their equivalents.

Claims

1. A distribution network and data center collaborative planning method based on security domain analysis, characterized in that: The following steps are involved: S1. Construct a security domain model for the distribution network-data center system including distributed power sources; S2. Based on the constructed distribution network-data center system security domain model, a distribution network-data center system two-layer collaborative optimization model is constructed; S3. Solve the two-layer collaborative optimization model using an algorithm solver to obtain a collaborative planning result of the distribution network and the data center; In step S2, the distribution network-data center system security domain model represents the set of all operating points where the system satisfies the normal operation N-0 constraint and the N-1 safety constraint, and quantifies the practical safety boundary of the system; In step S3, the distribution network-data center system two-layer collaborative optimization model includes an upper planning layer and a lower operation layer; The upper planning layer optimizes the planning scheme with the goals of minimizing the full life cycle cost, maximizing the system's maximum energy supply capability index, and minimizing the full-dimensional safety margin standard deviation index; The lower operation layer solves the optimization dynamic adjustment strategy with the goal of minimizing the operation cost, and solves the maximum energy supply capacity index of the system.

2. A method for collaborative planning of distribution networks and data centers based on security domain analysis according to claim 1, characterized in that: In step S1, the construction of a distribution network-data center system security domain model including distributed power sources includes: The working point is defined as the vector consisting of the load power of all unbalanced nodes when the distribution network is operating normally, and the load power is limited to a specific range; Establish normal operation N-0 constraints, including but not limited to power flow constraints and data center constraints; Establish N-1 safety constraints, which take into account the fact that distributed generation output remains unchanged after N-1 failures and that data center loads can be reduced and transferred, and consider network bandwidth and task transfer time constraints; The distribution network-data center system security domain model is constructed based on the working point, normal operation N-0 constraints and N-1 safety constraints.

3. The method for collaborative planning of distribution network and data center based on security domain analysis according to claim 2 is characterized in that: The power flow constraint in the normal operation N-0 constraint includes a balance equation for calculating the power of the line and main transformer based on the downstream node load power and the output of the distributed generation. The specific formula is as follows: Where, are the power of line / main transformer i, i∈N; are the sets of nodes downstream of line / main transformer i respectively; is the output of all DGs connected to node j, in MW.

4. The method for collaborative planning of distribution network and data center based on security domain analysis according to claim 2, characterized in that: The data center constraints in the normal operation N-0 constraints include: A constraint on the total power consumption of the data center, where the total power consumption is determined by the power consumption of the servers and the power consumption of the cooling equipment, and a maximum value of the total power consumption is defined; Server energy consumption constraints, where the server energy consumption is related to the number of servers in the power-on state and the amount of data tasks, and limits the number of servers in the power-on state and the CPU utilization of a single server; A cooling system power consumption constraint is defined, wherein the cooling system power consumption is related to the cooling power of the data center and defines a maximum value for the cooling system power consumption.

5. The method for collaborative planning of distribution network and data center based on security domain analysis according to claim 2, characterized in that: Other constraints in the normal operation N-0 constraints include: Line capacity constraints; Main transformer capacity constraints; Distributed power generation output constraints.

6. The method for collaborative planning of distribution network and data center based on security domain analysis according to claim 2, characterized in that: The N-1 safety constraints include: The distributed generation keeps its instantaneous output unchanged after N-1 failure; After an N-1 failure, the data center can reduce some of its load and transfer some of it to other data centers, but it is also limited by the network bandwidth capacity and task transfer delay between data centers; N-1 line capacity constraint and N-1 main transformer capacity constraint, wherein the operating point at the N-1 fault moment is converted into the operating point at the normal operation moment through a mapping relationship.

7. The method for collaborative planning of distribution network and data center based on security domain analysis according to claim 1, characterized in that: The practical safety boundaries of the quantified system include: Define a safety upper boundary that reflects the combined energy supply capabilities of energy supply equipment, energy storage equipment, and distributed power sources, as well as the dynamic adjustment of data center load reduction and migration; Define a lower safety boundary that reflects the minimum load requirements of the power supply equipment and takes into account the task migration characteristics of the data center; Define the safety distance, which is the shortest distance from the current working point of the system to each safety boundary; The system safety performance is described using a full-dimensional safety margin vector, which is composed of the minimum value of the distance from the working point to each safety upper boundary in each time period; The full-dimensional safety margin standard deviation index is used to describe the balance of the system safety margin in the spatial dimension.

8. The method for collaborative planning of distribution network and data center based on security domain analysis according to claim 1, characterized in that: The upper planning layer aims to minimize the life cycle cost; The life cycle costs include: The investment cost obtained by annualizing the factors; Operating costs converted using an annualized factor.

9. The method for collaborative planning of distribution network and data center based on security domain analysis according to claim 8, characterized in that: The operating costs include: Equipment operation and maintenance costs; demand response costs; Cost of purchasing electricity from the higher-level power grid; Carbon emission costs; Network loss costs; The cost of curtailing wind and solar power.

10. The method for collaborative planning of distribution network and data center based on security domain analysis according to claim 1, characterized in that: The lower operation layer aims to minimize the operating cost, and its constraints include: Safety region constraint, i.e. ensuring that all elements of the full-dimensional safety margin vector are greater than zero; Distributed power generation output constraints; Distribution network node voltage constraints and line current constraints; The source node injects power constraints into the grid; Grid power flow constraints; Power purchase capacity constraints; Demand response constraints, including the relationship between data center load reduction and the reduction factor, as well as the upper and lower limits of the reduction factor; Energy storage constraints include the charge / discharge state, charge / discharge power, energy range, and sustainability of energy storage.

Images