A dynamic partitioning method and system for an electricity market

Abstract

This invention discloses a dynamic partitioning method and system for the electricity market, comprising: based on the network topology related to the electricity market, retaining tie-line sections and transmission channels with loads exceeding a threshold in historical operating data; performing preliminary partitioning of the network to determine initial partitions; and calculating and determining the final partitioning result of the electricity market based on the transmission channel information within each initial partition using sensitivity clustering and power flow calculation evaluation. The partitioning method of this invention retains important congestion sections, generating nodes, and load nodes of interest, resulting in a reasonable partitioning outcome.

Description

Technical Field

[0001] This invention relates to the field of power automation, and more specifically to a dynamic partitioning method and system for the power market. Background Technology

[0002] In the actual operation of electricity markets worldwide, three market pricing mechanisms based on marginal cost have emerged: system marginal price, regional marginal price, and nodal marginal price. The system marginal price, under the condition of satisfying unit operating constraints, ranks units by their bids from lowest to highest, with the bid from the unit whose load is satisfied last. The system marginal price does not consider the value difference of electricity in different geographical locations. The nodal marginal price, considering grid congestion, losses, and the marginal cost of each generating unit, represents the lowest cost per unit of new load at a node. The nodal marginal price reflects the value of electricity in time and space. While the nodal marginal price accurately reflects all transmission congestion in grid operation, it faces challenges such as computational complexity. Furthermore, in actual grid operation, transmission congestion typically occurs frequently and noticeably only between certain areas, while within these areas, the probability of congestion is low and the degree of congestion is mild. Therefore, regional marginal price was proposed to simplify and replace the nodal marginal price.

[0003] In markets employing zoning marginal pricing, the most crucial aspect is zoning. Given the complexity of transmission networks and power system operations, the rational delineation of zoning boundaries has always been a difficult and controversial issue. If zoning fails to reflect actual congestion levels, situations may arise where congestion is minimal between zoning areas but severe within zoning areas, leading to significant distortions in the zoning marginal pricing signal and impacting market efficiency.

[0004] The current electricity market zoning methods used in various countries have the following advantages and disadvantages: 1) Geographic zoning: This method is simple and has a clear physical meaning, but it is a static zoning method and cannot handle congestion problems within regions. 2) Social benefit maximization method: This method uses optimization principles, given a pricing objective function for each zone, and calculates the zoning method that maximizes social benefit through large-scale optimization. The computational workload is enormous, and it is prone to multiple solutions. 3) Probabilistic statistical zoning method: This method statistically analyzes historical price data and groups nodes with similar prices into the same zone. It requires collecting a large amount of data, and the data processing workload is very large. Complete data is difficult to obtain at the beginning of electricity market operation. Summary of the Invention

[0005] To address the aforementioned shortcomings in the existing technology, this invention provides a dynamic zoning method for the electricity market, comprising:

[0006] Based on the network topology related to the electricity market, retain tie-line sections and transmission channels with loads exceeding the threshold in historical operating data;

[0007] The network is initially divided to determine the initial partitions;

[0008] Based on the transmission channel information within each initial partition, sensitivity clustering and power flow calculation are used to determine the final partitioning result of the electricity market.

[0009] Preferably, the step of determining the final zoning result of the electricity market based on the transmission channel information within each initial zoning area, using sensitivity clustering and power flow calculation assessment, includes:

[0010] Step S1: Calculate the sensitivity of each node in each initial partition to the transmission channel and obtain the sensitivity matrix;

[0011] Step S2: Based on the sensitivity of each node in the sensitivity matrix, the K-means algorithm is used to cluster all nodes in each partition into multiple groups, determine the sensitivity of each group, and treat a group as a new node.

[0012] Step S3: Calculate the power of each transmission channel under the current partitioning result and the previous partitioning result based on the sensitivity of each new node and the DC power flow algorithm. When the power deviation of the same transmission channel meets the power flow calculation deviation accuracy, the loop ends and the current partitioning result is obtained; otherwise, the partitions at both ends of the transmission channel with the largest power deviation are decomposed to generate a new initial partition and step S1 is executed.

[0013] Preferably, the power deviation of the same transmission channel meets the power flow calculation deviation accuracy, including:

[0014] When the power deviation of the same transmission channel is lower than the accuracy of the power flow calculation deviation.

[0015] Preferably, the sensitivity based on each node employs the K-means algorithm to cluster all nodes within each partition into multiple groups, including:

[0016] Select any number of nodes as cluster centers;

[0017] The similarity between the remaining nodes and each cluster center is calculated based on the sensitivity of each node.

[0018] Based on the similarity between the remaining nodes and each cluster center, the remaining nodes are clustered with the cluster center with the lowest similarity to them;

[0019] In each class, a new cluster center is determined, and the solution is iterated until the convergence condition is met, thus clustering all nodes in each partition into multiple groups.

[0020] Preferably, the sensitivity matrix is ​​as shown in the following formula:

[0021]

[0022] Where: Ψ: sensitivity matrix; [H]: susceptance matrix corresponding to the transmission channel; [B']: susceptance matrix of all lines related to the node; Ψ l,i : The sensitivity of the output of node i to the active power flow of line l, i∈n-1, where n is the number of nodes.

[0023] Preferably, the preliminary division of the network includes:

[0024] The network can be divided according to geographical location or administrative affiliation, or based on blocked lines.

[0025] Based on the same inventive concept, this invention provides a dynamic zoning system for the electricity market, comprising:

[0026] The retention module is used to retain tie-line sections and transmission channels with loads exceeding a threshold in historical operating data, based on the network topology relevant to the electricity market.

[0027] The partitioning module is used to initially partition the network and determine the initial partitions;

[0028] The results module is used to calculate and determine the final partitioning results of the electricity market based on the transmission channel information in each initial partition, using sensitivity clustering and power flow calculation evaluation.

[0029] Preferably, the result module includes:

[0030] Sensitivity unit, used to calculate the sensitivity of each node in each initial partition to the transmission channel, to obtain the sensitivity matrix;

[0031] Clustering unit, used to use the K-means algorithm based on the sensitivity of each node in the sensitivity matrix to cluster all nodes in each partition into multiple groups, determine the sensitivity of each group, and treat a group as a new node;

[0032] The iterative unit is used to calculate the power of each transmission channel under the current partitioning result and the previous partitioning result based on the sensitivity of each new node and the DC power flow algorithm. When the power deviation of the same transmission channel meets the termination condition, the loop ends and the current partitioning result is obtained; otherwise, the partitions at both ends of the transmission channel with the largest power deviation are decomposed to generate a new initial partition and the sensitivity unit is called.

[0033] Preferably, the termination condition in the iterative unit includes:

[0034] When the power deviation of the same transmission channel is lower than the accuracy of the power flow calculation deviation.

[0035] Preferably, the clustering unit includes:

[0036] Selecting sub-units is used to arbitrarily select multiple nodes from each node as cluster centers;

[0037] The computational subunit is used to calculate the similarity between the remaining nodes and each cluster center based on the sensitivity of each node;

[0038] Clustering subunits are used to cluster the remaining nodes with the cluster centers that have the lowest similarity to each cluster center, based on the similarity between the remaining nodes and each cluster center.

[0039] The iterative sub-unit is used to determine new cluster centers in each class, and iteratively solves the problem until the convergence condition is met, clustering all nodes in each partition into multiple groups.

[0040] Compared with the prior art, the beneficial effects of the present invention are as follows:

[0041] The technical solution provided by this invention, based on the network topology related to the electricity market, retains tie-line sections and transmission channels with loads exceeding a threshold in historical operating data; performs preliminary division of the network to determine initial partitions; and, based on the transmission channel information within each initial partition, uses sensitivity clustering and power flow calculation evaluation to determine the final partitioning result of the electricity market. The final partitioning result retains only the transmission channels of interest, resulting in small active power flow deviations for the transmission channels of interest before and after partitioning, a smaller grid size after partitioning, and high performance for market clearing calculations using this partitioning result.

[0042] The technical solution provided by this invention has no impedance parameters for the lines in the partitioned power grid model, only a sensitivity matrix and a stability limit for the retained transmission channels, which is simple to calculate and highly efficient.

[0043] The technical solution provided by this invention decomposes the partitions at both ends of the power transmission channel with the largest power deviation, resulting in a small computational workload and a unique partitioning result.

[0044] The technical solution provided by this invention uses sensitivity clustering and power flow calculation evaluation for calculation, and uses the electrical susceptance data of the nodes. The data acquisition workload is small, easy to implement and reduces the amount of calculation. Attached Figure Description

[0045] Figure 1 A flowchart of a dynamic zoning method for the electricity market provided by this invention;

[0046] Figure 2 This is an overall schematic diagram of the IEEE-118 node power grid model before partitioning in an embodiment of the present invention;

[0047] Figure 3 This is a schematic diagram of IEEE-118 node power grid partition 1 before partitioning in an embodiment of the present invention;

[0048] Figure 4 This is a schematic diagram of IEEE-118 node power grid partition 2 before partitioning in an embodiment of the present invention;

[0049] Figure 5 This is a schematic diagram of IEEE-118 node power grid partition 3 before partitioning in an embodiment of the present invention;

[0050] Figure 6 The following is a detailed algorithm flowchart provided in the embodiments of the present invention;

[0051] In this context, numbers 1 to 118 represent 118 nodes; G represents a generator unit. Detailed Implementation

[0052] To better understand this invention, the following description, in conjunction with the accompanying drawings and examples, will further illustrate the invention.

[0053] The current electricity market zoning methods used in various countries have the following advantages and disadvantages: 1) Geographic zoning: This method is simple and has a clear physical meaning, but it is a static zoning method and cannot handle congestion problems within a region. 2) Social benefit maximization method: This method uses optimization principles, given a pricing objective function for each zone, and calculates the zoning method that maximizes social benefit through large-scale optimization. The computational workload is enormous, and it is prone to multiple solutions. 3) Probabilistic statistical zoning method: This method statistically analyzes historical price data and groups nodes with similar prices into the same zone. It requires collecting a large amount of data, and the data processing workload is very large. Complete data is difficult to obtain at the beginning of electricity market operation.

[0054] The purpose of this invention is to propose a dynamic zoning method for the power market based on sensitivity clustering, which can reflect the congestion of the power grid, retain the important congestion sections, generation nodes and load nodes of interest in the zoning, and minimize the power flow deviation before and after the aggregation of the sections of interest.

[0055] like Figure 1 As shown, the present invention provides a dynamic zoning method for the electricity market, comprising:

[0056] Step S1: Based on the network topology related to the electricity market, retain tie-line sections and transmission channels with loads exceeding the threshold in historical operating data;

[0057] Step S2: Perform preliminary division of the network to determine the initial partitions;

[0058] Step S3: Based on the transmission channel information in each initial partition, sensitivity clustering and power flow calculation are used to calculate and determine the final partitioning result of the electricity market.

[0059] Step S1, based on the network topology related to the electricity market, retains tie-line sections and transmission channels with loads exceeding a threshold in historical operating data, specifically including:

[0060] The transmission channels retained during the zoning process include tie line sections and heavy-load transmission channels with a load of over 95% based on historical operating data.

[0061] Step S2: Perform preliminary partitioning of the network to determine initial partitions, specifically including:

[0062] Traditional methods of zoning mainly include: ① zoning by geographical location or administrative affiliation; ② zoning based on congested lines.

[0063] Step S3: Based on the transmission channel information within each initial partition, sensitivity clustering and power flow calculation are used to determine the final partitioning result of the electricity market, specifically including:

[0064] For multiple nodes within a partition, they are clustered into multiple groups. The power generation and load on the nodes within a group are automatically assigned to the corresponding group, and the sensitivity factor of the group is determined according to the arithmetic mean and weighted average.

[0065] The initial partitioning results are evaluated as follows: ① Calculate the power flow of the full model to obtain the power of each reserved transmission channel; ② Calculate the power flow of the simplified model to obtain the power of each reserved transmission channel; ③ Calculate the deviation between the calculation results of the two models. If the deviation is lower than a pre-set threshold, the partitioning is considered to meet the requirements.

[0066] If the power error of a transmission channel exceeds a preset threshold, the sections at both ends of the transmission channel are decomposed, the number of sections is increased, and the previous sectioning process is repeated. The new sections are then re-evaluated until the deviation is lower than the preset threshold.

[0067] 1. Sensitivity Algorithm

[0068] Based on the DC power flow algorithm, the general expression for the active power flow of line ij is:

[0069] P ij =B ij (θ i -θ j (1)

[0070] In the formula, i and j are the first and last nodes of the line, respectively, and P ij For the active power flow at the starting point of line ij, θi and θ j Let B be the phase angle of the bus voltage at both ends of line ij. ij Let be the susceptance of line ij.

[0071] The injected power at a node is the sum of the power emitted by each branch of that node; therefore, the node power balance equation is:

[0072]

[0073] In the formula, P i Let be the injected power of node i, and n be the total number of nodes in the network.

[0074] Substituting equation (1) into equation (2), we get:

[0075]

[0076] In the formula, B is the sum of the susceptance values ​​of all lines directly connected to this node. i ' j =-B ij The susceptance value of the corresponding line is negative; B i ' i and B i ' j The above equation holds true for all n-1 nodes except the reference node, forming the diagonal and off-diagonal elements of matrix [B']. Therefore, the nodal power balance equation for an n-node power system can be expressed as:

[0077] [P]=[B'][θ] (4)

[0078] In the formula:

[0079] Where [P] and [θ] are the n-1 order active power injection and voltage phase angle vectors, respectively, excluding the quantities related to the balancing node as the angle reference point.

[0080] According to the formula (1) for the active power of a branch, its incremental form can be written as:

[0081]

[0082] Wherein, H is defined ij,i =B ij H ij,j =-B ij .

[0083] All branches have power increment equations of the form (5), which can be written in matrix form.

[0084] [ΔP l ]=[H][Δθ] (6)

[0085] In the formula:

[0086] (N l (Number of branches)

[0087] Among them, [ΔP l [H] is the branch active power increment vector, [Δθ] is the node voltage phase angle increment vector, and [H] is a k-row (n-1) matrix whose elements are defined as H k,i =B ij H k,j =-B ij This means that if a branch number k corresponds to start and end node numbers i and j, then the element in row k, column i of the [H] matrix is ​​equal to the susceptance H of that branch. k,i =B ij For an element in row k and column j equal to the electrical charge of that branch, take the negative sign H. k,j =-B ij .

[0088] Considering that the increments of the node voltage phase angle are all relative to the slack node, the equation does not include the slack node, and its dimension is n-1. The number of equations in equation (6) is equal to the number of branches N. l Moreover, these equations are linearly independent.

[0089] Linearizing and transforming equation (4) yields:

[0090] [Δθ]=[B'] -1 [ΔP] (7)

[0091] Substituting (7) into equation (6) yields:

[0092]

[0093]

[0094] Where [Ψ] has a dimension of N. l ×(n-1), called the branch active power sensitivity matrix, reflects the impact of the increase in active power injection at a single node on the active power of the branch. The corresponding element Ψ in [Ψ] l,i [ΔP] represents the sensitivity of the output of node i to the active power flow of line l, and [ΔP] represents the change in node power.

[0095] 2. Cluster analysis method

[0096] Clustering is the process of grouping data items into multiple classes or clusters. The differences between classes should be as large as possible, while the differences within classes should be as small as possible. This is the principle of "minimizing the similarity between classes and maximizing the similarity within classes".

[0097] Commonly used clustering algorithms include the K-means algorithm. The K-means algorithm works as follows: First, select k objects from n data objects as initial cluster centers. For the remaining objects, assign them to the clusters most similar to these cluster centers based on their similarity (distance). Then, calculate the cluster center of each new cluster (the mean of all objects in that cluster). Repeat this process until the convergence condition is met.

[0098] In the process of power market zoning, it is often desirable to retain transmission channels with the smallest possible power flow deviation. Sensitivity determines the distribution of power flow on the lines when the injected power at power network nodes changes. Therefore, sensitivity is used as an indicator to measure the contribution of nodes to the power flow of retained transmission channels.

[0099] First, calculate the sensitivity matrix Ψ of the full model:

[0100]

[0101] In the formula, ψ kN The sensitivity of branch k to active power injection at node N is represented by the sensitivity matrix Ψ, where each row represents a branch and each column represents a node.

[0102] Then, extract the rows corresponding to the transmission channels that need to be retained from the sensitivity matrix Ψ to form a new matrix Ψ′:

[0103]

[0104] ψ′ can be written as ψ′=[h1,h2,…,h i ], where each h i This represents the sensitivity column vector of a node i in the original network model to all retained transmission channels.

[0105] By inputting the Ψ' matrix into the K-means algorithm, the nodes can be clustered according to their sensitivity based on the specified number of clusters, and nodes with similar impact on the power flow of the retained transmission channel can be grouped together.

[0106] Example 2

[0107] The network model of the dynamic partitioning method for the electricity market designed in this invention is as follows: Figure 2 As shown, because Figure 2 The topology diagram shown is quite complex, so it needs to be clearly represented. Figure 2 ,use Figure 3 , Figure 4 and Figure 5 Will Figure 2 It is divided into three parts, with a total of 118 nodes, that is Figure 3 , Figure 4 and Figure 5 The number, and 188 branch roads; Figure 3 , Figure 4 and Figure 5 In this context, G represents the generator set.

[0108] Nodes 33, 34, 38, and 24 in partition 3 are connected to nodes 15, 19, 30, and 23 in partition 1, respectively; node 70 in partition 3 is connected to nodes 74 and 75 in partition 2; node 69 in partition 3 is connected to nodes 77 and 75 in partition 2; and node 68 in partition 3 is connected to nodes 77 and 81 in partition 2.

[0109] In this embodiment, the retained power transmission channels include the following branches: Node 23->Node 24, Node 15->Node 33, Node 19->Node 34, Node 30->Node 38, Node 70->Node 74, Node 70->Node 75, Node 69->Node 75, Node 69->Node 77, and Node 68->Node 81.

[0110] 1) The core processing flow of this invention is as follows: Figure 6 As shown;

[0111] 2) First, the power grid model is divided into three zones according to existing methods, namely... Figure 3 Partition 1 in Figure 4 Partition 2 and Figure 5 Partition 3 is in the middle; Partition 1 and Partition 3 are connected by 4 tie lines, and Partition 3 and Partition 2 are connected by 5 tie lines.

[0112] 3) Based on the full network model, the sensitivity matrix [Ψ] of the power of each branch to the nodes is calculated using the DC power flow model;

[0113] 4) Set the power flow deviation between the reserved transmission channel and the full model power flow to 4%. That is, if the power flow deviation between the reserved transmission channel and the full model power flow is less than 4% after partitioning, the partitioning is considered to meet the requirements.

[0114] 5) Repeat the sensitivity clustering partitioning and evaluate the power flow deviation between the retained transmission channels and the full model power flow after partitioning until the power flow deviation between the retained transmission channels and the full model meets the requirements.

[0115] 6) The power flow deviations between the reserved transmission channels in three zones, six zones, and twenty-four zones and the full model power flow are shown in Appendix Tables 1, 2, and 3:

[0116] Appendix 1: Power Flow Comparison of Retained Transmission Channels in the Initial Three Zones

[0117] Preserve transmission channels Full model branch power flow Simplified model trend relative error Branch road: 23->24 21.16 39.14 84.96％ Branch road: 15->33 11.1 17.24 55.31％ Branch road: 19->34 0.2 1.73 766.35％ Branch road: 30->38 80.55 54.89 -31.86％ Branch road: 70->74 16.18 8.88 -45.13％ Branch road: 70->75 0.26 -8.70 -3444.39％ Branch road: 69->75 96.37 73.93 -23.29％ Branch road: 69->77 40.42 -4.59 -111.36％ Branch road: 68->81 -57.23 26.49 -146.28％

[0118] Appendix 2: Power Flow Comparison of Retained Transmission Channels in the Six Zones

[0119] Branch Road Name Full model branch power flow Simplified model trend relative error Branch road: 23->24 21.16 12.45 -41.15％ Branch road: 15->33 11.1 16.35 47.34％ Branch road: 19->34 0.2 4.52 2158.71％ Branch road: 30->38 80.55 79.67 -1.09％ Branch road: 70->74 16.18 8.41 -48.04％ Branch road: 70->75 0.26 -9.27 -3665.58％ Branch road: 69->75 96.37 99.93 3.70％ Branch road: 69->77 40.42 31.08 -23.11％ Branch road: 68->81 -57.23 -34.15 -40.33％

[0120] Appendix 3: Power Flow Comparison of Retained Transmission Channels in Twenty-Four Districts

[0121] Branch Road Name Full model branch power flow Simplified model trend relative error Branch road: 23->24 21.16 20.74 -2.01％ Branch road: 15->33 11.1 10.99 -0.96％ Branch road: 19->34 0.2 -0.15 - Branch road: 30->38 80.55 81.42 1.08％ Branch road: 70->74 16.18 16.22 0.27％ Branch road: 70->75 0.26 0.31 - Branch road: 69->75 96.37 96.97 0.62％ Branch road: 69->77 40.42 41.91 3.68％ Branch road: 68->81 -57.23 -59.41 3.81％

[0122] The results show that after reaching twenty-four zones, the power flow deviation between the reserved transmission channels and the full model power flow reaches the given deviation requirement (4%). The reserved transmission channels in the 24 zones and the sensitivity of the zones to the reserved transmission channels are shown in Appendix Table 4.

[0123] Appendix Table 4 Results of 24 Partitions of IEEE 118 Nodes

[0124]

[0125]

[0126]

[0127] In this embodiment, the full model branch power flow represents the sensitivity of all nodes in the power grid model to the retention of transmission channels, while the simplified model power flow represents the sensitivity of each clustering result as a new node to the retention of transmission channels after clustering.

[0128] Based on the same inventive concept, this embodiment of the invention also provides a dynamic zoning system for an electricity market, comprising:

[0129] The retention module is used to retain tie-line sections and transmission channels with loads exceeding a threshold in historical operating data, based on the network topology relevant to the electricity market.

[0130] The partitioning module is used to initially partition the network and determine the initial partitions;

[0131] The results module is used to calculate and determine the final partitioning results of the electricity market based on the transmission channel information in each initial partition, using sensitivity clustering and power flow calculation evaluation.

[0132] In this embodiment, the result module includes:

[0133] Sensitivity unit, used to calculate the sensitivity of each node in each initial partition to the transmission channel, to obtain the sensitivity matrix;

[0134] Clustering unit, used to use the K-means algorithm based on the sensitivity of each node in the sensitivity matrix to cluster all nodes in each partition into multiple groups, determine the sensitivity of each group, and treat a group as a new node;

[0135] The iterative unit is used to calculate the power of each transmission channel under the current partitioning result and the previous partitioning result based on the sensitivity of each new node and the DC power flow algorithm. When the power deviation of the same transmission channel meets the termination condition, the loop ends and the current partitioning result is obtained; otherwise, the partitions at both ends of the transmission channel with the largest power deviation are decomposed to generate a new initial partition and the sensitivity unit is called.

[0136] In this embodiment, the termination condition in the iterative unit includes:

[0137] When the power deviation of the same transmission channel is lower than the accuracy of the power flow calculation deviation.

[0138] In this embodiment, the clustering unit includes:

[0139] Selecting sub-units is used to arbitrarily select multiple nodes from each node as cluster centers;

[0140] The computational subunit is used to calculate the similarity between the remaining nodes and each cluster center based on the sensitivity of each node;

[0141] Clustering subunits are used to cluster the remaining nodes with the cluster centers that have the lowest similarity to each cluster center, based on the similarity between the remaining nodes and each cluster center.

[0142] The iterative sub-unit is used to determine new cluster centers in each class, and iteratively solves the problem until the convergence condition is met, clustering all nodes in each partition into multiple groups.

[0143] Those skilled in the art will understand that embodiments of this application can be provided as methods, systems, or computer program products. Therefore, this application can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, this application can take the form of a computer program product embodied on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0144] This application is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of this application. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart... Figure 1 a process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0145] These computer program instructions may also be stored in a computer readable memory that can direct a computer or other programmable data processing device to work in a specific manner, so that the instructions stored in the computer readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 a process or multiple processes and / or boxes Figure 1 The function specified in one or more boxes.

[0146] The above are merely embodiments of the present invention and are not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention are included within the scope of the claims of the present invention pending approval.

Claims

1. A dynamic zoning method for an electricity market, characterized in that, include: Based on the network topology related to the electricity market, retain tie-line sections and transmission channels with loads exceeding the threshold in historical operating data; The network is initially divided to determine the initial partitions; Based on the transmission channel information within each initial partition, sensitivity clustering and power flow calculation evaluation are used to calculate and determine the final partitioning result of the electricity market. The final zoning result of the electricity market is determined by calculating and evaluating the transmission channel information within each initial zoning area using sensitivity clustering and power flow calculation, including: Step S1: Calculate the sensitivity of each node in each initial partition to the transmission channel and obtain the sensitivity matrix; Step S2: Based on the sensitivity of each node in the sensitivity matrix, the K-means algorithm is used to cluster all nodes in each partition into multiple groups, determine the sensitivity of each group, and treat a group as a new node. Step S3: Calculate the power of each transmission channel under the current partitioning result and the previous partitioning result based on the sensitivity of each new node and the DC power flow algorithm. When the power deviation of the same transmission channel meets the power flow calculation deviation accuracy, the loop ends and the current partitioning result is obtained; otherwise, the partitions at both ends of the transmission channel with the largest power deviation are decomposed to generate a new initial partition and step S1 is executed.

2. The method as described in claim 1, characterized in that, The power deviation of the same transmission channel meets the power flow calculation deviation accuracy requirements, including: When the power deviation of the same transmission channel is lower than the accuracy of the power flow calculation deviation.

3. The method as described in claim 1, characterized in that, The sensitivity based on each node employs the K-means algorithm, clustering all nodes within each partition into multiple groups, including: Select any number of nodes as cluster centers; The similarity between the remaining nodes and each cluster center is calculated based on the sensitivity of each node. Based on the similarity between the remaining nodes and each cluster center, the remaining nodes are clustered with the cluster center with the lowest similarity to them; In each class, a new cluster center is determined, and the solution is iterated until the convergence condition is met, thus clustering all nodes in each partition into multiple groups.

4. The method as described in claim 1, characterized in that, The sensitivity matrix is ​​shown in the following formula: In the formula: Ψ: sensitivity matrix; [H]: susceptance matrix corresponding to the transmission channel; [B']: susceptance matrix of all lines related to the node; Ψ l,i : The sensitivity of the output of node i to the active power flow of line l, i∈n-1, where n is the number of nodes.

5. The method as described in claim 1, characterized in that, The preliminary division of the network includes: The network can be divided according to geographical location or administrative affiliation, or based on blocked lines.

6. A dynamic zoning system for an electricity market, characterized in that, include: The retention module is used to retain tie-line sections and transmission channels with loads exceeding a threshold in historical operating data, based on the network topology relevant to the electricity market. The partitioning module is used to initially partition the network and determine the initial partitions; The results module is used to calculate and determine the final partitioning results of the electricity market based on the transmission channel information in each initial partition, using sensitivity clustering and power flow calculation evaluation. The result module includes: Sensitivity unit, used to calculate the sensitivity of each node in each initial partition to the transmission channel, to obtain the sensitivity matrix; Clustering unit, used to use the K-means algorithm based on the sensitivity of each node in the sensitivity matrix to cluster all nodes in each partition into multiple groups, determine the sensitivity of each group, and treat a group as a new node; The iterative unit is used to calculate the power of each transmission channel under the current partitioning result and the previous partitioning result based on the sensitivity of each new node and the DC power flow algorithm. When the power deviation of the same transmission channel meets the termination condition, the loop ends and the current partitioning result is obtained; otherwise, the partitions at both ends of the transmission channel with the largest power deviation are decomposed to generate a new initial partition and the sensitivity unit is called.

7. The system as described in claim 6, characterized in that, The termination conditions in the iterative unit include: When the power deviation of the same transmission channel is lower than the accuracy of the power flow calculation deviation.

8. The system as described in claim 6, characterized in that, The clustering unit includes: Selecting sub-units is used to arbitrarily select multiple nodes from each node as cluster centers; The computational subunit is used to calculate the similarity between the remaining nodes and each cluster center based on the sensitivity of each node; Clustering subunits are used to cluster the remaining nodes with the cluster centers that have the lowest similarity to each cluster center, based on the similarity between the remaining nodes and each cluster center. The iterative sub-unit is used to determine new cluster centers in each class, and iteratively solves the problem until the convergence condition is met, clustering all nodes in each partition into multiple groups.

Images