Urban power grid massive micro-resource partition stability control method, device, equipment, storage medium and program product

Abstract

This application relates to a method, apparatus, device, storage medium, and program product for zoned stability control of massive micro-resources in urban power grids. The method includes: performing equivalent modeling of the urban power grid based on its grid structure coupling characteristics and power balance index values ​​to obtain an equivalent model; dividing the urban power grid into multiple control zones based on the equivalent model; determining the importance score of each grid resource according to its resource characteristics within each control zone; classifying the grid resources according to their importance scores to obtain classified grid resources; constructing a multi-objective optimization function for the classified grid resources within each control zone, including new energy consumption targets and frequency stability targets; and generating a stability control strategy for the classified grid resources using the multi-objective optimization function. This method enables accurate coordinated control of massive micro-resources in urban power grids.

Description

Technical Field

[0001] This application relates to the field of smart grid technology, and in particular to a method, apparatus, computer equipment, computer-readable storage medium, and computer program product for the zoned stability control of massive micro-resources in urban power grids. Background Technology

[0002] With the acceleration of urbanization, the scale of mega-city power grids is growing rapidly, and load density is constantly increasing. At the same time, in response to the goal of green and low-carbon energy transition, massive amounts of micro-resources such as distributed photovoltaics, energy storage systems, and interruptible loads are being connected to urban distribution networks on a large scale. These micro-resources are characterized by their huge quantity, geographical dispersion, small individual capacity but large total amount, and significant differences in regulation characteristics.

[0003] In traditional technologies, the control of urban power grids often adopts a centralized unified control mode. However, with the rapid growth of the scale of micro-resources such as distributed photovoltaics, energy storage, and interruptible loads in ultra-large urban power grids, the traditional centralized unified control mode faces challenges such as the large scale of resources, geographical dispersion, and large differences in characteristics, making it difficult to achieve precise and rapid control. Therefore, there is a problem that it is impossible to accurately coordinate the control of the massive micro-resources in urban power grids. Summary of the Invention

[0004] Therefore, it is necessary to provide a method, device, computer equipment, computer-readable storage medium, and computer program product for the zoned stability control of massive micro-resources in urban power grids, which can solve the above-mentioned technical problems.

[0005] Firstly, this application provides a method for zonal stability control of massive micro-resources in urban power grids, including:

[0006] The urban power grid is modeled using equivalent values ​​based on its grid structure coupling characteristics and power balance index, resulting in an equivalent model. Based on the equivalent model, the urban power grid is divided into multiple control zones.

[0007] Based on the resource characteristics corresponding to each power grid resource within the control zone, an importance score is determined for each power grid resource, and each power grid resource is classified according to the importance score to obtain classified power grid resources;

[0008] For the classified power grid resources within the control zone, a multi-objective optimization function is constructed, which includes the renewable energy consumption target and the frequency stability target. Furthermore, a stability control strategy for the classified power grid resources is generated through the multi-objective optimization function.

[0009] In one embodiment, the equivalent modeling of the urban power grid based on the grid structure coupling characteristics and power balance index values ​​to obtain the equivalent model includes:

[0010] For the urban power grid, the urban power grid is divided into an internal system, boundary nodes and an external system. A block matrix of node voltage equations is constructed. By eliminating the internal nodes in the block matrix, an equivalent network equation containing the self-admittance of the boundary nodes, the mutual admittance between the boundary nodes and the external system and the equivalent injected current is constructed as the equivalent model of the urban power grid.

[0011] For the generators in the urban power grid, based on the generator power angle difference after disturbance, identify the co-tuning units with consistent power angle curves, aggregate the co-tuning units into equivalent generators, and construct the equivalent model corresponding to the generator based on the equivalent capacity, equivalent inertial time constant and equivalent transient reactance of the equivalent generators.

[0012] For the static load in the urban power grid, the proportional parameters of constant impedance, constant current and constant power are calculated by power weighted average, and an equivalent model corresponding to the static load is constructed.

[0013] For the dynamic load in the urban power grid, the equivalent motor parameters are calculated using capacity-weighted average to construct the equivalent model corresponding to the dynamic load.

[0014] In one embodiment, the division of the urban power grid into multiple control zones based on the equivalent model includes:

[0015] Construct a power grid weighted graph model; the set of nodes in the power grid weighted graph model is used to represent busbars or substations, the set of edges in the power grid weighted graph model is used to represent transmission lines or transformers, and the weights of the edges in the power grid weighted graph model are used to represent the electrical coupling strength between nodes;

[0016] Modularity is used as a quantitative indicator of the partition quality of the control partition. The Louvain algorithm is used to perform community detection on the power grid weighted graph model. Local optimization and network aggregation are performed iteratively until the modularity reaches the maximum value, thus obtaining the multiple control partitions.

[0017] In one embodiment, determining the importance score for each power grid resource based on its resource characteristics within the control partition includes:

[0018] Based on the resource type characteristics, resource user importance characteristics, voltage level characteristics, response speed characteristics, and interruption resource quantity characteristics of the power grid resources, multiple resource characteristics of the power grid resources are constructed.

[0019] The importance score corresponding to each of the resource characteristics is determined by weighted summation; the weight of each resource characteristic is adjusted according to the operation scenario of the urban power grid.

[0020] In one embodiment, constructing a multi-objective optimization function that includes a renewable energy consumption target and a frequency stability target includes:

[0021] Determine the decision variables; the decision variables include the charging and discharging power sequence of the energy storage system, the curtailment power sequence of new energy sources, and the power reduction sequence of interruptible loads;

[0022] Construct a first objective function, which is used to maximize the consumption of new energy sources;

[0023] Construct a second objective function to maximize frequency stability;

[0024] Based on the first objective function and the second objective function, a multi-objective optimization function is constructed to optimize the decision variables under the condition of satisfying constraints; the constraints include at least power balance constraints, thermal stability control and equipment safety constraints, and regulation resource constraints.

[0025] In one embodiment, generating a stability control strategy for the classified power grid resources through the multi-objective optimization function includes:

[0026] The comprehensive power grid disturbance index value is obtained, which is determined based on the frequency deviation rate and the frequency change rate.

[0027] The Sigmoid function is used to map the comprehensive power grid disturbance index value into a dynamic weighting factor;

[0028] The first objective function and the second objective function are weighted and merged into a single objective function using the dynamic weighting factor.

[0029] By solving the single objective function, a stable control strategy for the classified power grid resources is generated.

[0030] The smaller the comprehensive power grid disturbance index value, the more the dynamic weighting factor makes the optimization objective of the single objective function more biased towards new energy consumption; the larger the comprehensive power grid disturbance index value, the more the dynamic weighting factor makes the optimization objective of the single objective function more biased towards frequency stability.

[0031] Secondly, this application also provides a zoned stability control device for massive micro-resources in urban power grids, comprising:

[0032] The modeling and partitioning module is used to perform equivalent modeling of the urban power grid based on the grid structure coupling characteristics and power balance index values ​​of the urban power grid, to obtain the equivalent model, and to divide the urban power grid into multiple control partitions based on the equivalent model;

[0033] The feature classification module is used to determine the importance score of each power grid resource based on the resource characteristics of each power grid resource in the control partition, and to classify each power grid resource according to the importance score to obtain the classified power grid resources.

[0034] The stability control module is used to construct a multi-objective optimization function that includes a new energy consumption target and a frequency stability target for the classified grid resources within the control zone, and to generate a stability control strategy for the classified grid resources through the multi-objective optimization function.

[0035] Thirdly, this application also provides a computer device, including a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to implement the steps of the above-described method.

[0036] Fourthly, this application also provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the steps of the above-described method.

[0037] Fifthly, this application also provides a computer program product, including a computer program that, when executed by a processor, implements the steps of the above-described method.

[0038] The aforementioned method, apparatus, computer equipment, computer-readable storage medium, and computer program product for zonal stability control of massive micro-resources in urban power grids perform equivalent modeling of the urban power grid based on the grid structure coupling characteristics and power balance index values, obtaining an equivalent model. Based on the equivalent model, the urban power grid is divided into multiple control zones. According to the resource characteristics corresponding to each power grid resource within the control zone, the importance score corresponding to each power grid resource is determined. Based on the importance score, each power grid resource is classified, obtaining classified power grid resources. For the classified power grid resources within the control zone, a multi-objective optimization function including new energy consumption targets and frequency stability targets is constructed. Through the multi-objective optimization function, a stability control strategy for the classified power grid resources is generated. Thus, for ultra-large urban power grids integrating distributed photovoltaic, energy storage, and interruptible load micro-resources, through zonal equivalence, resource classification, and multi-level coordinated control, the massive micro-resources of the urban power grid can be accurately controlled collaboratively, improving grid stability and resource utilization efficiency. Attached Figure Description

[0039] To more clearly illustrate the technical solutions in the embodiments of this application or related technologies, the drawings used in the description of the embodiments of this application or related technologies will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other related drawings can be obtained based on these drawings without creative effort.

[0040] Figure 1 This is an application environment diagram of a zonal stability control method for massive micro-resources in an urban power grid, as shown in one embodiment.

[0041] Figure 2 This is a flowchart illustrating a method for zonal stability control of massive micro-resources in an urban power grid, as shown in one embodiment.

[0042] Figure 3 This is a flowchart illustrating a method for zonal stability control of massive micro-resources in an urban power grid, as described in another embodiment.

[0043] Figure 4 This is a structural block diagram of a zoned stability control device for massive micro-resources in an urban power grid, as shown in one embodiment.

[0044] Figure 5 This is an internal structural diagram of a computer device in one embodiment. Detailed Implementation

[0045] To make the objectives, technical solutions, and advantages of this application clearer, the following detailed description is provided in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the scope of this application.

[0046] The zonal stability control method for massive micro-resources in urban power grids provided in this application embodiment can be applied to, for example... Figure 1 In the application environment shown, terminal 102 communicates with server 104 via a network. A data storage system can store the data that server 104 needs to process. The data storage system can be integrated onto server 104, or it can be located in the cloud or on another network server.

[0047] The terminal 102 may be, but is not limited to, various personal computers, laptops, smartphones, tablets, IoT devices, etc.

[0048] Server 104 can be a standalone physical server, a server cluster or distributed system consisting of multiple physical servers, or a cloud server providing cloud computing services. For example, server 104 can be a power system dispatch server, a cloud computing platform, etc.

[0049] Server 104 performs equivalent modeling of the urban power grid based on the grid structure coupling characteristics and power balance index values, obtaining an equivalent model. Based on the equivalent model, the urban power grid is divided into multiple control zones. Server 104 determines the importance score of each power grid resource according to its resource characteristics within each control zone, and classifies the power grid resources according to their importance scores, obtaining classified power grid resources. For the classified power grid resources within the control zones, Server 104 constructs a multi-objective optimization function that includes new energy consumption targets and frequency stability targets, and generates a stable control strategy for the classified power grid resources through the multi-objective optimization function.

[0050] In one exemplary embodiment, such as Figure 2 As shown, a method for zoned stability control of massive micro-resources in urban power grids is provided, and this method is applied to... Figure 1 Taking server 104 as an example, the explanation includes:

[0051] Step S202: Based on the grid structure coupling characteristics and power balance index values ​​of the urban power grid, perform equivalent modeling of the urban power grid to obtain the equivalent model, and divide the urban power grid into multiple control zones based on the equivalent model.

[0052] Urban power grids can refer to the power transmission and distribution network within a megacity, including generator sets, substations, transmission lines, distribution equipment, and various loads. Urban power grids may contain multiple voltage levels, such as 220 kV, 110 kV, and 10 kV, and are connected to a large number of distributed photovoltaic systems, energy storage systems, interruptible loads, and other micro-resources.

[0053] The coupling characteristics of the power grid structure refer to the electrical connections between nodes in an urban power grid through transmission lines, transformers, and other equipment. These connections determine the power transmission capacity and voltage impact between nodes. The coupling characteristics of the power grid structure can be quantitatively characterized by parameters such as the impedance relationship between nodes and the power transmission distribution factor.

[0054] The power balance index refers to the degree of balance between power generation and load in the power grid, which can include active power balance and reactive power balance. The power balance index is used to measure whether each region of the power grid has the ability to operate relatively independently.

[0055] Equivalent modeling refers to simplifying a complex urban power grid into an equivalent network model. This equivalent model retains the main electrical characteristics of the power grid while reducing its complexity. Equivalent modeling includes two methods: static equivalent modeling and dynamic equivalent modeling. Static equivalent modeling is used for the power grid, while dynamic equivalent modeling is used for generator aggregation and load aggregation.

[0056] Control zoning refers to dividing the urban power grid into several relatively independent regions based on the electrical coupling characteristics of the power grid. The electrical coupling strength between nodes within each region is high, while the electrical coupling strength between regions is low. Through zoning, coordinated control of massive micro-resources can be achieved.

[0057] For example, the process of equivalent modeling an urban power grid includes: static equivalent processing of the network structure of the urban power grid, and dynamic equivalent processing of the generators and loads in the urban power grid.

[0058] The specific process of static equivalence is as follows:

[0059] S.1.1, based on Ward's equivalent model, the power grid is divided into an internal system (I), boundary nodes (B), and an external system (E). By eliminating the internal nodes, a model is obtained that retains only the boundary nodes and the equivalent injection.

[0060] S.1.2, the nodal voltage equations of the detailed network are written in block matrix form:

[0061] ;

[0062] in, These refer to the injected current at the boundary, internal, and external nodes, respectively. These represent the voltage values ​​at the boundary, internal, and external nodes, respectively; Y is the admittance matrix, with subscripts indicating subarrays between corresponding node sets.

[0063] S.1.3, assuming no injection occurs in internal nodes ( Then the second line of the equation is:

[0064] ;

[0065] From this, the internal node voltages can be solved. :

[0066] ;

[0067] Substitute the above equation into the first row of the equation. ,get:

[0068] ;

[0069] This leads to the equivalent network equation:

[0070] ;

[0071] in, It is the self-integration matrix of the boundary nodes after equivalence, representing the connection relationship between nodes inside the equivalence network. The equivalent admittance of the boundary nodes and the external system. This is the equivalent injected current term.

[0072] The specific process of dynamic equivalence is as follows:

[0073] S.2.1, the specific equivalent process of the generator is as follows:

[0074] First, by comparing the generator power angle difference after the disturbance, we can find the units whose power angle curves remain consistent (coordinated):

[0075] ;

[0076] in It is a small threshold. A generator that satisfies the power angle condition. and They are considered to be in sync and can aggregate.

[0077] The synchronized generator units are aggregated into a single equivalent generator. The equivalent parameters of the generator can be calculated as follows:

[0078] ;

[0079] ;

[0080] ;

[0081] in, Let i be the rated capacity of the i-th unit. For equivalent capacity, The equivalent inertial time constant, Let be the inertial time constant of the i-th unit. Let i be the transient reactance of the i-th unit. It is the equivalent transient reactance.

[0082] S.2.2, the specific equivalent process of the load is as follows:

[0083] For the aggregation of static loads, the ZIP model parameters (the ratio of constant impedance Z, constant current I, and constant power P) are weighted by power:

[0084] ;

[0085] in, The parameters obtained after aggregation are: These represent the constant impedance ratio, constant current ratio, and constant power ratio after polymerization, respectively.

[0086] For the aggregation of dynamic loads, such as the aggregation of induction motors, the parameters of the equivalent motor (such as inertia time constant) are... Stator Reactor Rotor reactance The capacity-weighted average method is used for calculation.

[0087] In practice, after obtaining the equivalent model, the urban power grid can be divided into multiple control zones based on the equivalent model. The specific steps for dividing the power grid into these zones may include:

[0088] S.3.1, Construct a weighted graph model of the power grid ,in This is a set of nodes, representing busbars and substations; Let it be a set of edges, representing transmission lines and transformers; The weight of the edge represents the electrical coupling strength between nodes.

[0089] The weight definition methods include:

[0090] S.3.1.1, Weights based on electrical distance: Electrical distance It reflects the electrical tightness between nodes, and its reciprocal or negative exponent can be used as the coupling strength weight.

[0091] ;

[0092] ;

[0093] in, It is the nodal impedance matrix. It is the attenuation coefficient. The smaller the distance (the closer the electrical distance), the higher the weight. The larger the value, the stronger the coupling.

[0094] S.3.1.2, Weights based on power flow entropy / PTDF:

[0095] ;

[0096] in, It is a line For nodes The power transfer distribution factor of the injected power. The smaller this value, the better the node... and The more similar the impact on power flow in the lines, the closer the electrical connection. The weight can also be defined by taking the reciprocal of the value.

[0097] S.3.1.3, Weights based on synchronization coefficients:

[0098] ;

[0099] That is, calculating the change in voltage phase angle between two nodes using data from a wide-area measurement system (WAMS). The correlation coefficient is the highest correlation coefficient between the two nodes. A higher correlation coefficient indicates a greater consistency in the dynamic process of the two nodes' swings and a stronger coupling.

[0100] S.3.2, Modularity Q is used as a quantitative indicator of partition quality, and its calculation formula is as follows:

[0101] ;

[0102] in, For nodes and The edge weights between them. For nodes The sum of the weights of all connected edges (i.e., node strength). This is the sum of the weights of all edges in the graph. In a randomly connected network, nodes and The expected weights between them. For the Kronecker function, when the node and When they belong to the same partition (i.e.) ), Otherwise .

[0103] S.3.3, the Louvain optimization algorithm is used to perform community detection on the graph model. Through iterative local optimization and network aggregation, the partitioning scheme that maximizes the modularity Q is found.

[0104] The Louvain algorithm iterative process is as follows:

[0105] a. Initialization: Initialize each node as an independent partition.

[0106] b. The first stage is local optimization, and the specific process is as follows:

[0107] b.1 Traverse each node .

[0108] b.2 Try to move the node Move from the original partition to each of its neighboring nodes. The partition it belongs to.

[0109] b.3 Calculate the modularity gain for each move. The modularity gain formula is:

[0110] ;

[0111] in The sum of the weights of all edges within the target partition. This is the sum of the weights of the edges connected to all nodes within the target partition. For nodes The sum of the weights of all connected edges. For nodes The sum of edge weights between the target partition and some nodes within the partition.

[0112] b.4 will node Move to the position that yields the maximum positive gain The neighboring partition. If all gains are non-positive, then the node Remain in the original partition.

[0113] b.5 Repeat this process until no node movement can improve performance. until.

[0114] c. The second stage is mesh aggregation, and the specific steps are as follows:

[0115] c.1 Aggregate the partitions obtained in the first stage into new "supernodes".

[0116] c.2 The edge weights between “supernodes” in the new graph are equal to the sum of the weights of all edges between the corresponding two partitions in the original graph.

[0117] c.3 The edge weights within a partition will form a self-loop, representing the internal connection strength of the original partition.

[0118] d. Iteration: On the new graph after aggregation, repeat the first and second phases until the modularity is reached. It will no longer change.

[0119] S.3.4, introduce functional constraints to verify and correct the partitioning scheme, including:

[0120] a. Active / Reactive Power Balance Constraint:

[0121] ;

[0122] ;

[0123] in, These are partitions The generators, loads, and reactive power compensation equipment inside the unit. It is the permissible power imbalance threshold, which ensures that each zone is electrically relatively autonomous.

[0124] b. Connectivity constraint: Each partition must be connected in the network topology.

[0125] Step S204: Based on the resource characteristics of each power grid resource within the control zone, determine the importance score of each power grid resource, and classify each power grid resource according to the importance score to obtain the classified power grid resources.

[0126] Among them, grid resources refer to various equipment and loads within the control zone that can participate in grid regulation, including energy storage systems, distributed photovoltaics, interruptible loads, traditional loads, and other micro-resources of the urban power grid.

[0127] Among them, resource characteristics refer to the multi-dimensional characteristics that describe the attributes of power grid resources, including resource type characteristics, resource user importance characteristics, voltage level characteristics, response speed characteristics, and interruption resource quantity characteristics.

[0128] The importance score refers to the comprehensive score calculated based on multiple resource characteristics of power grid resources. This score is used to measure the importance of power grid resources in power grid regulation.

[0129] Among them, the classified power grid resources refer to the set of resources obtained after classifying power grid resources according to their importance scores. Different categories of power grid resources have different priorities and control strategies in stability control.

[0130] In practice, a feature vector can be used for each power grid resource i. express:

[0131] ;

[0132] in, Based on resource type characteristics, unique thermal coding is used to represent them, such as energy storage, photovoltaic, interruptible load, and traditional load; Assign values ​​to features related to importance based on the social attributes of resource users; Assign a value based on the voltage level characteristic of the resource, according to the voltage level to which it is connected; For response speed, Both represent interruption costs and response technology characteristics.

[0133] S.4.2, Calculate the importance score using the weighted summation model formula as follows:

[0134] ;

[0135] in, The weight vector W is a scalarized aggregate value of resource type characteristics, and can be dynamically adjusted according to the power grid operation scenario.

[0136] S.4.3, Comprehensive Classification: Based on Importance Score A threshold is set for initial classification (important, secondary important, general), and then the classification is corrected by combining the pre-set hard rules based on the characteristics of the resource type to obtain the final classification result, thus forming the classified power grid resources.

[0137] Step S206: For the classified grid resources within the control zone, construct a multi-objective optimization function that includes the renewable energy consumption target and the frequency stability target; and generate a stability control strategy for the classified grid resources through the multi-objective optimization function.

[0138] The renewable energy consumption target refers to maximizing the utilization rate of renewable energy and minimizing the amount of renewable energy wasted, while ensuring the stable operation of the power grid.

[0139] The frequency stability objective refers to maintaining the power grid frequency within the allowable range and minimizing frequency deviation and frequency change rate through coordinated control of regulatory resources.

[0140] Among them, the multi-objective optimization function refers to the optimization problem that simultaneously considers the goals of new energy consumption and frequency stability. This optimization problem requires finding a control strategy that optimizes both goals as much as possible while satisfying various constraints.

[0141] Among them, the stability control strategy refers to the specific control instructions generated for the classified grid resources, including the charging and discharging power sequence of energy storage systems, the curtailment power sequence of new energy sources, and the power reduction sequence of interruptible loads.

[0142] For example, a stable control policy can be generated by solving a multi-objective optimization problem. The specific process may include:

[0143] S.5.1, Define decision variables, assuming the control period is... The formula is:

[0144] ;

[0145] in, The charging and discharging power sequence for the energy storage system;

[0146] in, This is a sequence of abandoned renewable energy power.

[0147] in, This is a power reduction sequence for interruptible loads.

[0148] S.5.2, Define the objective function, including:

[0149] a. Objective 1: Maximize the consumption of new energy sources, i.e. minimize the total amount of abandoned electricity.

[0150] ;

[0151] b. Objective 2: Maximize frequency stability by minimizing the operational costs of adjusting resources and the rate of power change, thereby providing smooth frequency support.

[0152] ;

[0153] Here, X represents the control of resources.

[0154] S.5.3 sets constraints, including power balance equation constraints, thermal stability control and equipment safety inequality constraints, and control resource self-constraints.

[0155] a. Power balance equation constraint:

[0156] ;

[0157] in, It includes the output of conventional generating units and renewable energy sources that have not been abandoned.

[0158] b. Inequality constraint between thermal stability control and equipment safety

[0159] b.1 Line power flow constraints:

[0160] ;

[0161] in It can be expressed as a linear function of the decision variables by DC power flow or sensitivity coefficient:

[0162] ;

[0163] in Inject power into the node.

[0164] b.2 Transformer capacity constraints:

[0165] ;

[0166] b.3 Node voltage safety constraints:

[0167] ;

[0168] c. Regulating resource self-constraints

[0169] c.1 Constraints of Energy Storage Systems:

[0170] ;

[0171] ;

[0172] c.2 Constraints on Curtailment of New Energy Power:

[0173] ;

[0174] c.3 Interruptible load constraints:

[0175] ;

[0176] S.5.4, the weighted summation method is used to transform the multi-objective problem into a single objective, which is then solved using a linear programming or quadratic programming solver. By adjusting the weight factor λ, the Pareto optimal solution set is obtained, and the two objective functions are merged into a single objective function.

[0177] ;

[0178] Among them, the weighting factor λ is a dynamic weight, the value of which is determined by the real-time disturbance level of the power grid:

[0179] S.6.1, Constructing a comprehensive power grid disturbance index This indicator is composed of frequency deviation rate. and frequency change rate It is a weighted combination after normalization;

[0180] ;

[0181] ;

[0182] ;

[0183] in, It is the real-time frequency. It is the rated frequency. and These are the maximum allowable or typical values ​​for these indicators. It is a weighting coefficient, and .

[0184] S.6.2, the perturbation index D(t) is mapped to the dynamic weight factor λ(t) using a sigmoid function:

[0185] ;

[0186] in, The disturbance threshold is determined by power grid stability calculations. This is the curve steepness coefficient. The larger the value, the steeper the curve, and the faster the weight switching. For scenarios requiring millisecond-level response times, A large value should be chosen to achieve a near "step" switch.

[0187] When the system disturbance is small ( Hour, When the value is close to 1, the optimization target is biased towards the consumption of new energy sources; when the system disturbance is large ( When (large), Approaching 0, the optimization target is biased towards frequency stability, thereby achieving adaptive switching of the control strategy.

[0188] In the aforementioned method for zonal stability control of massive micro-resources in urban power grids, the urban power grid is modeled using equivalent methods based on its grid structure coupling characteristics and power balance index values. This yields an equivalent model, which is then used to divide the urban power grid into multiple control zones. Based on the resource characteristics of each grid resource within a control zone, an importance score is determined for each resource. These resources are then classified according to their importance scores, resulting in categorized grid resources. For the categorized grid resources within each control zone, a multi-objective optimization function is constructed, incorporating both renewable energy absorption and frequency stability objectives. This multi-objective optimization function is then used to generate a stability control strategy for the categorized grid resources. Thus, for ultra-large urban power grids integrating distributed photovoltaic, energy storage, and interruptible load micro-resources, zonal equivalence, resource classification, and multi-level coordinated control can accurately and collaboratively control the massive micro-resources of the urban power grid, improving grid stability and resource utilization efficiency.

[0189] In another embodiment, the urban power grid is modeled using equivalent methods based on its grid structure coupling characteristics and power balance index values ​​to obtain an equivalent model, including:

[0190] For urban power grids, the grid is divided into internal systems, boundary nodes, and external systems. A block matrix of node voltage equations is constructed. By eliminating internal nodes in the block matrix, an equivalent network equation is constructed, including the self-admittance of boundary nodes, the mutual admittance between boundary nodes and the external system, and the equivalent injected current. This serves as the equivalent post-model for the urban power grid. For generators in the urban power grid, based on the generator power angle difference after disturbance, co-tuning units with consistent power angle curves are identified. These co-tuning units are aggregated into equivalent generators. Based on the equivalent capacity, equivalent inertial time constant, and equivalent transient reactance of the equivalent generators, an equivalent post-model for the generators is constructed. For static loads in the urban power grid, power-weighted averaging is used to calculate the proportional parameters of constant impedance, constant current, and constant power, and an equivalent post-model for static loads is constructed. For dynamic loads in the urban power grid, capacity-weighted averaging is used to calculate the equivalent motor parameters, and an equivalent post-model for dynamic loads is constructed.

[0191] In practical implementation, for the network structure of urban power grids, the Ward equivalence method can be used to divide the urban power grid into three parts: internal system, boundary nodes, and external system. The internal system refers to the area that needs simplification, the boundary nodes are the connection nodes between the internal and external systems, and the external system is the area where detailed information needs to be retained.

[0192] Then, by eliminating internal nodes, an equivalent network model is constructed that retains only boundary nodes and equivalent injections. Specifically, the nodal voltage equations of the urban power grid can be written in block matrix form, which includes the current injection and voltage quantities of boundary nodes, internal nodes, and external nodes, as well as the admittance relationships between nodes.

[0193] For example, the block matrix can be represented as:

[0194] ;

[0195] Then, assuming no injected current at the internal nodes, the relationship between the internal node voltage, the boundary node voltage, and the external node voltage is solved from the second row of equations in the block matrix. Substituting this relationship into the first row of equations in the block matrix, the equivalent self-admittance of the boundary nodes, the mutual admittance between the boundary nodes and the external system, and the equivalent injected current are obtained, thus constructing the equivalent network equation.

[0196] The equation of the iso-network can be expressed as:

[0197] ;

[0198] The injected current at the boundary node is equal to the product of the equivalent boundary node self-admittance and the boundary node voltage, plus the product of the equivalent mutual admittance between the boundary node and the external system and the external node voltage, plus the equivalent injected current term. Here, the equivalent boundary node self-admittance represents the connection relationship between nodes within the equivalent network, and the equivalent boundary node mutual admittance represents the electrical coupling relationship between the boundary node and the external system.

[0199] For generators in urban power grids, dynamic equivalence can be achieved through the aggregation of co-regulating units. By comparing the power angle change curves of each generator after being disturbed, co-regulating units with consistent power angle curves are identified. Specifically, when the norm of the power angle difference between two generators is less than a preset threshold, these two generators are determined to be co-regulating units.

[0200] Then, the identified co-tuning units are aggregated into an equivalent generator. Specifically, the equivalent capacity of the equivalent generator is calculated based on the rated capacity of each co-tuning unit, and the equivalent capacity is equal to the sum of the rated capacities of all co-tuning units; the equivalent inertial time constant of the equivalent generator is calculated based on the inertial time constant and rated capacity of each co-tuning unit, and the equivalent inertial time constant is equal to the sum of the products of the inertial time constant and rated capacity of each co-tuning unit divided by the equivalent capacity; the equivalent transient reactance of the equivalent generator is calculated based on the transient reactance and rated capacity of each co-tuning unit, and the equivalent transient reactance is equal to the sum of the products of the transient reactance and rated capacity of each co-tuning unit divided by the equivalent capacity.

[0201] For static loads in urban power grids, the ZIP model parameters of the equivalent load are calculated using a power-weighted averaging method. The ZIP model parameters include the scaling factors for constant impedance, constant current, and constant power components, with the sum of these three scaling factors equal to 1. Specifically, the equivalent ZIP model parameters of the equivalent load are calculated based on the power of each static load and the ZIP model parameters.

[0202] For dynamic loads in urban power grids, equivalent motor parameters are calculated using a capacity-weighted average method. Equivalent motor parameters include inertia time constant, stator reactance, and rotor reactance. Specifically, the equivalent motor parameters are calculated based on the capacity and motor parameters of each dynamic load.

[0203] The technical solution of this embodiment obtains a simplified equivalent model through the equivalent modeling process. This equivalent model retains the main electrical characteristics of the urban power grid while reducing the complexity of the model, laying the foundation for subsequent zoning processing.

[0204] In another embodiment, the urban power grid is divided into multiple control zones based on the equivalence model, including: constructing a power grid weighted graph model; the node set of the power grid weighted graph model is used to represent buses or substations, the edge set of the power grid weighted graph model is used to represent transmission lines or transformers, and the weight of the edge of the power grid weighted graph model is used to represent the electrical coupling strength between nodes; using modularity as a quantitative index value of the partition quality of the control zone, the Louvain algorithm is used to perform community detection on the power grid weighted graph model, and local optimization and network aggregation are performed iteratively until the modularity reaches the maximum value, resulting in multiple control zones.

[0205] First, a weighted graph model of the power grid is constructed, which includes a set of nodes, a set of edges, and edge weights. The set of nodes represents the busbars or substations in the urban power grid, the set of edges represents the transmission lines or transformers, and the edge weights represent the electrical coupling strength between nodes.

[0206] As an example, edge weights can be defined based on electrical distance. Electrical distance reflects the degree of electrical tightness between nodes; a smaller electrical distance indicates stronger electrical coupling between the two nodes. Specifically, the electrical distance between node i and node j is calculated based on the node impedance matrix. This electrical distance equals the self-impedance of node i plus the self-impedance of node j minus twice the mutual impedance between nodes i and j. The reciprocal or negative exponent of the electrical distance is used as the edge weight, resulting in a larger weight between node pairs with smaller electrical distances.

[0207] As another example, edge weights can be defined based on the power transmission distribution factor (DPF). The DPF represents the degree of influence of node-injected power on line power flow. The sum of the absolute values ​​of the differences in DPF between node i and node j across all lines is calculated; the smaller this sum, the more similar the influence of the two nodes on line power flow, and the closer their electrical connection. The reciprocal of this sum is used as the edge weight.

[0208] As another example, edge weights are defined using a synchronization coefficient-based approach. Voltage phase angle change data for each node are acquired through a wide-area measurement system, and the correlation coefficient between the voltage phase angle changes of node i and node j is calculated. A higher correlation coefficient indicates greater consistency in the swing of the two nodes during the dynamic process, and stronger electrical coupling. This correlation coefficient is then used as the edge weight.

[0209] In practical implementation, after constructing the power grid weighted graph model, the Louvain algorithm can be used to perform community detection on the power grid weighted graph model to achieve the partitioning of the urban power grid. The Louvain algorithm is a community detection algorithm based on modularity optimization. It finds the partitioning scheme that maximizes modularity by iteratively executing two stages: local optimization and network aggregation.

[0210] Modularity is a quantitative indicator of partition quality; a higher modularity value indicates better partition quality. The calculation of modularity considers the difference between the connection weights between nodes within the same partition in a real network and the expected connection weights in a random network. A larger modularity value is achieved when the actual connection weights between nodes within the same partition are significantly higher than the random expected values.

[0211] In its implementation, local optimization is performed in the first stage of the Louvain algorithm. First, each node is initialized as an independent partition. Then, each node is traversed, attempting to move it from its original partition to the partitions of each of its neighboring nodes, and the modularity gain for each move is calculated. The modularity gain reflects the degree to which node movement improves the partition quality.

[0212] When calculating the modularity gain, factors such as the sum of the weights of all edges within the target partition, the sum of the weights of edges connected to all nodes within the target partition, the sum of the weights of all edges connected to the moved node, and the sum of the edge weights between the moved node and some nodes within the target partition are considered. The node is moved to the neighboring partition that yields the maximum positive gain. If all gains are non-positive, the node remains in its original partition.

[0213] Repeat the node movement process described above until no node movement can improve modularity. At this point, the first stage of local optimization is complete.

[0214] Then, network aggregation is performed in the second stage of the Louvain algorithm. The partitions obtained in the first stage are aggregated into new supernodes, constructing a new network graph. In the new network graph, the edge weights between supernodes are equal to the sum of the weights of all edges between corresponding partitions in the original graph, and the edge weights within a partition form self-loops of the supernodes, representing the internal connection strength of the original partition.

[0215] The first and second phases are repeated on the new network graph after aggregation until the modularity no longer changes. At this point, a partitioning scheme that maximizes the modularity is obtained, and multiple control partitions are obtained based on this partitioning scheme.

[0216] After obtaining the zoning scheme, functional constraints are introduced to verify and modify the scheme to ensure that it meets the actual needs of power grid operation. Functional constraints include active power balance constraints, reactive power balance constraints, and connectivity constraints.

[0217] Among them, the active power balance constraint requires that the difference between the total generator output and the total load demand in each zone shall not exceed the preset active power imbalance threshold as a proportion of the total load demand.

[0218] Among these constraints, the reactive power balance requirement stipulates that the ratio of the difference between the total reactive power output of generators and reactive power compensation equipment in each zone and the total reactive power demand of the load does not exceed a preset reactive power imbalance threshold. These two constraints ensure that each zone has relative electrical autonomy.

[0219] The connectivity constraint requires that each partition be connected in the network topology, meaning that there must be a path connecting any two nodes within a partition through other nodes within the partition. The partitioning scheme is checked to ensure it meets the connectivity constraint. If a partition is not connected, it is split or merged until all partitions meet the connectivity constraint.

[0220] Through the above zoning process, the urban power grid is divided into multiple control zones. The electrical coupling strength between nodes within each control zone is high, while the electrical coupling strength between control zones is low. This achieves the zoning objective of strong coupling within zones and weak coupling between zones, laying the foundation for subsequent zoned stability control.

[0221] In another embodiment, the importance score of each power grid resource is determined based on the resource characteristics of each power grid resource within the control zone. This includes: constructing multiple resource characteristics of the power grid resource based on the resource type characteristics, resource user importance characteristics, voltage level characteristics, response speed characteristics, and interruption resource quantity characteristics; performing a weighted summation of each resource characteristic to determine the importance score of the power grid resource; and adjusting the weight of each resource characteristic according to the operation scenario of the urban power grid.

[0222] In the specific implementation, the resource feature vector of each power grid resource is first constructed. This resource feature vector includes resource type features, resource user importance features, voltage level features, response speed features, and interruption resource quantity features.

[0223] Among them, the resource type feature is used to characterize the type of power grid resources, and the resource type is encoded using one-hot encoding. For example, for energy storage systems, the resource type feature is set as a vector [1,0,0,0]; for distributed photovoltaic systems, the resource type feature is set as a vector [0,1,0,0]; for interruptible loads, the resource type feature is set as a vector [0,0,1,0]; and for traditional loads, the resource type feature is set as a vector [0,0,0,1].

[0224] Among them, the resource user importance feature is used to characterize the importance of the users served by the power grid resources, and is assigned a value based on the social attributes of the users. For example, important users such as hospitals and schools are assigned a higher importance feature value, general industrial users are assigned a medium importance feature value, and commercial users who can be flexibly adjusted are assigned a lower importance feature value.

[0225] Among them, the voltage level characteristic is used to characterize the voltage level to which the power grid resources are connected, and a value is assigned according to the voltage level. For example, resources connected to the 220 kV voltage level are assigned a higher voltage level characteristic value; resources connected to the 110 kV voltage level are assigned a medium voltage level characteristic value; and resources connected to the 10 kV or lower voltage level are assigned a lower voltage level characteristic value.

[0226] The response speed characteristic is used to characterize the response speed of power grid resources to control commands, and is assigned a value based on the technical characteristics of the resources. For example, a higher response speed characteristic value is assigned to energy storage systems with response times in the millisecond range; a medium response speed characteristic value is assigned to interruptible loads with response times in the second range; and a lower response speed characteristic value is assigned to traditional loads with response times in the minute range.

[0227] The interruption resource quantity characteristic is used to characterize the impact of interrupting the power grid resource on users, and is assigned a value based on the resource's capacity and the user's sensitivity. For example, resources with large capacity and insensitive to interruptions are assigned a lower interruption resource quantity characteristic value; resources with small capacity but sensitive to interruptions are assigned a higher interruption resource quantity characteristic value.

[0228] After constructing the resource feature vector, the importance score of each resource feature is determined by weighted summation.

[0229] Specifically, weighting coefficients are set for each resource characteristic, and these coefficients can be dynamically adjusted according to the operating scenarios of the urban power grid. For example, when the urban power grid is in normal operation, higher weighting coefficients are assigned to resource type characteristics and response speed characteristics to prioritize the use of fast-response energy storage systems and distributed photovoltaics for regulation. When the urban power grid is in an emergency, higher weighting coefficients are assigned to resource user importance characteristics and voltage level characteristics to prioritize power supply to important users and high-voltage level resources.

[0230] Optionally, a scalarized aggregate value for resource type characteristics can be calculated, which converts the uniquely thermally encoded resource type characteristics into a single numerical value. For example, different scalar values ​​can be preset for different resource types: 1.0 for energy storage systems, 0.8 for distributed photovoltaic systems, 0.6 for interruptible loads, and 0.4 for traditional loads.

[0231] In practice, the scalar aggregated value of the resource type feature is multiplied by the corresponding weight coefficient, plus the resource user importance feature multiplied by the corresponding weight coefficient, plus the voltage level feature multiplied by the corresponding weight coefficient, plus the response speed feature multiplied by the corresponding weight coefficient, plus the interruption resource quantity feature multiplied by the corresponding weight coefficient, to obtain the importance score of the power grid resource.

[0232] For example, the importance score can be expressed as:

[0233] .

[0234] After obtaining the importance scores of each power grid resource, the resources are classified according to their importance scores. Multiple importance score thresholds are set to divide the power grid resources into three categories: important resources, less important resources, and general resources.

[0235] For example, the first importance score threshold is set to 0.7, and the second importance score threshold is set to 0.4. When the importance score of a power grid resource is greater than 0.7, the power grid resource is classified as an important resource. When the importance score of a power grid resource is greater than 0.4 and less than or equal to 0.7, the power grid resource is classified as a secondary important resource. When the importance score of a power grid resource is less than or equal to 0.4, server 104 classifies the power grid resource as a general resource.

[0236] The technical solution of this application embodiment, through the above-described resource feature extraction and classification process, achieves effective classification of massive micro-resources within the control partition, providing a foundation for the subsequent generation of stable control strategies.

[0237] In another embodiment, a multi-objective optimization function is constructed, which includes a renewable energy consumption target and a frequency stability target. This includes: determining decision variables; constructing a first objective function to maximize renewable energy consumption; constructing a second objective function to maximize frequency stability; and constructing a multi-objective optimization function based on the first and second objective functions to optimize the decision variables under certain constraints. The constraints include at least power balance constraints, thermal stability control and equipment safety constraints, and regulation resource constraints.

[0238] The decision variables include the charging and discharging power sequence of the energy storage system, the curtailment power sequence of new energy sources, and the power reduction sequence of interruptible loads.

[0239] Among them, the charging and discharging power sequence of the energy storage system represents the charging or discharging power of the energy storage system at each moment during the control period; the curtailed power sequence of new energy sources represents the curtailed power of new energy sources at each moment during the control period; and the power reduction sequence of interruptible loads represents the power reduction of interruptible loads at each moment during the control period.

[0240] In the specific implementation, a first objective function is constructed to maximize the absorption of renewable energy. Specifically, the total amount of power wasted by renewable energy at each moment within the control period is accumulated to obtain the total amount of power wasted; minimizing the total amount of power wasted is used as the first objective function, that is, the first objective function is equal to the sum of the total amount of power wasted by renewable energy at each moment within the control period.

[0241] For example, the first objective function can be expressed as:

[0242] .

[0243] In the specific implementation, a second objective function is constructed to maximize frequency stability. Specifically, smooth frequency support is provided by minimizing the operational cost of adjusting resources and the power change rate; the operational cost of the energy storage system is obtained by calculating the square of the difference in charging and discharging power between adjacent time points, multiplying it by the weighting coefficient of the energy storage system, and summing the results; the operational cost of the interruptible load is obtained by calculating the square of the difference in power reduction between adjacent time points of the interruptible load, multiplying it by the weighting coefficient of the interruptible load, and summing the results; the operational cost of the energy storage system and the operational cost of the interruptible load are added together as the second objective function.

[0244] For example, the second objective function can be expressed as:

[0245] .

[0246] Then, various constraints are set to ensure that the generated control strategy meets the actual needs of power grid operation. These constraints include power balance constraints, thermal stability control and equipment safety constraints, and regulation resource constraints.

[0247] The power balance constraint is an equality constraint, requiring that at any given moment during the control period, the total generator output minus the power curtailed from renewable energy sources plus the charging and discharging power of the energy storage system equals the total load demand minus the power reduction from interruptible loads plus network losses. The total generator output includes both conventional units and the output of renewable energy sources that have not been curtailed.

[0248] Among them, thermal stability control and equipment safety constraints are inequality constraints, including line power flow constraints, transformer capacity constraints, and node voltage safety constraints. Line power flow constraints require that the absolute value of the power flow of each transmission line at any given time does not exceed the maximum allowable power flow of that line. Specifically, this can be expressed as a linear function of the node injected power, and thus as a linear function of the decision variables, through a DC power flow model or sensitivity coefficients. Transformer capacity constraints require that the absolute value of the transmitted power of each transformer at any given time does not exceed the maximum capacity of that transformer. Node voltage safety constraints require that the voltage value of each node at any given time is between the minimum and maximum allowable voltage.

[0249] Among these, resource constraints include energy storage system constraints, renewable energy curtailment constraints, and interruptible load constraints. Energy storage system constraints require that the state of charge (SOC) of the energy storage system at each time point be between the allowable minimum and maximum SOC, and that the charging and discharging power of the energy storage system be between the allowable maximum charging power and maximum discharging power. Specifically, recursive constraints on the SOC can be established based on the relationship between the changes in SOC, charging and discharging power, charging and discharging efficiency, and rated capacity of the energy storage system at adjacent times. Renewable energy curtailment constraints require that the curtailed power of renewable energy at each time point be between 0 and the predicted output of renewable energy at that time point. Interruptible load constraints require that the power reduction of interruptible loads at each time point be between 0 and the maximum reducible power of the interruptible load.

[0250] In another embodiment, a stability control strategy for classified grid resources is generated through a multi-objective optimization function, including: obtaining a comprehensive grid disturbance index value, which is determined based on the frequency deviation rate and the frequency change rate; mapping the comprehensive grid disturbance index value to a dynamic weighting factor using a Sigmoid function; weighting and merging the first objective function and the second objective function into a single objective function using the dynamic weighting factor; and generating a stability control strategy for classified grid resources by solving the single objective function. Specifically, the smaller the comprehensive grid disturbance index value, the more the dynamic weighting factor biases the optimization objective of the single objective function towards renewable energy consumption; the larger the comprehensive grid disturbance index value, the more the dynamic weighting factor biases the optimization objective of the single objective function towards frequency stability.

[0251] In practice, after constructing the first objective function, the second objective function, and various constraints, a multi-objective optimization function is constructed based on the first and second objective functions to optimize the decision variables under the constraints.

[0252] For example, a stability control strategy for classified power grid resources is generated through a multi-objective optimization function.

[0253] First, the comprehensive power grid disturbance index value can be obtained, which is determined based on the frequency deviation rate and the frequency change rate.

[0254] The frequency deviation rate is the absolute value of the difference between the real-time frequency and the rated frequency, divided by the rated frequency; the frequency change rate is the derivative of the real-time frequency with respect to time. Specifically, the normalized frequency deviation rate is obtained by dividing the frequency deviation rate by the maximum allowable value; the normalized frequency change rate is obtained by dividing the frequency change rate by the maximum allowable value. Then, the normalized frequency deviation rate is multiplied by a first weighting coefficient, and the normalized frequency change rate is multiplied by a second weighting coefficient to obtain the comprehensive power grid disturbance index value, where the sum of the first and second weighting coefficients equals 1. For example, if the first weighting coefficient is 0.6 and the second weighting coefficient is 0.4, it means that the influence weight of the frequency deviation rate on the comprehensive power grid disturbance index value is greater than the influence weight of the frequency change rate.

[0255] In the specific implementation, the Sigmoid function is used to map the comprehensive power grid disturbance index value to a dynamic weighting factor. The Sigmoid function is an S-shaped curve function, and its output value is between 0 and 1. A disturbance threshold and a curve steepness coefficient are set, and the dynamic weighting factor is calculated using the Sigmoid function.

[0256] The disturbance threshold, determined by grid stability calculations, represents the critical point at which the grid transitions from a normal state to a disturbed state. The curve steepness coefficient determines the steepness of the Sigmoid function curve; the larger the curve steepness coefficient, the more rapidly the dynamic weighting factor changes near the disturbance threshold.

[0257] For example, when the comprehensive power grid disturbance index is much smaller than the disturbance threshold, the output of the Sigmoid function is close to 1, and the dynamic weighting factor is close to 1; when the comprehensive power grid disturbance index is much larger than the disturbance threshold, the output of the Sigmoid function is close to 0, and the dynamic weighting factor is close to 0; when the comprehensive power grid disturbance index is close to the disturbance threshold, the output of the Sigmoid function smoothly transitions between 0 and 1.

[0258] In practice, a dynamic weighting factor is used to weight and merge the first and second objective functions into a single objective function. The single objective function is obtained by multiplying the first objective function by the dynamic weighting factor and adding the second objective function multiplied by 1 minus the dynamic weighting factor.

[0259] When the comprehensive disturbance index of the power grid is smaller, the dynamic weighting factor is close to 1, and the optimization objective of the single objective function is biased towards minimizing the amount of renewable energy abandoned, that is, biased towards the renewable energy consumption objective; when the comprehensive disturbance index of the power grid is larger, the dynamic weighting factor is close to 0, and the optimization objective of the single objective function is biased towards minimizing the cost of adjusting resources, that is, biased towards the frequency stability objective.

[0260] In practice, a stable control strategy for the classified power grid resources is generated by solving a single objective function. Optionally, a linear programming solver or a quadratic programming solver can be used to solve the single objective function. Under the premise of satisfying power balance constraints, thermal stability control and equipment safety constraints, and regulation resource constraints, the decision variable value that minimizes the single objective function is found, and this decision variable value is the stable control strategy.

[0261] Optionally, the single objective function and its constraints can be converted into a standard linear programming problem or a quadratic programming problem, and then the optimization solver can be called to solve it.

[0262] In practical applications, after obtaining the optimal solution from the optimization solver, the charging and discharging power sequences of the energy storage system, the curtailed power sequences of new energy sources, and the power reduction sequences of interruptible loads are extracted as stability control strategies. These strategies are then distributed to the corresponding grid resources within the control zone, achieving zoned stability control of the massive micro-resources of the urban power grid.

[0263] In another embodiment, the dynamic weighting factor can be defined in a different way. For example, based on the actual operating state of the power grid, multiple disturbance levels can be set, each disturbance level corresponding to a fixed weighting factor. When the comprehensive disturbance index value of the power grid is in different ranges, the weighting factor corresponding to the disturbance level is selected. For example, when the comprehensive disturbance index value of the power grid is less than 0.2, a weighting factor of 0.9 is selected; when the comprehensive disturbance index value of the power grid is between 0.2 and 0.5, a weighting factor of 0.5 is selected; and when the comprehensive disturbance index value of the power grid is greater than 0.5, a weighting factor of 0.1 is selected.

[0264] In another embodiment, different multi-objective optimization methods can be used to generate stable control strategies. For example, instead of weighting and merging the first and second objective functions into a single objective function, the multi-objective optimization problem can be solved directly to obtain a Pareto optimal solution set. Each solution in the Pareto optimal solution set satisfies the constraints, and no other solution can increase the value of one objective function without decreasing the value of another.

[0265] Optionally, a multi-objective optimization algorithm, such as an evolutionary algorithm, can be used to solve the multi-objective optimization problem. A solution that meets the current power grid operation requirements is selected from the Pareto optimal solution set as the stability control strategy. Specifically, solutions biased towards renewable energy consumption or frequency stability can be selected from the Pareto optimal solution set based on the preferences of power grid operators or the real-time state of the power grid.

[0266] In another embodiment, such as Figure 3 As shown, a method for zoned stability control of massive micro-resources in urban power grids is provided, and this method is applied to... Figure 1 Taking server 104 as an example, the following steps are included:

[0267] Step S302: For the urban power grid, the urban power grid is divided into internal system, boundary nodes and external system. A block matrix of node voltage equations is constructed. By eliminating the internal nodes in the block matrix, an equivalent network equation containing the self-admittance of the boundary nodes, the mutual admittance between the boundary nodes and the external system and the equivalent injected current is constructed as the equivalent post-model of the urban power grid.

[0268] Step S304: For generators in the urban power grid, based on the generator power angle difference after disturbance, identify the co-tuning units with consistent power angle curves, aggregate the co-tuning units into equivalent generators, and construct the equivalent post-model corresponding to the generator based on the equivalent capacity, equivalent inertial time constant and equivalent transient reactance of the equivalent generator.

[0269] Step S306: For static loads in the urban power grid, the proportional parameters of constant impedance, constant current and constant power are calculated using power weighted average to construct the equivalent model corresponding to the static load; for dynamic loads in the urban power grid, the equivalent motor parameters are calculated using capacity weighted average to construct the equivalent model corresponding to the dynamic load.

[0270] Step S308: Construct a power grid weighted graph model; use modularity as a quantitative indicator of the partition quality of the control partition, and use the Louvain algorithm to perform community detection on the power grid weighted graph model. Iteratively perform local optimization and network aggregation until the modularity reaches the maximum value to obtain multiple control partitions.

[0271] In this power grid weighted graph model, the set of nodes is used to represent busbars or substations, the set of edges is used to represent transmission lines or transformers, and the weights of the edges are used to represent the electrical coupling strength between nodes.

[0272] Step S310: Based on the resource type characteristics, resource user importance characteristics, voltage level characteristics, response speed characteristics, and interruption resource quantity characteristics of power grid resources, construct multiple resource characteristics of power grid resources.

[0273] Step S312: Perform a weighted summation of each resource characteristic to determine the importance score of the power grid resource, and classify each power grid resource according to the importance score to obtain the classified power grid resources.

[0274] The weights of each resource feature are adjusted according to the operating scenario of the urban power grid.

[0275] Step S314: For the classified power grid resources within the control zone, determine the decision variables; construct a first objective function; construct a second objective function; based on the first and second objective functions, construct a multi-objective optimization function to optimize the decision variables under the constraints.

[0276] The decision variables include the charging and discharging power sequence of the energy storage system, the curtailment power sequence of new energy sources, and the power reduction sequence of interruptible loads; the first objective function is used to maximize the absorption of new energy sources; the second objective function is used to maximize frequency stability; the constraints include at least power balance constraints, thermal stability control and equipment safety constraints, and regulation resource constraints.

[0277] Step S316: Obtain the comprehensive disturbance index value of the power grid, and use the Sigmoid function to map the comprehensive disturbance index value of the power grid to a dynamic weighting factor; use the dynamic weighting factor to weight and merge the first objective function and the second objective function into a single objective function; by solving the single objective function, generate a stability control strategy for the classified power grid resources.

[0278] Among them, the comprehensive disturbance index value of the power grid is determined based on the frequency deviation rate and the frequency change rate.

[0279] The smaller the comprehensive power grid disturbance index value, the more the dynamic weighting factor makes the optimization objective of the single objective function more biased towards new energy consumption; the larger the comprehensive power grid disturbance index value, the more the dynamic weighting factor makes the optimization objective of the single objective function more biased towards frequency stability.

[0280] It should be noted that the specific limitations of the above steps can be found in the specific limitations of a zoned stability control method for massive micro-resources in an urban power grid described above.

[0281] It should be understood that although the steps in the flowcharts of the embodiments described above are shown sequentially according to the arrows, these steps are not necessarily executed in the order indicated by the arrows. Unless explicitly stated herein, there is no strict order restriction on the execution of these steps, and they can be executed in other orders. Moreover, at least some steps in the flowcharts of the embodiments described above may include multiple steps or multiple stages. These steps or stages are not necessarily completed at the same time, but can be executed at different times. The execution order of these steps or stages is not necessarily sequential, but can be performed alternately or in turn with other steps or at least some of the steps or stages of other steps.

[0282] Based on the same inventive concept, this application also provides a device for implementing the above-described method for regional stability control of massive micro-resources in urban power grids. The solution provided by this device is similar to the implementation described in the above method. Therefore, the specific limitations of one or more embodiments of the device for regional stability control of massive micro-resources in urban power grids provided below can be found in the limitations of the method for regional stability control of massive micro-resources in urban power grids described above, and will not be repeated here.

[0283] In one exemplary embodiment, such as Figure 4 As shown, a zoned stability control device for massive micro-resources in an urban power grid is provided, comprising:

[0284] The modeling and partitioning module 410 is used to perform equivalent modeling of the urban power grid based on the grid structure coupling characteristics and power balance index values ​​of the urban power grid, to obtain an equivalent model, and to divide the urban power grid into multiple control partitions based on the equivalent model.

[0285] The feature classification module 420 is used to determine the importance score of each power grid resource according to the resource characteristics of each power grid resource in the control partition, and to classify each power grid resource according to the importance score to obtain the classified power grid resources;

[0286] The stability control module 430 is used to construct a multi-objective optimization function that includes a new energy consumption target and a frequency stability target for the classified power grid resources within the control zone, and to generate a stability control strategy for the classified power grid resources through the multi-objective optimization function.

[0287] In one embodiment, the modeling partitioning module 410 is specifically used to divide the urban power grid into an internal system, boundary nodes, and an external system, construct a block matrix of node voltage equations, and construct an equivalent network equation containing the self-admittance of the boundary nodes, the mutual admittance between the boundary nodes and the external system, and the equivalent injected current by eliminating the internal nodes in the block matrix, as the equivalent post-model corresponding to the urban power grid; for the generators in the urban power grid, based on the generator power angle difference after disturbance, identify the co-tuning units with consistent power angle curves, aggregate the co-tuning units into equivalent generators, and construct the equivalent post-model corresponding to the generators based on the equivalent capacity, equivalent inertial time constant, and equivalent transient reactance of the equivalent generators; for the static loads in the urban power grid, use power weighted averaging to calculate the proportional parameters of constant impedance, constant current, and constant power, and construct the equivalent post-model corresponding to the static loads; for the dynamic loads in the urban power grid, use capacity weighted averaging to calculate the equivalent motor parameters, and construct the equivalent post-model corresponding to the dynamic loads.

[0288] In one embodiment, the modeling partitioning module 410 is specifically used to construct a power grid weighted graph model; the node set of the power grid weighted graph model is used to represent a bus or substation, the edge set of the power grid weighted graph model is used to represent a transmission line or transformer, and the weight of the edge of the power grid weighted graph model is used to represent the electrical coupling strength between nodes; the modularity is used as a quantitative index value of the partitioning quality of the control partition, and the Louvain algorithm is used to perform community detection on the power grid weighted graph model. Through iterative execution of local optimization and network aggregation, the modularity reaches the maximum value, and the multiple control partitions are obtained.

[0289] In one embodiment, the feature classification module 420 is specifically used to construct multiple resource features of the power grid resources based on the resource type features, resource user importance features, voltage level features, response speed features, and interruption resource quantity features of the power grid resources; to perform a weighted summation of each resource feature to determine the importance score corresponding to the power grid resource; and to adjust the weight of each resource feature according to the operation scenario of the urban power grid.

[0290] In one embodiment, the stability control module 430 is specifically used to determine decision variables; the decision variables include the charging and discharging power sequence of the energy storage system, the curtailment power sequence of new energy sources, and the power reduction sequence of interruptible loads; construct a first objective function, which is used to maximize the absorption of new energy sources; construct a second objective function, which is used to maximize frequency stability; and construct a multi-objective optimization function to optimize the decision variables under the constraints of the first objective function and the second objective function; the constraints include at least power balance constraints, thermal stability control and equipment safety constraints, and regulation resource constraints.

[0291] In one embodiment, the stability control module 430 is specifically used to acquire the comprehensive power grid disturbance index value, which is determined based on the frequency deviation rate and the frequency change rate; map the comprehensive power grid disturbance index value to a dynamic weighting factor using a Sigmoid function; use the dynamic weighting factor to weight and merge the first objective function and the second objective function into a single objective function; and generate a stability control strategy for the classified power grid resources by solving the single objective function; wherein, the smaller the comprehensive power grid disturbance index value, the more the dynamic weighting factor makes the optimization objective of the single objective function more biased towards new energy consumption; the larger the comprehensive power grid disturbance index value, the more the dynamic weighting factor makes the optimization objective of the single objective function more biased towards frequency stability.

[0292] The modules in the aforementioned urban power grid's massive micro-resource zone stability control device can be implemented entirely or partially through software, hardware, or a combination thereof. These modules can be embedded in or independent of the processor in a computer device, or stored in the computer device's memory as software, so that the processor can call and execute the corresponding operations of each module.

[0293] In one exemplary embodiment, a computer device is provided, which may be a server, and its internal structure diagram may be as follows: Figure 5As shown, this computer device includes a processor, memory, input / output (I / O) interfaces, and a communication interface. The processor, memory, and I / O interfaces are connected via a system bus, and the communication interface is also connected to the system bus via the I / O interfaces. The processor provides computational and control capabilities. The memory includes non-volatile storage media and internal memory. The non-volatile storage media stores the operating system, computer programs, and a database. The internal memory provides the environment for the operating system and computer programs stored in the non-volatile storage media. The database stores power grid data. The I / O interfaces are used for exchanging information between the processor and external devices. The communication interface is used for communication with external terminals via a network connection. When the computer program is executed by the processor, it implements a method for zoned stability control of massive micro-resources in an urban power grid.

[0294] Those skilled in the art will understand that Figure 5 The structure shown is merely a block diagram of a portion of the structure related to the present application and does not constitute a limitation on the computer device to which the present application is applied. Specific computer devices may include more or fewer components than those shown in the figure, or combine certain components, or have different component arrangements.

[0295] In one exemplary embodiment, a computer device is provided, including a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to implement the steps in the above-described method embodiments.

[0296] In one embodiment, a computer-readable storage medium is provided having a computer program stored thereon, which, when executed by a processor, implements the steps in the above method embodiments.

[0297] In one embodiment, a computer program product is provided, including a computer program that, when executed by a processor, implements the steps in the above method embodiments.

[0298] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for analysis, data stored, data displayed, etc.) involved in this application are all information and data authorized by the user or fully authorized by all parties, and the collection, use and processing of the relevant data must comply with relevant regulations.

[0299] Those skilled in the art will understand that all or part of the processes in the methods of the above embodiments can be implemented by a computer program instructing related hardware. The computer program can be stored in a non-volatile computer-readable storage medium, and when executed, it can include the processes of the embodiments of the above methods. Any references to memory, databases, or other media used in the embodiments provided in this application can include at least one of non-volatile memory and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical memory, high-density embedded non-volatile memory, resistive random access memory (ReRAM), magnetic random access memory (MRAM), ferroelectric random access memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memory can include random access memory (RAM) or external cache memory, etc. By way of illustration and not limitation, RAM can take many forms, such as Static Random Access Memory (SRAM) or Dynamic Random Access Memory (DRAM). The databases involved in the embodiments provided in this application may include at least one type of relational database and non-relational database. Non-relational databases may include, but are not limited to, blockchain-based distributed databases. The processors involved in the embodiments provided in this application may be general-purpose processors, central processing units, graphics processing units, digital signal processors, programmable logic devices, quantum computing-based data processing logic devices, artificial intelligence (AI) processors, etc., and are not limited to these.

[0300] The technical features of the above embodiments can be combined in any way. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this application.

[0301] The embodiments described above are merely illustrative of several implementation methods of this application, and while the descriptions are specific and detailed, they should not be construed as limiting the scope of this patent application. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of this application, and these all fall within the protection scope of this application. Therefore, the protection scope of this application should be determined by the appended claims.

Claims

1. A method for zoned stability control of massive micro-resources in urban power grids, characterized in that, The method includes: The urban power grid is modeled using equivalent values ​​based on its grid structure coupling characteristics and power balance index, resulting in an equivalent model. Based on the equivalent model, the urban power grid is divided into multiple control zones. Based on the resource characteristics corresponding to each power grid resource within the control zone, an importance score is determined for each power grid resource, and each power grid resource is classified according to the importance score to obtain classified power grid resources; For the classified power grid resources within the control zone, a multi-objective optimization function is constructed, which includes the renewable energy consumption target and the frequency stability target. Furthermore, a stability control strategy for the classified power grid resources is generated through the multi-objective optimization function.

2. The method according to claim 1, characterized in that, The equivalent model of the urban power grid is obtained by performing a power grid structure coupling characteristics and power balance index values ​​based on the urban power grid, including: For the urban power grid, the urban power grid is divided into an internal system, boundary nodes and an external system. A block matrix of node voltage equations is constructed. By eliminating the internal nodes in the block matrix, an equivalent network equation containing the self-admittance of the boundary nodes, the mutual admittance between the boundary nodes and the external system and the equivalent injected current is constructed as the equivalent model of the urban power grid. For the generators in the urban power grid, based on the generator power angle difference after disturbance, identify the co-tuning units with consistent power angle curves, aggregate the co-tuning units into equivalent generators, and construct the equivalent model corresponding to the generator based on the equivalent capacity, equivalent inertial time constant and equivalent transient reactance of the equivalent generators. For the static load in the urban power grid, the proportional parameters of constant impedance, constant current and constant power are calculated by power weighted average, and an equivalent model corresponding to the static load is constructed. For the dynamic load in the urban power grid, the equivalent motor parameters are calculated using capacity-weighted average to construct the equivalent model corresponding to the dynamic load.

3. The method according to claim 1, characterized in that, The equivalent model divides the urban power grid into multiple control zones, including: Construct a power grid weighted graph model; the set of nodes in the power grid weighted graph model is used to represent busbars or substations, the set of edges in the power grid weighted graph model is used to represent transmission lines or transformers, and the weights of the edges in the power grid weighted graph model are used to represent the electrical coupling strength between nodes; Modularity is used as a quantitative indicator of the partition quality of the control partition. The Louvain algorithm is used to perform community detection on the power grid weighted graph model. Local optimization and network aggregation are performed iteratively until the modularity reaches the maximum value, thus obtaining the multiple control partitions.

4. The method according to claim 1, characterized in that, The step of determining the importance score for each power grid resource based on its resource characteristics within the control partition includes: Based on the resource type characteristics, resource user importance characteristics, voltage level characteristics, response speed characteristics, and interruption resource quantity characteristics of the power grid resources, multiple resource characteristics of the power grid resources are constructed. The importance score corresponding to each of the resource characteristics is determined by weighted summation; the weight of each resource characteristic is adjusted according to the operation scenario of the urban power grid.

5. The method according to claim 1, characterized in that, The construction of the multi-objective optimization function, which includes the renewable energy consumption target and the frequency stability target, includes: Determine the decision variables; the decision variables include the charging and discharging power sequence of the energy storage system, the curtailment power sequence of new energy sources, and the power reduction sequence of interruptible loads; Construct a first objective function, which is used to maximize the consumption of new energy sources; Construct a second objective function to maximize frequency stability; Based on the first objective function and the second objective function, a multi-objective optimization function is constructed to optimize the decision variables under the condition of satisfying constraints; the constraints include at least power balance constraints, thermal stability control and equipment safety constraints, and regulation resource constraints.

6. The method according to claim 1, characterized in that, The step of generating a stability control strategy for the classified power grid resources through the multi-objective optimization function includes: The comprehensive power grid disturbance index value is obtained, which is determined based on the frequency deviation rate and the frequency change rate. The Sigmoid function is used to map the comprehensive power grid disturbance index value into a dynamic weighting factor; The first objective function and the second objective function are weighted and merged into a single objective function using the dynamic weighting factor. By solving the single objective function, a stable control strategy for the classified power grid resources is generated. The smaller the comprehensive power grid disturbance index value, the more the dynamic weighting factor makes the optimization objective of the single objective function more biased towards new energy consumption; the larger the comprehensive power grid disturbance index value, the more the dynamic weighting factor makes the optimization objective of the single objective function more biased towards frequency stability.

7. A zoned stability control device for massive micro-resources in an urban power grid, characterized in that, The device includes: The modeling and partitioning module is used to perform equivalent modeling of the urban power grid based on the grid structure coupling characteristics and power balance index values ​​of the urban power grid, to obtain the equivalent model, and to divide the urban power grid into multiple control partitions based on the equivalent model; The feature classification module is used to determine the importance score of each power grid resource based on the resource characteristics of each power grid resource in the control partition, and to classify each power grid resource according to the importance score to obtain the classified power grid resources. The stability control module is used to construct a multi-objective optimization function that includes a new energy consumption target and a frequency stability target for the classified grid resources within the control zone, and to generate a stability control strategy for the classified grid resources through the multi-objective optimization function.

8. A computer device comprising a memory and a processor, wherein the memory stores a computer program, characterized in that, When the processor executes the computer program, it implements the steps of the method according to any one of claims 1 to 6.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by a processor, it implements the steps of the method according to any one of claims 1 to 6.

10. A computer program product, comprising a computer program, characterized in that, When the computer program is executed by a processor, it implements the steps of the method according to any one of claims 1 to 6.

Images