Power grid security boundary feature extraction method and system based on data driving
Through a data-driven method, the power grid safety boundary features are extracted using weighted oblique decision trees and trend data, and the problem of inaccurate power grid safety boundary capture in the existing technology is solved, and the rapid and accurate safety identification of the operating status of the power system is achieved.
Patent Information
- Application Number
- CN202411034115.2
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-07-30
- Publication Date
- 2024-11-22
AI Technical Summary
The prior art cannot capture the safety boundaries of the power grid in real time and accurately, resulting in difficulty in identifying the safety status of the power system. The feature extraction method has high-dimensional and noise interference problems, which affects its effectiveness.
Using a data-driven method, the power grid trend data is generated through trend calculation, the power grid safety boundary is divided using a weighted oblique decision tree to form a safety rule group, and the boundary characteristics are tightened through unsafe data to reduce the pseudo-positive category rate and achieve rapid and accurate security identification.
It improves the accuracy and real-time performance of the power grid safety boundary feature extraction, reduces the misjudgment rate of unsafe operating points, and enhances the safety identification ability of the power system operating status.
Smart Images

Figure CN119005502A_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the field of safe operation of power systems, and in particular to a data-driven method and system for extracting power grid safety boundary features. Background Art
[0002] Grid safety boundary feature extraction is a key link to ensure the stable operation of the power system and prevent faults. However, there are still some shortcomings in the current grid safety boundary feature extraction, mainly including the following aspects: The feature dimension is too high: The grid system is complex and involves many parameters, which leads to the extracted feature dimension being too high. This not only increases the computational complexity, but also may introduce redundant information, affecting the effectiveness of feature extraction. The grid is disturbed by various noises during operation, such as equipment noise, environmental noise, etc. These noises may mask or distort the true characteristics of the grid safety boundary, resulting in inaccurate feature extraction. The increase in the penetration rate of new energy has a great impact on the safe operation of the grid. The high-dimensional observations of large-scale power grids make it difficult to determine whether a given operating point is safe. Existing feature extraction methods cannot capture the grid safety boundary in real time and accurately, which limits their effectiveness in practical applications. Summary of the invention
[0003] The purpose of the present invention is to provide a data-driven power grid safety boundary feature extraction method and system to overcome the problem that the feature extraction method in the prior art cannot capture the power grid safety boundary in real time and accurately.
[0004] In order to achieve the above object, the present invention adopts the following technical solution:
[0005] In a first aspect, the present invention provides a data-driven power grid security boundary feature extraction method, comprising the following steps:
[0006] S1, using power flow calculation to generate power grid power flow data;
[0007] S2, using a weighted oblique decision tree to divide the power grid flow data to obtain the power grid safety boundary of the weighted oblique decision tree node;
[0008] S3, based on the path of the weighted oblique decision tree, the power grid safety boundary of all nodes in the weighted oblique decision tree is extracted to form a safety rule group, and the unsafe power grid data near the power grid safety boundary is used to tighten the safety rule group to obtain the power grid safety boundary characteristics.
[0009] Preferably, power flow calculation is used to generate power grid power flow data, which specifically includes the following steps:
[0010] The simulation is used to calculate the power flow and generate the grid operation status data with safety tags, and the branch head end power and branch end power corresponding to different operation points are obtained;
[0011] Calculate branch load factor based on branch active transmission capacity, branch head end power and branch end power;
[0012] Calculate the grid status label based on the branch load rate and the set threshold;
[0013] The power grid operation state data including the safety tag is generated based on the power grid state tag and the power grid operation state vector.
[0014] Preferably, the power grid operation state vector selects unit output and load as features.
[0015] Preferably, using a weighted oblique decision tree to divide the power grid flow data to obtain the power grid safety boundary of the weighted oblique decision tree node specifically includes the following steps:
[0016] The solution corresponding to minimizing the weighted information entropy is used to obtain the parameters of the linear partition in each node of the weighted oblique decision tree;
[0017] The weighted information entropy is used as the objective function, and the parameters of the linear division in the nodes of the weighted oblique decision tree are used as the coefficients before the features in the power grid security boundary to obtain the power grid security boundary.
[0018] Preferably, extracting the power grid safety boundaries of all nodes in the weighted oblique decision tree based on the path of the weighted oblique decision tree to form a safety rule group specifically includes the following steps:
[0019] All nodes from the root node to a leaf node are called a path. With the help of tree traversal, the divisions on the oblique decision tree path are extracted to form an inequality group, thereby extracting a safety rule group representing the relationship between the unit output and the safety of power grid operation.
[0020] Preferably, the nodes of the tree include serial number, division, tree depth, whether it is a leaf node, left child node serial number and right child node serial number; the division includes feature index, feature coefficient and threshold.
[0021] Preferably, tightening the power grid safety boundary according to the unsafe power grid data near the power grid safety boundary specifically includes the following steps:
[0022] a. Calculate the Euclidean distance from the pseudo-positive samples in path p to each hyperplane in the path;
[0023] b. Obtain the hyperplane corresponding to the minimum Euclidean distance and update the threshold of the hyperplane;
[0024] c. Determine whether other pseudo positive samples belong to path p based on the hyperplane of the updated threshold, and remove pseudo positive samples that do not belong to path p;
[0025] d. Calculate the Euclidean distances from other pseudo-positive samples in path p to each hyperplane in the path, and repeat steps b and c above until the pseudo-positive samples in path p are empty. This completes the tightening of the security rule group, and the tightened security rule group is the obtained power grid security boundary feature.
[0026] In a second aspect, the present invention provides a data-driven power grid safety boundary feature extraction system, comprising a data sorting module, a safety boundary module and a feature extraction module;
[0027] Data sorting module, used for power flow calculation to generate power grid power flow data;
[0028] The safety boundary module uses a weighted oblique decision tree to divide the power grid flow data to obtain the power grid safety boundary of the weighted oblique decision tree node;
[0029] The feature extraction module extracts the power grid safety boundaries of all nodes in the weighted oblique decision tree based on the path of the weighted oblique decision tree to form a safety rule group, and uses the unsafe power grid data near the power grid safety boundary to tighten the safety rule group to obtain the power grid safety boundary characteristics.
[0030] Preferably, power flow calculation is used to generate power grid power flow data, which specifically includes the following steps:
[0031] The simulation is used to calculate the power flow and generate the grid operation status data with safety tags, and the branch head end power and branch end power corresponding to different operation points are obtained;
[0032] Calculate branch load factor based on branch active transmission capacity, branch head end power and branch end power;
[0033] Calculate the grid status label based on the branch load rate and the set threshold;
[0034] The power grid operation state data including the safety tag is generated based on the power grid state tag and the power grid operation state vector.
[0035] Preferably, the power grid operation state vector selects unit output and load as features.
[0036] Preferably, using a weighted oblique decision tree to divide the power grid flow data to obtain the power grid safety boundary of the weighted oblique decision tree node specifically includes the following steps:
[0037] The solution corresponding to minimizing the weighted information entropy is used to obtain the parameters of the linear partition in each node of the weighted oblique decision tree;
[0038] The weighted information entropy is used as the objective function, and the parameters of the linear division in the nodes of the weighted oblique decision tree are used as the coefficients before the features in the power grid security boundary to obtain the power grid security boundary.
[0039] Preferably, extracting the power grid safety boundaries of all nodes in the weighted oblique decision tree based on the path of the weighted oblique decision tree to form a safety rule group specifically includes the following steps:
[0040] All nodes from the root node to a leaf node are called a path. With the help of tree traversal, the divisions on the oblique decision tree path are extracted to form an inequality group, thereby extracting a safety rule group representing the relationship between the unit output and the safety of power grid operation.
[0041] Preferably, the nodes of the tree include serial number, division, tree depth, whether it is a leaf node, left child node serial number and right child node serial number; the division includes feature index, feature coefficient and threshold.
[0042] Preferably, tightening the power grid safety boundary according to the unsafe power grid data near the power grid safety boundary specifically includes the following steps:
[0043] a. Calculate the Euclidean distance from the pseudo-positive samples in path p to each hyperplane in the path;
[0044] b. Obtain the hyperplane corresponding to the minimum Euclidean distance and update the threshold of the hyperplane;
[0045] c. Determine whether other pseudo positive samples belong to path p based on the hyperplane of the updated threshold, and remove pseudo positive samples that do not belong to path p;
[0046] d. Calculate the Euclidean distances from other pseudo-positive samples in path p to each hyperplane in the path, and repeat steps b and c above until the pseudo-positive samples in path p are empty. This completes the tightening of the security rule group, and the tightened security rule group is the obtained power grid security boundary feature.
[0047] In a third aspect, the present invention provides a computer device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements the steps of the above-mentioned data-driven power grid safety boundary feature extraction method when executing the computer program.
[0048] In a fourth aspect, the present invention provides a computer-readable storage medium, wherein the computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the steps of the above-mentioned data-driven power grid safety boundary feature extraction method are implemented.
[0049] Compared with the prior art, the present invention has the following beneficial technical effects:
[0050] The invention discloses a data-driven power grid safety boundary feature extraction method, which generates power grid flow data by power flow calculation, builds a power grid safety boundary identification model based on the power grid flow data, extracts power grid safety boundary features from the power system operation state based on the power grid safety boundary identification model, tightens the power grid safety rules by using the power grid safety boundary features, and reduces the false positive rate of the model by tightening the safety boundary, thereby realizing rapid and accurate identification of the safety of the power system operation state. Compared with the traditional axis-parallel decision tree method, the safety rules extracted by the proposed method have a higher degree of matching with the actual operation state.
[0051] Preferably, a data-driven power grid safety boundary identification model is constructed based on a tightened weighted oblique decision tree to achieve safety rule extraction and rule tightening. Based on the weighted oblique decision tree, the safety boundary is fitted to the power flow data with power grid safety labels, and the rule extraction is achieved by traversing the tree. The safe area can be found by traversing the tree from the root node, and the parameters of the oblique decision tree model are fine-tuned considering critical samples to avoid misjudging the unsafe operating state as a safe state. BRIEF DESCRIPTION OF THE DRAWINGS
[0052] Figure 1 The figure is a flow chart of a data-driven power grid safety boundary feature extraction method in an embodiment of the present invention.
[0053] Figure 2 Schematic diagram of a power grid security rule group in an embodiment of the present invention.
[0054] Figure 3 Schematic diagram of the feasible domain of power grid security in an embodiment of the present invention.
[0055] Figure 4 Schematic diagram of power grid security rule results in an embodiment of the present invention.
[0056] Figure 5 Schematic diagram of a data-driven power grid safety boundary feature extraction system in an embodiment of the present invention. DETAILED DESCRIPTION
[0057] In order to enable those skilled in the art to better understand the scheme of the present invention, the technical scheme in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work should fall within the scope of protection of the present invention.
[0058] It should be noted that the terms "first", "second", etc. in the specification and claims of the present invention and the above-mentioned drawings are used to distinguish similar objects, and are not necessarily used to describe a specific order or sequence. It should be understood that the data used in this way can be interchanged where appropriate, so that the embodiments of the present invention described herein can be implemented in an order other than those illustrated or described herein. In addition, the terms "including" and "having" and any variations thereof are intended to cover non-exclusive inclusions, for example, a process, method, system, product or device that includes a series of steps or units is not necessarily limited to those steps or units that are clearly listed, but may include other steps or units that are not clearly listed or inherent to these processes, methods, products or devices.
[0059] like Figure 1 As shown, the present invention provides a data-driven power grid safety boundary feature extraction method, which can more accurately obtain power grid safety boundary features, reduce the safety boundary of the hyperplane group fitting, effectively reduce the number of constraints, and tighten the power grid safety rules according to the power grid safety boundary features, effectively reducing the false positive rate of misjudging an unsafe state as safe, and specifically includes the following steps:
[0060] S1, using power flow calculation to generate power grid power flow data;
[0061] S2, using a weighted oblique decision tree to divide the power grid flow data to obtain the power grid safety boundary of the weighted oblique decision tree node;
[0062] S3, based on the path of the weighted oblique decision tree, the power grid safety boundary of all nodes in the weighted oblique decision tree is extracted to form a safety rule group, namely, the power grid safety rule; the unsafe power grid data near the power grid safety boundary is used to tighten the safety rule group to obtain the power grid safety boundary characteristics, which are used to quickly and accurately judge the power grid operation status and reduce the misjudgment of unsafe operation points.
[0063] This application uses simulation to calculate the power flow and generate the power grid operation status data with safety tags. First, the head end power P of the branch corresponding to different operating points is obtained. ij and terminal power P ji :
[0064]
[0065]
[0066] In the formula, and Respectively represent the power injection of the unit and the load; V i and V j Respectively represent the voltage at the beginning and end of the branch; S ij and Sji Respectively represent the head end and the end branch flow; Y i s represents the parallel admittance at node i; Y ij represents the branch admittance; Indicates branch charging susceptance; T ij Indicates transformer ratio and phase shift; (·) * represents conjugation; E and E R Respectively represent the forward and reverse directions of the branch; P ij and P ji Respectively represent the active power at the beginning and end of the branch; It means taking the real part of a complex number.
[0067] The branch load rate is calculated based on the active transmission capacity of the branch, as shown in formula (5):
[0068]
[0069] Where P ij Indicates the active power of the branch; Represents the active transmission capacity of the branch; ρ k Indicates the active load rate of the branch; max(P ij ,P ji ) represents the maximum value of the active power of the head-end and the tail-end branches.
[0070] The safety of the power grid is line-oriented. If all lines are not overloaded, the power grid is safe; if one branch is overloaded, the power grid is unsafe. Therefore, the safety of the operation state is evaluated by giving a threshold, and the power grid state label, that is, the power grid flow data containing the safety label, is calculated, as shown in formula (6):
[0071]
[0072] In the formula, y i Indicates the grid status label, where 0 indicates a safe state and 1 indicates an unsafe state; N br Indicates the number of branches, Indicates rounding down.
[0073] Based on the acquired grid state label and grid operation state vector, the grid flow data D of the power system state with a safety label is generated, as shown in equations (7)-(8):
[0074]
[0075] In the formula, the power grid operation state vector Select unit output p g and load p d As a feature.
[0076] The safety of the power system operation state requires a safe definition of the power system operation state. This application uses the power grid operation safety boundary to define the power system operation state; the power grid operation safety boundary refers to a hyperplane that can divide the safe operation point and the unsafe operation point of the power system operation state. The process of identifying the power system operation state is to continuously adjust the parameters of the hyperplane so that the hyperplane has the classification performance that meets the requirements for safe operation points and unsafe operation points. As a classifier for power grid safety assessment, weighted oblique decision trees (WODT) have strong learning ability and can linearly combine multiple power injection features to form a safety boundary until a complete tree is fitted. The parameters of the linear division in each node of the weighted oblique decision tree are obtained by minimizing the solution corresponding to the weighted information entropy; the coefficient before the feature in the power grid safety boundary is the parameter θ of the linear division in the node of the weighted oblique decision tree. Therefore, the power grid safety boundary fitted from the power grid flow data can accurately reflect the actual operation of the system.
[0077] The present application divides power grid flow data based on a weighted oblique decision tree to obtain a power grid safety boundary of a weighted oblique decision tree node, which is used to generate a hyperplane that divides safe and unsafe operating points.
[0078] The specific steps include:
[0079] Step 2-1: Calculate the instance weight. Assume that the linear partition parameter θ in the nodes of the weighted oblique decision tree is used, and use the sigmoid function to calculate the weight of each sample, as shown in equations (9)-(11):
[0080]
[0081] represents the weight of each sample in the left child node to the hyperplane, represents the weight of each sample in the right child node to the hyperplane, σ represents the sigmoid function, x i represents the sample, θ T represents the transpose of the hyperplane parameters;
[0082] Step 2-2: Binary splitting of a root node can generate two child nodes, left and right, and calculate the weight W of the left child node respectively. L (θ) and the weight W of the right child node R (θ), as shown in equations (12)-(13):
[0083]
[0084] S LRepresents the left child node set, S R Represents the right child node set, y i Indicates the label of the sample;
[0085] Step 2-3: Calculate the weight W of the left child node of each class k L (θ) and the weight W of the right child node k R (θ), as shown in equations (14)-(16):
[0086]
[0087] k represents the category; I represents the indicator function, which is 1 if they are equal and 0 if they are not equal;
[0088] Step 2-4: Calculate the weight ratio of each category in the left child node and the right child node, as shown in equations (17)-(18):
[0089]
[0090] Indicates the weight ratio of the k-th class samples in the left child node, Indicates the weight ratio of the k-th class samples in the right child node;
[0091] Step 2-5: Calculate the information entropy H of the left child node L (θ) and the information entropy H of the right child node R (θ), as shown in equations (19)-(20):
[0092]
[0093] C represents the total number of categories.
[0094] Step 2-6: Use weighted information entropy as the objective function to obtain the power grid security boundary; simplify it as shown in formula (21):
[0095]
[0096] E(θ) represents the weighted information entropy, which is used to measure the homogeneity of the target variable within the subset.
[0097] The above power grid safety boundaries are stored in the nodes of the tree, and one node corresponds to one power grid safety boundary. The safety rule extraction is based on the traversal of the tree to extract the safety boundaries in the nodes to form a safety rule group, such as Figure 2 shown.
[0098] The path based on the weighted oblique decision tree extracts the power grid safety boundaries of all nodes in the weighted oblique decision tree to form a safety rule group, that is, power grid safety rule generation, which specifically includes the following steps:
[0099] All nodes from the root node to a leaf node are called a path. With the help of tree traversal, the divisions on the oblique decision tree path are extracted to form an inequality group, thereby extracting a safety rule group representing the relationship between the unit output and the safety of power grid operation.
[0100] The node information of the tree includes sequence number, partition, tree depth, whether it is a leaf node, left child node sequence number and right child node sequence number. The partition includes feature index (corresponding to the selected power grid feature), feature coefficient (coefficient corresponding to the selected power grid feature, i.e. hyperplane parameter) and threshold.
[0101] The above division adopts a soft division method. The above weighted oblique decision tree method cannot guarantee that the false positive rate in the classification problem is reduced to zero, and there will be a phenomenon of judging unsafe samples as safe. In the safe operation of the power grid, the dispatcher cannot accept that the line is unsafe and is judged as safe; therefore, it is necessary to translate the power grid safety boundary (hyperplane) according to the unsafe power grid data near the power grid safety boundary to achieve the tightening of the above safety rules, and obtain the power grid safety boundary characteristics after tightening.
[0102] The specific steps of tightening the power grid security boundary based on the unsafe power grid data near the power grid security boundary include:
[0103] The unsafe power grid data near the power grid safety boundary are pseudo positive samples in the path from the root node to a leaf node; the pseudo positive samples are sample data inconsistent with the test data in the data of the power grid safety boundary division of the weighted oblique decision tree node obtained by dividing the power grid flow data using the weighted oblique decision tree.
[0104] Step 4-1: Calculate the pseudo-positive samples in path p to the i-th hyperplane π in the path i The Euclidean distance is:
[0105]
[0106] in, represents the oth pseudo-positive sample in path p; θ i Parameters of the linear partition among the nodes in the i-th hyperplane.
[0107] Step 4-2: Get the i-th hyperplane π corresponding to the minimum Euclidean distance i And update its threshold:
[0108]
[0109] Among them, θ * represents the non-constant parameter of the hyperplane π; t i represents the i-th hyperplane πi The constant term of , that is, the threshold.
[0110] Step 4-3: Remove the pseudo-positive samples from the queue: Determine other pseudo-positive samples based on the updated threshold Does it belong to path p? If Then remove Otherwise, it is not removed;
[0111] Step 4-4: Repeat the above steps until the pseudo-positive samples in path p are empty, and the tightening of the security rule group is completed. The tightened security rule group is the obtained power grid safety boundary feature.
[0112] This application fits the safety boundary of the power flow data with power grid safety labels based on the weighted oblique decision tree, realizes the extraction of power grid safety rules with the help of the traversal of the weighted oblique decision tree, and can find the safe area by traversing the tree from the root node, tightens the power grid safety boundary to reduce the false positive rate of the model, and adjusts the model parameters of the oblique decision tree considering the critical samples (unsafe power grid data near the power grid safety boundary) to avoid misjudging the unsafe operating state as a safe state.
[0113] Example
[0114] A power grid is selected for case analysis.
[0115] The visualization of the weighted oblique decision tree helps to deeply understand how the algorithm predicts the power grid state, and it is also easy to explain to the dispatcher. The root node in the weighted oblique decision tree represents the power grid operation training set, which contains 615 points belonging to the safe state and 185 points belonging to the unsafe state. If the points in the power grid operation training set meet the division conditions, they are assigned to the left child node; if they do not meet the division conditions, they are assigned to the right child node. The weighted oblique decision tree minimizes the information entropy at each node to achieve a linear combination of power grid characteristics. Figure 3 The following figure shows two selected tests. Divide the plane into a safe region in the upper left corner and an unsafe region in the lower right corner. Then use the test Refine again and remove some unsafe domains at the top. Since more grid characteristics are considered in each division, the weighted oblique decision tree can achieve higher accuracy of grid safety boundary identification with a shallower tree depth, and the number of rows of extracted rules is also less.
[0116] The visualization of the weighted oblique decision tree helps us to deeply understand how the algorithm predicts the power grid state and is also easy to explain to the dispatcher. Figure 4 shown.
[0117] In another embodiment of the present invention, Figure 5 As shown, a data-driven power grid safety boundary feature extraction system is provided, including a data sorting module, a safety boundary module and a feature extraction module;
[0118] A data sorting module is used to generate power grid flow data based on flow calculation;
[0119] Safety boundary module, which builds a power grid safety boundary identification model based on power grid flow data;
[0120] The feature extraction module extracts the grid security boundary features of the power system operation status based on the grid security boundary identification model, and uses the grid security boundary features to tighten the grid security rules and obtain the grid security boundary features.
[0121] In another embodiment of the present invention, a terminal device is provided, the terminal device includes a processor and a memory, the memory is used to store a computer program, the computer program includes program instructions, and the processor is used to execute the program instructions stored in the computer storage medium. The processor may be a central processing unit (CPU), or other general-purpose processors, digital signal processors (DSP), application-specific integrated circuits (ASIC), field-programmable gate arrays (FPGA) or other programmable logic devices, discrete gates or transistor logic devices, discrete hardware components, etc., which are the computing core and control core of the terminal, which are suitable for implementing one or more instructions, and are specifically suitable for loading and executing one or more instructions to implement the corresponding method flow or corresponding functions; the processor described in the embodiment of the present invention can be used for the operation of the data-driven power grid security boundary feature extraction method.
[0122] In another embodiment of the present invention, the present invention also provides a storage medium, specifically a computer-readable storage medium (Memory), which is a memory device in a terminal device for storing programs and data. It is understandable that the computer-readable storage medium here can include both the built-in storage medium in the terminal device and the extended storage medium supported by the terminal device. The computer-readable storage medium provides a storage space, which stores the operating system of the terminal. In addition, one or more instructions suitable for being loaded and executed by the processor are also stored in the storage space, and these instructions can be one or more computer programs (including program codes). It should be noted that the computer-readable storage medium here can be a high-speed RAM memory or a non-volatile memory, such as at least one disk memory. The processor can load and execute one or more instructions stored in the computer-readable storage medium to implement the corresponding steps of the data-driven power grid security boundary feature extraction method in the above embodiment.
[0123] Those skilled in the art will appreciate that the embodiments of the present application may be provided as methods, systems, or computer program products. Therefore, the present application may adopt the form of a complete hardware embodiment, a complete software embodiment, or an embodiment in combination with software and hardware. Moreover, the present application may adopt the form of a computer program product implemented in one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) that include computer-usable program code.
[0124] The present application is described with reference to the flowcharts and / or block diagrams of the methods, devices (systems), and computer program products according to the embodiments of the present application. It should be understood that each process and / or box in the flowchart and / or block diagram, as well as the combination of the processes and / or boxes in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to generate a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.
[0125] These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing device to operate 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 A function specified in one or more boxes.
[0126] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process. Figure 1 A process or multiple processes and / or boxes Figure 1 The steps for the functions specified in one or more boxes.
[0127] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention rather than to limit it. Although the present invention has been described in detail with reference to the above embodiments, ordinary technicians in the relevant field should understand that the specific implementation methods of the present invention can still be modified or replaced by equivalents. Any modification or equivalent replacement that does not depart from the spirit and scope of the present invention should be covered within the scope of protection of the claims of the present invention.
Claims
1. A data-driven power grid security boundary feature extraction method, It is characterized in that The following steps are involved: S1, using power flow calculation to generate power grid power flow data; S2, using a weighted oblique decision tree to divide the power grid flow data to obtain the power grid safety boundary of the weighted oblique decision tree node; S3, based on the path of the weighted oblique decision tree, the power grid safety boundary of all nodes in the weighted oblique decision tree is extracted to form a safety rule group, and the unsafe power grid data near the power grid safety boundary is used to tighten the safety rule group to obtain the power grid safety boundary characteristics.
2. According to claim 1, a data-driven power grid security boundary feature extraction method, It is characterized in that The power flow data of the power grid is generated by power flow calculation, which specifically includes the following steps: The simulation is used to calculate the power flow and generate the grid operation status data with safety tags, and the branch head end power and branch end power corresponding to different operation points are obtained; Calculate branch load factor based on branch active transmission capacity, branch head end power and branch end power; Calculate the grid status label based on the branch load rate and the set threshold; The power grid operation state data including the safety tag is generated based on the power grid state tag and the power grid operation state vector.
3. According to the data-driven power grid security boundary feature extraction method of claim 2, It is characterized in that The power grid operation state vector selects unit output and load as features.
4. According to claim 1, a data-driven power grid security boundary feature extraction method, It is characterized in that The method of using a weighted oblique decision tree to divide the power grid flow data and obtain the power grid safety boundary of the weighted oblique decision tree node specifically includes the following steps: The solution corresponding to minimizing the weighted information entropy is used to obtain the parameters of the linear partition in each node of the weighted oblique decision tree; The weighted information entropy is used as the objective function, and the parameters of the linear division in the nodes of the weighted oblique decision tree are used as the coefficients before the features in the power grid security boundary to obtain the power grid security boundary.
5. According to claim 1, a data-driven power grid security boundary feature extraction method, It is characterized in that The path extraction of the weighted oblique decision tree based on the power grid safety boundary of all nodes in the weighted oblique decision tree to form a safety rule group specifically includes the following steps: All nodes from the root node to a leaf node are called a path. With the help of tree traversal, the divisions on the oblique decision tree path are extracted to form an inequality group, thereby extracting a safety rule group representing the relationship between the unit output and the safety of power grid operation.
6. According to claim 5, a data-driven power grid security boundary feature extraction method, It is characterized in that The nodes of the tree include serial number, partition, tree depth, whether it is a leaf node, left child node serial number and right child node serial number; the partition includes feature index, feature coefficient and threshold.
7. According to claim 1, a data-driven power grid security boundary feature extraction method, It is characterized in that The specific steps of tightening the power grid security boundary based on the unsafe power grid data near the power grid security boundary include: a. Calculate the Euclidean distance from the pseudo-positive samples in path p to each hyperplane in the path; b. Obtain the hyperplane corresponding to the minimum Euclidean distance and update the threshold of the hyperplane; c. Determine whether other pseudo positive samples belong to path p based on the hyperplane of the updated threshold, and remove pseudo positive samples that do not belong to path p; d. Calculate the Euclidean distances from the remaining pseudo-positive samples in path p to the hyperplanes in the path, and repeat steps b and c above until the pseudo-positive samples in path p are empty. This completes the tightening of the security rule group, and the tightened security rule group is the obtained power grid security boundary feature.
8. A data-driven power grid security boundary feature extraction system, It is characterized in that Including data sorting module, security boundary module and feature extraction module; Data sorting module, used for power flow calculation to generate power grid power flow data; The safety boundary module uses a weighted oblique decision tree to divide the power grid flow data to obtain the power grid safety boundary of the weighted oblique decision tree node; The feature extraction module extracts the power grid safety boundaries of all nodes in the weighted oblique decision tree based on the path of the weighted oblique decision tree to form a safety rule group, and uses the unsafe power grid data near the power grid safety boundary to tighten the safety rule group to obtain the power grid safety boundary characteristics.
9. A computer device comprising a memory, a processor and a computer program stored in the memory and executable on the processor, It is characterized in that When the processor executes the computer program, the steps of the data-driven power grid safety boundary feature extraction method as described in any one of claims 1 to 7 are implemented.
10. A computer-readable storage medium storing a computer program. It is characterized in that When the computer program is executed by a processor, the steps of the data-driven power grid safety boundary feature extraction method according to any one of claims 1 to 7 are implemented.
Citation Information
Cited By
Method and device for generating high-dimensional security boundary of power system adaptive to online state
CN120016459A
An online state power system high-dimensional security boundary generation method and device
CN120016459B
Power grid safe operation boundary extraction method and device, storage medium, program product and computer equipment
CN120127623A
