A Threat Assessment Method and System for Inferring Power Flow Information of Power Grid Lines

Abstract

The present invention belongs to the field of power system information security and data inference, and specifically provides a threat assessment method and system for power grid line power flow information inference, including the following steps: obtaining node features and edge features from pre-disclosed data and using them as inputs to a pre-constructed power grid line power flow information inference model to obtain power grid line power flow information inference values; calculating the error value between the power grid line power flow information inference values and preset labels, and evaluating the threat degree of power grid line power flow information leakage according to the error value, wherein the power grid line power flow information inference model includes: a feature projection module, a graph representation learning module, a deep feature extraction module, and an inference module; the present invention improves the recognition accuracy of the threat of power grid line power flow information leakage and quantifies the possibility of inferring the key operating state of the power grid using limited public information.

Description

Technical Field

[0001] The present invention belongs to the fields of power system information security and data inference, and particularly relates to a threat assessment method and system for power grid line power flow information inference. Background Art

[0002] With the deepening of the digital transformation, power data has become a key production factor, and its strategic value has become increasingly prominent. To maximize the value of power data, the disclosure of power market information has become an important measure to ensure the transparent operation of the market. However, due to the significant physical coupling and business association relationships among power data, attackers may infer key sensitive information such as power grid line power flow from public data, which poses a potential threat to the security of the power system.

[0003] Currently, there are multiple deficiencies in the research on power system security, such as:

[0004] First, existing research mainly confines the problem of inferring non-public sensitive data from public data to the category of parameter identification, focusing on inferring system sensitive parameters, while the in-depth research on system sensitive operation data is relatively lacking.

[0005] Second, there is a complex non-linear association relationship between public data and power grid line power flow information. It is difficult to accurately infer the power flow state under different combinations of public data only relying on physical models, and the physical mechanisms of some scenarios have not been fully revealed. Moreover, relying solely on machine learning methods cannot effectively integrate power grid physical constraints, and the generalization ability is limited.

[0006] Therefore, the modern power system needs a method to evaluate the possibility of inferring power grid line power flow information from public data, providing a reference for improving the power market information disclosure mechanism. Summary of the Invention

[0007] The purpose of the present invention is to provide a threat assessment method and system for power grid line power flow information inference to solve the problem of difficult threat assessment of inferential leakage of current power grid line power flow information.

[0008] To achieve the above purpose, the technical solution adopted by the present invention is:

[0009] In the first aspect, a threat assessment method for power grid line power flow information inference provided by the present invention includes the following steps:

[0010] Obtain node features, edge features, and power grid topology features from the pre-disclosed data of the power market, and input the node features, edge features, and power grid topology features into a pre-constructed power grid line power flow information inference model to obtain the power grid line power flow information inference value;

[0011] Calculate the error value between the inferred value of the power grid line power flow information and the preset label, and evaluate the threat degree of the power grid line power flow information leakage according to the error value. Among them, the pre-constructed power grid line power flow information inference model includes:

[0012] A feature projection module for performing linear transformation on the input node features and edge features;

[0013] A graph representation learning module for updating the output of the feature projection module in combination with the power grid topology features;

[0014] A deep feature extraction module for successively splicing, feature transformation, and merging the output of the graph representation learning module;

[0015] An inference module for performing dimensional mapping on the output of the deep feature extraction module and outputting the inferred value of the power grid line power flow information corresponding to the pre-disclosed data.

[0016] Preferably, the loss function of the pre-constructed power grid line power flow information inference model is constructed according to the prior knowledge combined with the penalty term of the power grid physical property constraints.

[0017] Preferably, the loss function of the pre-constructed power grid line power flow information inference model is:

[0018]

[0019] Among them, are all non-negative weight coefficients; is the mean square error between the inferred value and the actual value; The penalty term for violating the line capacity constraint; The penalty term for violating the shadow price constraint.

[0020] Preferably, the expression of the graph representation learning module:

[0021]

[0022]

[0023]

[0024]

[0025]

[0026]

[0027]

[0028] Among them, Denote the concatenation operation of features; Are multi-layer perceptrons of the activation function PReLU that match the pre-activation values, node features, and edge feature dimensions respectively; And Are two linear layers with the activation function Sigmoid and a single scalar output respectively; Is the BatchNorm layer; Is the Layer node 's feature; Is the Layer node 's feature; Is the Layer edge 's feature; Is the set of nodes connected to the node ; Are all pre-activation values of features; Is the Layer node 's feature; The Layer edge 's feature; Is the set of nodes; Is the set of edges; Is the Layer node's feature; Is the Layer edge's feature.

[0029] Preferably, the concatenation in the deep feature extraction module is performed by the following formula:

[0030]

[0031]

[0032] Where Is the feature after concatenating the node feature and edge feature output by the graph representation learning module; Is the Node's feature output by the graph representation learning module; Is the Node's feature output by the graph representation learning module; Is the Edge's feature output by the graph representation learning module; Denote the concatenation operation of features; Is the set of edges; Is the feature after concatenating the Node's feature, Node's feature, and Edge's feature output by the graph representation learning module.

[0033] Preferably, after evaluating the threat degree of power grid line power flow information leakage according to the error value, it further includes identifying the data that has a key impact on the inference of power grid line power flow information in the public data of the power market by using the error value between the inferred value of the power grid line power flow information and the preset label.

[0034] In a second aspect, the present invention provides a threat assessment system for power grid line power flow information inference, including:

[0035] A power flow inference value acquisition unit, configured to obtain node features, edge features, and power grid topology features from the pre-disclosed data of the power market, input the node features, edge features, and power grid topology features into a pre-constructed power grid line power flow information inference model, and obtain an inferred value of the power grid line power flow information;

[0036] A threat degree assessment unit, configured to calculate the error value between the inferred value of the power grid line power flow information and the preset label, and evaluate the threat degree of power grid line power flow information leakage according to the error value, wherein the pre-constructed power grid line power flow information inference model includes:

[0037] A feature projection module, configured to perform a linear transformation on the input node features and edge features;

[0038] A graph representation learning module, configured to update the output of the feature projection module in combination with the power grid topology features;

[0039] A deep feature extraction module, configured to sequentially perform splicing, feature transformation, and merging on the output of the graph representation learning module;

[0040] An inference module, configured to perform a dimension mapping on the output of the deep feature extraction module and output the inferred value of the power grid line power flow information corresponding to the pre-disclosed data.

[0041] In a third aspect, the present invention provides a computer device, including:

[0042] A processor, suitable for executing a computer program;

[0043] A computer-readable storage medium, in which a computer program is stored, and when the computer program is executed by the processor, the method described above is executed.

[0044] In a fourth aspect, the present invention provides a computer-readable storage medium, in which a computer program is stored, and when the computer program is executed by a processor, the method described above is implemented.

[0045] In a fifth aspect, the present invention provides a computer program product, characterized in that the computer program product includes a computer program, and when the computer program is executed by a processor, the method described above is implemented.

[0046] Compared with the prior art, the beneficial effects of the present invention are as follows:

[0047] A threat assessment method for inferring power grid line power flow information provided by the present invention pre-constructs a power grid line power flow information inference model including a feature projection module, a graph representation learning module, a deep feature extraction module, and an inference module. It obtains node features and edge features from pre-disclosed data, and uses the feature projection module to perform a linear transformation on the input node features and edge features to obtain a feature representation with a unified dimension. Then, the graph representation learning module is used to effectively learn and update the node features and edge features with the unified dimension simultaneously, while traditional networks only focus on the update of node features, resulting in insufficient utilization of edge features. Then, the deep feature extraction module is used to integrate the output features of the graph representation learning module according to the system topology structure, and then a residual connection mechanism is adopted for deep feature learning to effectively alleviate the problem of gradient disappearance. Finally, the inference module is used to map the extracted high-dimensional features to the target output dimension. Through this power grid line power flow information inference model, node features (such as power generation, load demand, nodal marginal price) and edge features (such as capacity, shadow price) are integrated, revealing the implicit association between the disclosed data and the power grid line power flow information, and quantifying the possibility of inferring the key operating state of the power grid using limited disclosed information.

[0048] Furthermore, the present invention proposes to fuse a penalty term of power grid physical property constraints in the loss function of the power grid line power flow information inference model to improve the power flow inference accuracy under data sparse conditions, ensure the physical feasibility and operation consistency of the inference results. Compared with traditional physical models and pure machine learning methods, this method has stronger generalization ability and robustness.

[0049] Furthermore, the present invention can also identify and quantify the low-sensitivity information in the disclosed data that has a key impact on the power grid line power flow inference. Through sensitivity analysis, it reveals the contribution degree of different types of data to the inference accuracy, clarifies the sensitive data categories that need to be protected with emphasis, and helps to optimize the power market information disclosure strategy. BRIEF DESCRIPTION OF THE DRAWINGS

[0050] Figure 1 It is a flowchart of an embodiment of the present invention;

[0051] Figure 2 It is a schematic diagram of the specific process of an embodiment of the present invention;

[0052] Figure 3 It is an architecture diagram of the power grid line power flow neural network of an embodiment of the present invention;

[0053] Figure 4 It is a topology diagram of the IEEE 14-node test system in an embodiment of the present invention. DETAILED DESCRIPTION OF THE INVENTION

[0054] In the following description, specific details such as specific system architectures, technologies, etc. are presented for the purpose of illustration rather than limitation, so as to provide a thorough understanding of the embodiments of the present application. However, those skilled in the art should clearly understand that the present application can also be implemented in other embodiments without these specific details. In other cases, detailed descriptions of well-known systems, devices, circuits, and methods are omitted to avoid unnecessary details from interfering with the description of the present application.

[0055] It should be understood that when used in the specification and appended claims of the present application, the term "comprising" indicates the presence of the described features, wholes, steps, operations, elements, and / or components, but does not exclude the presence or addition of one or more other features, wholes, steps, operations, elements, components, and / or their combinations.

[0056] It should also be understood that the term "and / or" used in the specification and appended claims of the present application refers to any combination and all possible combinations of one or more of the associated listed items, and includes these combinations.

[0057] As used in the specification and appended claims of the present application, the term "if" can be interpreted as "when", "once", "in response to determining", or "in response to detecting" depending on the context. Similarly, the phrase "if determined" or "if [the described condition or event] is detected" can be interpreted as meaning "once determined", "in response to determining", "once [the described condition or event] is detected", or "in response to detecting [the described condition or event]" depending on the context.

[0058] In addition, in the description of the specification and appended claims of the present application, the terms "first", "second", "third", etc. are only used for distinguishing descriptions and cannot be understood as indicating or implying relative importance.

[0059] The reference to "one embodiment" or "some embodiments" etc. described in the specification of the present application means that a specific feature, structure, or characteristic described in connection with that embodiment is included in one or more embodiments of the present application. Thus, statements such as "in one embodiment", "in some embodiments", "in other some embodiments", "in still other embodiments", etc. that appear in different places in this specification do not necessarily all refer to the same embodiment, but mean "one or more but not all embodiments", unless otherwise specifically emphasized in other ways. The terms "comprising", "including", "having", and their variants all mean "including but not limited to", unless otherwise specifically emphasized in other ways.

[0060] Embodiment 1

[0061] AsFigure 1 As shown in the figure, a threat assessment method for inferring power grid line flow information provided in this embodiment includes the following steps:

[0062] Step 1: Obtain the historical public data corresponding to the power market and the unpublicized power grid line flow information. Among them, obtain node features, edge features, and power grid topology features from the historical public data, and construct a training set and a validation set with the obtained node features, edge features, and power grid topology features; and use the unpublicized power grid line flow information as a preset label ;

[0063] Step 2: Construct a neural network architecture for inferring power grid line flow, and embed a penalty term constrained by the physical characteristics of the power grid into the loss function of the neural network architecture for inferring power grid line flow in combination with prior knowledge to generate a threat assessment model for inferring power grid line flow information;

[0064] Step 3: Optimize the model parameters using the training set, and supervise the training process through the validation set to obtain a trained threat assessment model for inferring power grid line flow information;

[0065] Step 4: Obtain node features and edge features from the pre-public data, input them into the threat assessment model for inferring power grid line flow information to obtain an inferred value of the power grid line flow information, calculate the error value between the inferred value of the power grid line flow information and the preset label, and evaluate the threat degree of power grid line flow information leakage according to the error value.

[0066] Embodiment 2

[0067] As Figure 2 shown in the figure, a threat assessment method for inferring power grid line flow information provided in this embodiment includes the following steps:

[0068] Step 1: Obtain the historical public data corresponding to the power market and the unpublicized power grid line flow information. Among them, obtain node features, edge features, and power grid topology features from the historical public data, and construct a training set and a validation set with the obtained node features, edge features, and power grid topology features; use the unpublicized power grid line flow information as a preset label ; Specifically:

[0069] S101: According to the threat assessment objective of inferring power grid line flow, obtain the historical public data corresponding to the power market and the unpublicized power grid line flow information. Among them, the historical public data includes power grid construction parameters, market transaction data, and system operation data, where:

[0070] The power grid construction parameters include: topological structure, node distribution (power generation / substation / load), voltage level, line capacity, etc.;

[0071] Market transaction data includes: nodal marginal price, shadow price, marginal energy price, marginal loss price, etc.;

[0072] System operation data includes: unit status, actual load, system reserve information, transmission status of important channels, transmission section constraint conditions, inter-provincial tie-line power flow, key equipment power flow data, etc.

[0073] S102: Structurally process the obtained historical public data to obtain node features, edge features, and power grid topology features, where:

[0074] Set the topological structure diagram of the power grid as , where, represents the set of nodes, represents the set of edges; the number of nodes in the power grid is , and the number of edges is ;

[0075] Extract the data corresponding to each node in the topological structure diagram from the historical public data to construct node features ;

[0076] Extract the data corresponding to each edge in the topological structure diagram from the historical public data to construct edge features ;

[0077] Construct the power grid topology features based on the connection relationship between nodes in the topological structure diagram;

[0078] S103: Dataset division and storage

[0079] Construct a training set and a validation set from the obtained node features, edge features, and power grid topology features; save the processed dataset in an efficient storage format for easy training and validation of the model.

[0080] Step 2, construct a power grid line power flow inference neural network architecture, embed the penalty term of the power grid physical characteristics constraint into the loss function in combination with prior knowledge, and generate a power grid line power flow information inference model. The specific method is:

[0081] S201: Construct a power grid line power flow inference neural network architecture. The input of this power grid line power flow inference neural network architecture is node features , edge features and power grid topology features, and the output is the power grid line power flow information inference value . The power grid line power flow inference neural network architecture includes:

[0082] Feature projection module: Perform a linear transformation on the input node features and edge features through a fully connected layer, and output hidden features;

[0083] Graph representation learning module: The hidden features output by the feature projection module and the power grid topology features are updated using a two - level cascaded XENet network. Among them, each layer of the XENet network updates the input hidden features using the following formula:

[0084]

[0085]

[0086]

[0087]

[0088]

[0089]

[0090]

[0091] Among them, represents the concatenation operation of features; are multi - layer perceptrons of the activation function PReLU that match the pre - activation value, node features, and edge feature dimensions respectively; and are two linear layers with activation functions Sigmoid and a single scalar output respectively; is the BatchNorm layer; is the th layer node 's feature; is the th layer node 's feature; is the th layer edge 's feature; is the node connected to node ; are all pre - activation values of features; is the th layer node 's feature; The th layer edge 's feature; is the th layer node's feature; is the th layer edge's feature.

[0092] Deep feature extraction module: It is used to splice the updated node features and edge features output by the graph representation learning module through the following formula; then, the spliced features are subjected to feature transformation through two cascaded fully connected layers. Each fully connected layer is followed by a ReLU activation function, and a residual connection is introduced to add the input features to the transformed features, and the merged features are output.

[0093]

[0094]

[0095] Among them, is the feature after splicing the node features and edge features output by the graph representation learning module; is the node feature output by the graph representation learning module; is the node feature output by the graph representation learning module; is the edge feature output by the graph representation learning module; represents the splicing operation of features; is the set of edges; is the node feature, node feature and edge feature after splicing the features output by the graph representation learning module.

[0096] Inference module: Use a single fully connected layer to map the dimension of the merged features output by the deep feature extraction module to the target output dimension.

[0097] Step S202, construct the loss function of the power grid line power flow inference neural network architecture. This loss function is obtained by embedding the penalty term constrained by the physical characteristics of the power grid based on prior knowledge, and the power grid line power flow information inference model is obtained.

[0098] In addition to the traditional data fitting term, that is, the mean squared error (MSE) between the inferred value and the actual value, the following two penalty terms based on the physical characteristics constraints of the power grid are introduced to describe the mechanism of inferring the line power flow from the publicly available data:

[0099]

[0100]

[0101]

[0102] Penalty term for violating the line capacity constraint : According to the security operation constraints of the power market model, the power flow of the power grid line should not exceed its capacity limit. Therefore, let the maximum capacity of the power grid line power flow be , there is:

[0103]

[0104] Penalty term for violating the shadow price constraint : Used to characterize the relationship between the power flow of the power grid line, the line capacity and the shadow price. Let the shadow price of the line be , there is:

[0105]

[0106] In summary, the overall loss function can be expressed as:

[0107]

[0108] Among them, are all non - negative weight coefficients, used to balance the magnitudes of each loss component and regulate their relative contributions to model training, and adjust the weights so that each loss component is on the same order of magnitude.

[0109] Optionally, other known physical constraints related to the power flow of the power grid line in historical public data can also be analyzed, and penalty terms are constructed and added to the loss function.

[0110] Step 3, use the training set to optimize the model parameters of the power grid line power flow information inference model, and supervise the training process through the validation set to obtain the trained power grid line power flow information inference model;

[0111] S301, determine the set of model parameters to be optimized in the power grid line power flow inference model;

[0112] S302, in each training iteration, obtain a batch of training data through random sampling, and input it into the power grid line power flow information inference model to get the corresponding estimated values;

[0113] S303, optimize and update the model parameters based on the loss function until the loss value converges to the error threshold Hereinafter, the model training uses the training set data, while the hyperparameter optimization of the model is adjusted through the validation set data.

[0114] Step 4, process the pre - public data obtained from the power market according to Step 1 to obtain the node features, edge features and power grid topology features corresponding to the pre - public data; use the node features, edge features and power grid topology features as the input of the power grid line power flow information inference model to obtain the power grid line power flow information inference value, and calculate the power grid line power flow information inference value and the preset label The error value between them is used to evaluate the threat level of power grid line power flow information leakage in the pre-disclosed data according to the error value.

[0115] Embodiment 3

[0116] Based on Embodiment 1 or 2, a method for inferring threat assessment of power grid line power flow information provided in this embodiment further includes using the error value between the inferred value of power grid line power flow information and a preset label to identify the data that has a key impact on the inference of power grid line power flow information in the public data of the power market. Specifically:

[0117] Adjust the input data, including selectively adding or removing specific categories of data, retrain the model based on the adjusted input data, and perform power grid line power flow inference;

[0118] Calculate the error value between the inferred value of power grid line power flow information obtained using the adjusted input data and the preset label between;

[0119] By comparing the change in the inference error before and after the adjustment of the input data, identify the data that has a key impact on the inference of power grid line power flow information.

[0120] Embodiment 4

[0121] This embodiment conducts simulation verification based on the IEEE 14-node test system, and its topology diagram is as Figure 4 shown.

[0122] According to the current power grid data disclosure rules, a benchmark data combination is set as shown in Table 1. Among them, in this example, the node features are composed of the node type, voltage level, and the blocked component of the node marginal electricity price, and the edge features include important line power flow data, transformer power flow data, line capacity, and shadow price.

[0123] Table 1 Description of Benchmark Known Data

[0124]

[0125] To simulate the load fluctuation over time, the load data announced in the 2012 Global Energy Forecasting Competition is used. Its real-time load is normalized to obtain the load fluctuation trend, and the fluctuation is multiplied by the active power value of the system nodes to obtain the real-time active power data of the nodes. With a sampling interval of one hour, a total of 10,000 samples are generated and divided into a training set, a validation set, and a test set according to a ratio of 8:1:1.

[0126] To evaluate the inference effect, this embodiment uses the following two indicators:

[0127] (a) Weighted Mean Absolute Percentage Error (wMAPE): It is used to evaluate the relative error. In the actual operation of the power grid, due to the dynamic changes in load levels and power generation outputs, there will be cases where the power flow values of individual power grid lines are very small at certain moments. In view of this characteristic, this embodiment selects wMAPE instead of the traditional MAPE as the relative error evaluation index because when the actual power flow value approaches zero, MAPE may produce a relative error that tends to infinity, affecting the reliability of the overall evaluation index; the calculation formula of wMAPE is:

[0128]

[0129] where and are the actual value and the inferred value of the power flow of the power grid line respectively.

[0130] (b) Mean Absolute Error (MAE): It is used to evaluate the absolute error between the inferred value and the actual value, and the calculation formula is:

[0131]

[0132] The set baseline method is the Graph Attention Network (GAT) to verify the effectiveness of the proposed method.

[0133] The first experiment: Inferred threat assessment

[0134] Train the model on the IEEE 14-node test system under the condition of the benchmark data combination to infer the power flow values of all network lines. The experimental results on the test set show that this embodiment achieves a wMAPE of 5.16% and an MAE of 0.2694, and shows extremely high inference accuracy at multiple moments, which reflects that there is a significant leakage risk caused by inferring from public data in the power flow data of the line. In contrast, if the baseline method is selected, its wMAPE is 13.51% and its MAE is 0.7056, which indicates that this application has obvious performance advantages over the baseline method in the line power flow inference task.

[0135] The second experiment: Identification of important low-sensitivity / public data during the inference process

[0136] To evaluate the threats of different types of low-sensitivity level data to the line power flow inference and identify the key low-sensitivity data during the inference process, this embodiment designs five different data combinations for comparative experiments, as shown in Table 2.

[0137] Table 2 Known Data Description

[0138]

[0139] Table 3 Line Power Flow Inference Errors under Different Combinations of Known Data in the 14 - Node System

[0140]

[0141] For the IEEE 14 - node test system, the known data was adjusted and the model was trained multiple times, and the inference accuracy of the line power flow was obtained as shown in Table 3. The analysis is as follows:

[0142] The experimental results show that this method is always superior to the baseline method in terms of inference accuracy. Specifically, this method obtains lower wMAPE values in all data combinations, ranging from 5.16% to 11.67%, while the wMAPE values of the baseline method are significantly higher, ranging from 13.51% to 65.83%. This performance advantage stems from the better learning ability of this method for edge features.

[0143] Analysis finds that introducing various types of low - sensitive data will increase the leakage risk of high - sensitive data. By comparing the results of combination 1 with combinations 2 to 4, it can be seen that different types of low - sensitive data can improve the inference accuracy of the line power flow to varying degrees. Specifically, when only partial line power flows are known in combination 1, the wMAPE inferred by this method is 11.67% and the MAE is 0.6097. When additional node types and voltage levels are known in combination 2 compared to combination 1, the wMAPE of this method drops to 6.97% and the MAE drops to 0.3639. This indicates that grid construction parameters play a significant role in the task of grid line power flow inference, which is highly consistent with the physical property that the line power flow flows from the power generation nodes to the load nodes, suggesting that system operators should avoid disclosing too many grid construction parameters such as the distribution of power generation / load nodes and voltage levels. The result of combination 5 further confirms this, where the wMAPE of this method further drops to 5.16% and the MAE drops to 0.2695, indicating that the collaborative use of multiple low - sensitive data can produce a cumulative effect, resulting in a significant increase in the leakage threat of high - sensitive data.

[0144] In summary, this embodiment verifies the superiority of this method in the threat assessment of grid power flow information inference.

[0145] Embodiment 5

[0146] A threat assessment system for grid line power flow information provided in this embodiment includes:

[0147] The grid line power flow information inference value acquisition unit is used to obtain node features, edge features, and grid topology features from the pre-disclosed data of the power market, and input the node features, edge features, and grid topology features into a pre-constructed grid line power flow information inference model to obtain the grid line power flow information inference value;

[0148] The threat level assessment unit is used to calculate the error between the grid line power flow information inference value and a preset label, and evaluate the threat level of grid line power flow information leakage according to the error, where:

[0149] The pre-constructed grid line power flow information inference model includes:

[0150] The feature projection module is used to perform a linear transformation on the input node features and edge features;

[0151] The graph representation learning module is used to update the output of the feature projection module in combination with the grid topology features;

[0152] The deep feature extraction module is used to sequentially splice, perform feature transformation, and merge the output of the graph representation learning module;

[0153] The inference module is used to perform dimensionality mapping on the output of the deep feature extraction module and output the grid line power flow information inference value corresponding to the pre-disclosed data.

[0154] Embodiment 6

[0155] This embodiment provides a computer device, including: a memory for storing a computer program; a processor for implementing the steps of a computer method when executing the computer program.

[0156] When the processor executes the computer program, it implements the steps of the above computer method.

[0157] Alternatively, when the processor executes the computer program, it implements the functions of each module in the above system,

[0158] The computer device may be a desktop computer, a notebook, a palm computer, a cloud server, or other computing devices. The computer device may include, but is not limited to, a processor and a memory. Those skilled in the art can understand that the above are examples of computer devices and do not constitute a limitation on the computer device. It may include more components than the above, or combine some components, or different components. For example, the computer device may further include input / output devices, network access devices, a bus, etc.

[0159] The so-called processor may be a central processing unit (CPU), or may also be other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. The general-purpose processor may be a microprocessor, or the processor may also be any conventional processor, etc. The processor is the control center of the computer device, and connects various parts of the entire computer device through various interfaces and circuits.

[0160] The memory can be used to store the computer program and / or module. By running or executing the computer program and / or module stored in the memory, and calling the data stored in the memory, the processor realizes various functions of the computer device.

[0161] The memory may mainly include a program storage area and a data storage area. Among them, the program storage area can store an operating system, application programs required for at least one function (such as a sound playback function, an image playback function, etc.); the data storage area can store data created according to the use of the mobile phone (such as audio data, phone book, etc.). In addition, the memory may include high-speed random access memory, and may also include non-volatile memory, such as a hard disk, a memory, a plug-in hard disk, a smart media card (SMC), a secure digital (SD) card, a flash card, at least one magnetic disk storage device, a flash memory device, or other volatile solid-state storage devices.

[0162] Embodiment 7

[0163] This embodiment also provides a computer-readable storage medium. The computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the steps of the method are realized.

[0164] If the modules / units integrated in the computer system are implemented in the form of software function units and sold or used as independent products, they can be stored in a computer-readable storage medium.

[0165] Based on such understanding, all or part of the processes in the above method of the present invention can also be completed by instructing relevant hardware through a computer program. The computer program can be stored in a computer-readable storage medium. When the computer program is executed by a processor, the steps of the above computer method can be implemented. Among them, the computer program includes computer program code, and the computer program code can be in the form of source code, object code, executable file, or preset intermediate form, etc.

[0166] The computer-readable storage medium can include: any entity or device capable of carrying the computer program code, recording medium, USB flash drive, mobile hard disk, magnetic disk, optical disc, computer memory, read-only memory (ROM), random access memory (RAM), electrical carrier signal, telecommunication signal, and software distribution medium, etc.

[0167] It should be noted that the content included in the computer-readable storage medium can be appropriately increased or decreased according to the requirements of legislation and patent practice within the jurisdiction. For example, in some jurisdictions, according to legislation and patent practice, the computer-readable storage medium does not include electrical carrier signals and telecommunication signals.

[0168] Embodiment 8

[0169] This embodiment provides a computer product. The computer program product includes a computer program, and the computer program is stored in a computer-readable storage medium; the processor of the computer device reads the computer program from the computer-readable storage medium, and the processor executes the computer program, so that the computer device can execute the method in Embodiment 1, which will not be elaborated here.

[0170] It should be noted that those of ordinary skill in the art can understand that all or part of the processes in the above method of the embodiment can be completed by instructing relevant hardware through a computer program. The program can be stored in a computer-readable storage medium. When the program is executed, it can include the processes of the above embodiments of each method.

[0171] The above embodiments are only used to illustrate the technical solutions of the present application, rather than to limit it; although the present application has been described in detail with reference to the foregoing embodiments, those of ordinary skill in the art should understand that they can still modify the technical solutions recorded in the foregoing embodiments, or perform equivalent replacements for some of the technical features; and these modifications or replacements do not make the essence of the corresponding technical solutions deviate from the spirit and scope of the technical solutions of the embodiments of the present application, and should all be included in the protection scope of the present application.

Claims

1. A method for assessing power grid line flow information inference threat, characterized in that: The following steps are involved: Obtain node features, edge features, and grid topology features from pre-public data of the power market, input the node features, edge features, and grid topology features into a pre-built grid line power flow information inference model, and obtain grid line power flow information inference values; The error value between the power line flow information inference value and the preset label is calculated, and the threat level of power line flow information leakage is evaluated according to the error value. The pre-built power line flow information inference model includes: Feature projection module, used to perform linear transformation on input node features and edge features; The graph representation learning module is used to update the output of the feature projection module combined with the grid topology features using a two-layer cascaded XENet network; A deep feature extraction module is used to sequentially concatenate, transform and merge the outputs of the graph representation learning module; An inference module is used to perform dimension mapping on the output of the deep feature extraction module and output an inferred value of the power grid line flow information corresponding to the pre-public data; The graph represents the expression of the learning module: in, Represents the concatenation operation of features; They are multi-layer perceptrons with activation functions PReLU that match the pre-activation value, node feature, and edge feature dimensions respectively; and They are two linear layers with activation function Sigmoid and a single scalar output; It is the BatchNorm layer; For the Layer Node Features; For the Layer Node Features; For the Layer Edge Features; For the node The set of connected nodes; are the pre-activation values ​​of the features; For the Layer Node Features; No. Layer Edge Features; is a collection of nodes; is the set of edges; For the Characteristics of layer nodes; For the Features of layer edges.

2. A method for assessing power grid line flow information inference threats according to claim 1, characterized in that: The loss function of the pre-constructed power grid line flow information inference model is constructed based on prior knowledge combined with penalty terms constrained by the physical characteristics of the power grid.

3. A method for assessing power grid line flow information inference threats according to claim 1 or 2, characterized in that: The loss function of the pre-built power line flow information inference model for: in, All are non-negative weight coefficients; is the mean square error between the inferred value and the actual value; Penalty term for violating line capacity constraints; The penalty term for violating the shadow price constraint.

4. A method for assessing power grid line flow information inference threat according to claim 1, characterized in that: The deep feature extraction module is spliced ​​by the following formula: in, It is the concatenation of node features and edge features output by the graph representation learning module; Nodes output by the graph representation learning module Features; Nodes output by the graph representation learning module Features; is the edge output by the graph representation learning module Features; Represents the concatenation operation of features; is the set of edges; Nodes output by the graph representation learning module Features and nodes Features and edges The features after splicing.

5. A method for assessing power grid line flow information inference threats according to claim 1, characterized in that: After evaluating the threat level of power grid line flow information leakage according to the error value, the method further includes: The error value between the inferred value of power grid line flow information and the preset label is used to identify the data in the pre-public data of the power market that has a key impact on the inference of power grid line flow information.

6. A power grid line flow information inference threat assessment system, characterized in that: Based on the evaluation method according to any one of claims 1 to 5, the system comprises: A power flow inference value acquisition unit is used to acquire node characteristics, edge characteristics and power grid topology characteristics from pre-public data of the power market, input the node characteristics, edge characteristics and power grid topology characteristics into a pre-built power grid line power flow information inference model, and obtain the power grid line power flow information inference value; The threat level assessment unit is used to calculate the error value between the power line flow information inference value and the preset label, and assess the power line flow information leakage threat level according to the error value, wherein the pre-built power line flow information inference model includes: Feature projection module, used to perform linear transformation on input node features and edge features; The graph representation learning module is used to update the output of the feature projection module combined with the grid topology features using a two-layer cascaded XENet network; A deep feature extraction module is used to sequentially concatenate, transform and merge the outputs of the graph representation learning module; The inference module is used to perform dimension mapping on the output of the deep feature extraction module and output the inferred value of the power grid line flow information corresponding to the pre-public data.

7. A computer device, characterized in that: include: a processor suitable for executing a computer program; A computer-readable storage medium, wherein a computer program is stored in the computer-readable storage medium, and when the computer program is executed by the processor, the method according to any one of claims 1 to 5 is performed.

8. A computer-readable storage medium storing a computer program, characterized in that: When the computer program is executed by a processor, the method according to any one of claims 1 to 5 is implemented.

9. A computer program product, characterized in that The computer program product comprises a computer program, and when the computer program is executed by a processor, the method according to any one of claims 1 to 5 is implemented.

Images