Power grid congestion risk assessment method and system considering node connectivity loss

By using node connectivity loss indicators and neural network models in the power grid, the problem that traditional grid safety warning technology cannot evaluate the grid operation risks in real time is solved, real-time online quantitative assessment and early warning of grid blockage risks is realized, and the safety of grid operation is improved Sex and accuracy.

CN114997566BActive Publication Date: 2024-12-03STATE GRID LIAONING ELECTRIC POWER CO LTD +2
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202210406215.8
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2022-04-18
Publication Date
2024-12-03
Estimated Expiration
2042-04-18

AI Technical Summary

Technical Problem

Traditional model-driven grid safety warning technology cannot evaluate the operating risks of power grids in real time, cannot adapt to the complex and changeable grid operation mode, and lacks clear expression of grid safety characteristics and real-time early warning capabilities.

Method used

The grid blocking risk assessment method based on actual grid simulation and operation data is adopted. By considering node connectivity loss indicators, combining feature extraction and neural network model, a blocking risk assessment sample set is generated, and dimensionality reduction processing and risk assessment are carried out to achieve grid blocking risk. Real-time online quantitative evaluation.

Benefits of technology

There is no need for accurate grid mathematical models, and offline and online learning can be performed simultaneously, which improves the accuracy of grid blocking risk assessment and early warning accuracy, significantly improves the safety of grid operation, and provides decision-making support.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN114997566B_ABST
    Figure CN114997566B_ABST
Patent Text Reader

Abstract

The present invention discloses a method and system for assessing the risk of grid congestion considering the loss of node connectivity. The method utilizes the grid structure of the target grid and the given operation mode data, combines the simulation data generated by the power system analysis and calculation model and the historical data of the actual operation of the grid, calibrates the data according to the data preprocessing technology and the index considering the loss of node connectivity, and then generates a multi-source congestion risk assessment sample set; then utilizes the feature extraction technology to perform data dimension reduction processing; and then combines the machine learning algorithm to form a grid congestion risk assessment method. The method of the present invention does not require an accurate mathematical model of the grid, and only mines based on simulation data and historical operation data, continuously improves the accuracy of machine learning grid congestion risk assessment, provides decision support for the prevention of static operation risks and dispatching control of the grid, and significantly improves the safety of the actual operation of the grid.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention belongs to the field of steady-state analysis of power systems, and in particular relates to a method and system for evaluating power grid congestion risk taking into account node connectivity loss. Background Art

[0002] The traditional field of power grid security early warning mainly focuses on some offline methods based on model drive. For example, the earlier concept of security early warning is based on a power grid security early warning technology of the "triage-consultation" mode. In the "triage", the hidden dangers of power grid security are found in multiple dimensions, the security level of power grid is determined, and a comprehensive early warning is made in the "consultation", which achieves a high degree of automation. There are also technologies that design power grid security early warning and decision support systems from three levels: time dimension, space dimension and object dimension. There are technologies that design a blackout defense system based on the three lines of defense of the power system, and the power grid security early warning is the basis of the system. The three lines of defense focus on protection and control after the fault, and cannot be warned in advance. There are also technologies that comprehensively consider static safety issues, transient safety issues, voltage safety issues and relay protection setting verification based on the energy management system data platform to build a power grid security early warning system. There are also technologies that build an online dynamic safety assessment and early warning system for the power system, which can realize multiple types of online safety and stability analysis such as static stability, transient stability, voltage stability, and small interference stability. Some scholars have developed a large-scale power grid safety analysis, early warning and control system that can preliminarily rehearse, analyze, warn and prevent various types of power grid failures and accidents.

[0003] With the continuous development of smart grids, on the one hand, the scale of power grids is getting larger and larger, and the high proportion of new energy and AC / DC hybrid connection have led to increasingly complex operation modes of power grids and close to the boundary of stable operation; on the other hand, the accumulation of a large number of measurement methods and multi-temporal and spatial time scale data has also brought new challenges to the operation analysis and evaluation of power grids. The traditional "model-driven" power grid risk assessment technology has the following problems: ① Critical lines or transmission sections are important means for dispatching and operating personnel to "reduce the dimension of the power grid". In the traditional "model-driven" power grid safety warning, the transmission section is formulated offline and cannot be updated online, which makes it difficult to adapt to the current complex and changeable operation mode of the power grid. ② The section limit transmission capacity is an important basis for dispatchers to monitor the section, and the section flow needs to be controlled below the section limit transmission capacity. However, in the traditional "model-driven" power grid safety warning, the operation rules are relatively extensive, lacking the selection of power grid safety characteristics, and unable to clearly express the key factors affecting power grid safety, and it is even more difficult to pre-warn and pre-control the power grid according to these key factors. ③ Real-time is the basic requirement for safety warning. Due to the limitation of computing power, it is difficult for "model-driven" power grid safety warning to achieve real-time warning, let alone real-time or advanced warning.

[0004] Therefore, there is an urgent need to carry out research on data generation and preprocessing technologies, congestion risk warning and other issues related to data-driven grid congestion risk assessment in the cyber-physical environment of smart grids. Summary of the invention

[0005] The purpose of the present invention is to provide a method and system for evaluating the risk of power grid congestion taking into account the loss of node connectivity, so as to overcome the defect that the prior art cannot perform real-time online quantitative evaluation of the risk of real-time power grid operation. The present invention provides a certain theoretical basis and quantitative analysis technical support for power grid risk assessment and dispatching operation based on actual power grid simulation and operation data, thereby quantifying the degree of future power grid congestion risk.

[0006] In order to achieve the above object, the present invention adopts the following technical scheme:

[0007] The grid congestion risk assessment method considering node connectivity loss includes:

[0008] The data to be processed are formed by using the grid structure and operation mode data of the target power grid, combining the simulation data generated by the power system analysis and calculation model and the historical data of the actual operation of the power grid;

[0009] Preprocess the data to be processed;

[0010] An index that considers node connectivity loss is added to the preprocessed data as a data attribute. The grid congestion risk index is established using the data that considers node connectivity loss to calibrate the preprocessed data and generate a congestion risk assessment sample set.

[0011] The feature extraction technology is used to reduce the dimension of the blocking risk assessment sample set;

[0012] For the congestion risk assessment sample set after dimensionality reduction processing, a power grid congestion risk assessment neural network model is built, and the power grid congestion risk assessment neural network model is trained. The trained power grid congestion risk assessment neural network model is used to evaluate the power grid congestion risk.

[0013] Furthermore, the grid structure of the target power grid refers to various normal, maintenance, newly commissioned lines and main transformers, as well as all grid structures under the operation and scheduling conditions considering topology control;

[0014] The operation mode data of the target power grid include various power flow operation modes under conventional generator sets, new energy generator sets, loads and DC transmission power parameters.

[0015] Furthermore, the simulation data generated by the power system analysis and calculation model refers to steady-state calculation simulation data generated by various analysis software capable of performing steady-state power flow calculations on power systems;

[0016] The historical data of actual operation of the power grid refers to the historical steady-state power flow data recorded from actual operation under various grid structures and various operation modes.

[0017] Furthermore, the specific process of preprocessing the data to be processed is: performing data denoising, abnormal data cleaning, missing data completion and data normalization on the data to be processed in sequence.

[0018] Furthermore, the index considering the node connectivity loss refers to the loss of system node connectivity at two consecutive running moments: represents the node connectivity loss of the system at time t, C t , C t-1 They represent the node connectivity of the system at time t and time t-1 respectively.

[0019] Furthermore, the establishment of the grid congestion risk index to calibrate the preprocessed data is specifically as follows: the grid congestion risk label under the preprocessed data is defined as the line power overload index R, that is: in: K refers to the total number of power grid lines, ρ k is the power overload effect of line k; p k is the transmission power of line k, is the rated maximum transmission power of line k, μ k Represents the importance factor of line k.

[0020] Furthermore, the blocking risk assessment sample set refers to a machine learning training sample set after data calibration under different grid structures and different operating modes.

[0021] Furthermore, the use of feature extraction technology to perform dimensionality reduction processing on the blocking risk assessment sample set is specifically: using a feature extraction algorithm to perform data feature dimensionality reduction, and the feature extraction algorithm adopts a minimum redundancy maximum correlation method or a principal component analysis method.

[0022] Furthermore, the grid congestion risk assessment neural network model adopts a composite model including a long short-term memory neural network and a back propagation neural network.

[0023] The grid congestion risk assessment system considering node connectivity loss includes:

[0024] The module for acquiring data to be processed is used to utilize the grid structure and operation mode data of the target power grid, and combine the simulation data generated by the power system analysis and calculation model and the historical data of the actual operation of the power grid to form the data to be processed;

[0025] Preprocessing module: used to preprocess the data to be processed;

[0026] Blockage risk assessment sample set generation module: used to add an indicator that considers node connectivity loss as a data attribute to the preprocessed data, and establish a power grid blockage risk indicator to calibrate the preprocessed data using the data of the indicator that considers node connectivity loss, and generate a blockage risk assessment sample set;

[0027] Dimensionality reduction module: used to reduce the dimension of the blocking risk assessment sample set using feature extraction technology;

[0028] Evaluation module: It is used to build a power grid congestion risk assessment neural network model for the congestion risk assessment sample set after dimensionality reduction processing, and to train the power grid congestion risk assessment neural network model, and to use the trained power grid congestion risk assessment neural network model to evaluate the power grid congestion risk.

[0029] Compared with the prior art, the present invention has the following beneficial technical effects:

[0030] The method of the present invention does not require an accurate mathematical model of the power grid, but only mines based on simulation data and historical operation data. It can perform offline learning and online learning at the same time, and continuously improve the accuracy of the machine learning power grid congestion risk assessment method, thereby overcoming the shortcoming that the traditional N-1 static safety analysis cannot provide online rapid and accurate early warning of system safety; and by adding the consideration of the node connectivity loss index as a data attribute, it is possible to more fully mine the correlation between power grid topology information and operation risk, improve the accuracy of power grid congestion risk early warning, and provide decision support for operation scheduling for power grid risks, which can significantly improve the safety of actual power grid operation and has significant expected economic and social benefits. BRIEF DESCRIPTION OF THE DRAWINGS

[0031] The drawings in the specification are used to provide further understanding of the present invention and constitute a part of the present invention. The schematic embodiments of the present invention and their descriptions are used to explain the present invention and do not constitute improper limitations on the present invention.

[0032] Figure 1 Flow chart of the grid congestion risk assessment method considering node connectivity loss.

[0033] Figure 2 Schematic diagram of node connectivity calculation. DETAILED DESCRIPTION

[0034] 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.

[0035] 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.

[0036] Embodiment 1

[0037] See also Figure 1 , a grid congestion risk assessment method considering node connectivity loss, utilizes the grid structure of the target grid and the given operating mode data, combines the simulation data generated by the power system analysis and calculation model and the historical data of the actual operation of the grid, calibrates the data according to the data preprocessing technology and the indicators considering the node connectivity loss, and generates a congestion risk assessment sample set; then uses the feature extraction technology to perform data dimensionality reduction processing; and then combines the machine learning algorithm to form a grid congestion risk assessment method.

[0038] Among them, the grid structure of the target power grid refers to various normal, maintenance, newly commissioned lines and main transformers, as well as all grid structure information under the operation and scheduling conditions considering topology control; the operation mode data of the target power grid refers to various power flow operation modes under different conventional generating units, new energy units, loads and DC transmission power parameters, such as winter large, winter small, summer large, summer small, and beginning of the year modes; the simulation data generated by the power system analysis and calculation model refers to steady-state calculation simulation data based on various analysis software that can perform steady-state power flow calculations on power systems, such as PSASP, PSD-BPA, PSS / E, etc.; the historical data of actual operation of the power grid refers to the historical steady-state power flow data recorded from actual operations under various grid structures and various operation modes.

[0039] Data preprocessing technology refers to data denoising, abnormal data cleaning, missing data completion, and data normalization. The indicator considering node connectivity loss refers to the loss of system node connectivity at two consecutive running moments: represents the node connectivity loss of the system at time t, C t , C t-1 Respectively represent the node connectivity of the system at time t and t-1; calibrating the data means that the grid congestion risk label under the data is defined as the line power overload index R, that is: in: K refers to the total number of power grid lines, ρ k is the power overload effect of line k; p k is the transmission power of line k. is the rated maximum transmission power of line k. k It represents the importance factor of line k, which is influenced by factors such as line topology, voltage level, and supplied load level.

[0040] Generating a congestion risk assessment sample set refers to a machine learning training sample set after data calibration under different grid structures and different operating modes; using feature extraction technology to perform data dimensionality reduction processing refers to using feature extraction algorithms such as minimum redundancy maximum correlation and principal component analysis to reduce the dimensionality of data features; combining machine learning algorithms to form a power grid congestion risk assessment method refers to using machine learning algorithms such as back propagation neural networks and long short-term memory neural networks to perform machine learning modeling training and testing for power grid congestion risk assessment.

[0041] Embodiment 2

[0042] The present invention also provides a power grid congestion risk assessment system considering node connectivity loss, comprising:

[0043] The module for acquiring data to be processed is used to utilize the grid structure and operation mode data of the target power grid, and combine the simulation data generated by the power system analysis and calculation model and the historical data of the actual operation of the power grid to form the data to be processed;

[0044] Preprocessing module: used to preprocess the data to be processed;

[0045] Blockage risk assessment sample set generation module: used to add an indicator that considers node connectivity loss as a data attribute to the preprocessed data, and establish a power grid blockage risk indicator to calibrate the preprocessed data using the data of the indicator that considers node connectivity loss, and generate a blockage risk assessment sample set;

[0046] Dimensionality reduction module: used to reduce the dimension of the blocking risk assessment sample set using feature extraction technology;

[0047] Evaluation module: For the congestion risk assessment sample set after dimensionality reduction processing, a power grid congestion risk assessment neural network model is built, and the power grid congestion risk assessment neural network model is trained. The trained power grid congestion risk assessment neural network model is used to evaluate the power grid congestion risk.

[0048] Embodiment 3

[0049] Source-load node connectivity is a quantification of the connectivity level between generator nodes and load nodes in the power system. It can be divided into two types: power supply side and load side. The specific definition is as follows:

[0050] 1) Node connectivity-load side definition:

[0051]

[0052]

[0053]

[0054] Among them C load Represents the node connectivity-load side overall index, represents the source-load connectivity of load i, Gen j Represents the active output of generator j, Load i represents the active power demand of load i, α ij Represents the connectivity coefficient between load i and generator j. ij It represents the electrical distance between the busbars to which load i and generator j are connected, which is represented by the number of the shortest transmission lines between the two busbars plus one.

[0055] 2) Node connectivity - power supply side definition:

[0056]

[0057]

[0058]

[0059] Among them C gen Represents the overall indicator of system node connectivity-power supply side, where represents the source-load connectivity of generator j, Gen j Represents the active output of generator j, Load i represents the active power demand of load i, α ij Represents the connectivity coefficient between load i and generator j. ij It represents the electrical distance between the busbars to which load i and generator j are connected, which is represented by the number of the shortest transmission lines between the two busbars plus one.

[0060] 3) Definition of system node connectivity:

[0061] C=λ load C load +λ gen C gen

[0062] Where C represents the connectivity of the system nodes, λ load ,λ gen They represent the node connectivity load factor and the generator weight respectively, which are set to 1.0 and 1.0 respectively in the present invention.

[0063] by Figure 2 Taking the schematic diagram as an example, the actual calculation of the system node connectivity is as follows:

[0064] Node connectivity-load side calculation:

[0065]

[0066]

[0067]

[0068]

[0069] Node connectivity - power supply side calculation:

[0070]

[0071]

[0072]

[0073]

[0074]

[0075] System node connectivity calculation:

[0076] C=λ load C load +λ gen C gen =1.0*C load +1.0*C gen

[0077] After the transmission line L3 is disconnected, the electrical distance d between the busbars in the system ij The system source-load connectivity index can reflect the changes in the system topology and assist in identifying the risk of power system congestion.

[0078] Then, the method of the present invention is tested according to the following steps:

[0079] Step 1: Generate a sample set for blocking risk assessment. Collect historical information of the IEEE-118 node standard test system, including network topology, active and reactive output of generators, active and reactive demand of loads, voltage of each bus node, load rate of each transmission line, load rate of generators and other information. Use PSASP and other power system simulation calculation software to build an IEEE-118 node system simulation model, and simulate and calculate the operation data of different grid structures and different operation modes. Combine historical data with simulation data to form data to be processed. Perform multi-source data preprocessing on the above data to be processed, including data denoising, abnormal data cleaning, missing data completion and data normalization. Calculate the line power overload index at each moment, and save the calculation results of the power overload index of each transmission line. According to the time sequence, calculate the node connectivity difference between the current moment and the previous moment, and save the calculation results of the system node connectivity loss and the connectivity loss of each load and generator node. Generate a sample set for blocking risk assessment based on the above calculation results. Note that the method of the present invention does not rely on an accurate mathematical model of the power grid, so when historical data is insufficient or there are few operating modes, any power system simulation platform can be used to generate simulation data and expand the data set. The system node connectivity indicators at the initial moment are shown in Tables 1 and 2.

[0080] Table 1 Node connectivity - power supply side

[0081]

[0082]

[0083] Table 2 Node connectivity - load side

[0084]

[0085]

[0086] Step 2: Use the minimum redundancy maximum correlation method to extract features and reduce data dimensionality. First, given a sufficiently large integer n, according to the mutual information I (S n ,Y), successively select features to generate n sequence feature sets Mutual Information I(S n ,Y) is defined as follows:

[0087]

[0088] Where: Y represents the grid congestion risk label, p(x1,x2,…,x n) represents the joint probability density function. Then, the models are trained according to the n feature sets for prediction, and the prediction error sequence E = {e1, e2, …, e n}, select k adjacent errors with smaller mean and variance to form an error set Ω; finally, search for the minimum value in Ω, and its corresponding feature set is the optimal feature set.

[0089] Step 3: Train the machine learning model. Since node connectivity loss is related to timing and the power system is in continuous operation, its risk status is related to the historical operating status. This method constructs a composite model that includes a long short-term memory neural network (LSTM) and a back propagation (BP) neural network. The LSTM inputs historical information to mine the operation rules of the power grid, and the BP neural network inputs the current moment information for correction. The data set is sorted according to the LSTM input parameters, and the neural network model for power grid congestion risk assessment is trained.

[0090] The method of the present invention builds a similar neural network model based on a sample set that does not use node connectivity for comparative experiments, and the results are shown in Table 3.

[0091] Table 3 Comparison results of blocking risk assessment models

[0092]

[0093]

[0094] It can be seen that the prediction accuracy of the method of the present invention is improved from 92.35% to 98.56%. Although the sample feature dimension increases and the single-round training time increases after using the source-load connectivity, its iterative convergence is greatly improved, and the total training time is shortened from 416 seconds to 176 seconds, and the performance is improved by 57.69%. This improvement will be more obvious for large power systems.

[0095] Those skilled in the art will appreciate that embodiments of the present invention may be provided as methods, systems, or computer program products. Therefore, the present invention may take the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware. Moreover, the present invention may take the form of a computer program product implemented 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.

[0096] The present invention is described with reference to flowcharts and / or block diagrams of methods, devices (systems), and computer program products according to embodiments of the present invention. It should be understood that each process and / or block in the flowchart and / or block diagram, as well as the combination of processes and / or blocks 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 produce 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.

[0097] 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 including 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.

[0098] 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.

[0099] 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 its protection scope. Although the present invention has been described in detail with reference to the above embodiments, ordinary technicians in the field should understand that after reading the present invention, those skilled in the art can still make various changes, modifications or equivalent substitutions to the specific implementation methods of the invention, but these changes, modifications or equivalent substitutions are all within the protection scope of the pending claims of the invention.

Claims

1. A grid congestion risk assessment method considering node connectivity loss, characterized in that: include: The data to be processed are formed by using the grid structure and operation mode data of the target power grid, combining the simulation data generated by the power system analysis and calculation model and the historical data of the actual operation of the power grid; Preprocess the data to be processed; An index considering node connectivity loss is added to the preprocessed data as a data attribute. The data considering the node connectivity loss index is used to establish a power grid congestion risk index to calibrate the preprocessed data and generate a congestion risk assessment sample set; wherein the index considering node connectivity loss refers to the loss of system node connectivity at two consecutive operating moments: represents the node connectivity loss of the system at time t, C t , C t-1 Respectively represent the node connectivity of the system at time t and t-1; the establishment of the grid congestion risk index to calibrate the preprocessed data is specifically as follows: the grid congestion risk label under the preprocessed data is defined as the line power overload index R, that is: in: K refers to the total number of power grid lines, ρ k is the power overload effect of line k; p k is the transmission power of line k, is the rated maximum transmission power of line k, μ k Indicates the importance factor of line k; The feature extraction technology is used to reduce the dimension of the blocking risk assessment sample set; For the congestion risk assessment sample set after dimensionality reduction processing, a power grid congestion risk assessment neural network model is built, and the power grid congestion risk assessment neural network model is trained, and the trained power grid congestion risk assessment neural network model is used to assess the power grid congestion risk, wherein the power grid congestion risk assessment neural network model adopts a composite model including a long short-term memory neural network and a back propagation neural network.

2. The method for evaluating grid congestion risk considering node connectivity loss according to claim 1, characterized in that: The grid structure of the target power grid refers to all normal, maintenance, newly commissioned lines and main transformers, as well as all grid structures under the operation and scheduling conditions considering topology control; The operation mode data of the target power grid include various power flow operation modes under conventional generator sets, new energy generator sets, loads and DC transmission power parameters.

3. The method for evaluating grid congestion risk considering node connectivity loss according to claim 2, characterized in that: The simulation data generated by the power system analysis and calculation model refers to steady-state calculation simulation data generated by various analysis software capable of performing steady-state power flow calculations on the power system; The historical data of actual operation of the power grid refers to the historical steady-state power flow data recorded from actual operation under various grid structures and various operation modes.

4. The method for evaluating grid congestion risk considering node connectivity loss according to claim 1, characterized in that: The specific process of preprocessing the data to be processed is: performing data denoising, abnormal data cleaning, missing data filling and data normalization on the data to be processed in sequence.

5. The method for evaluating grid congestion risk considering node connectivity loss according to claim 1, characterized in that: The blocking risk assessment sample set refers to a machine learning training sample set after data calibration under different grid structures and different operating modes.

6. The method for evaluating grid congestion risk considering node connectivity loss according to claim 1, characterized in that: The method of using feature extraction technology to reduce the dimension of the blocking risk assessment sample set specifically includes: using a feature extraction algorithm to reduce the dimension of data features, and the feature extraction algorithm adopts a minimum redundancy maximum correlation method or a principal component analysis method.

7. A grid congestion risk assessment system considering node connectivity loss, characterized in that: include: The module for acquiring data to be processed is used to utilize the grid structure and operation mode data of the target power grid, and combine the simulation data generated by the power system analysis and calculation model and the historical data of the actual operation of the power grid to form the data to be processed; Preprocessing module: used to preprocess the data to be processed; Blockage risk assessment sample set generation module: used to add an indicator that considers node connectivity loss as a data attribute to the preprocessed data, and establish a power grid blockage risk indicator to calibrate the preprocessed data using the data of the indicator that considers node connectivity loss, and generate a blockage risk assessment sample set; wherein the indicator that considers node connectivity loss refers to the loss of system node connectivity at two consecutive operating moments: represents the node connectivity loss of the system at time t, C t , C t-1 Respectively represent the node connectivity of the system at time t, t…1; the establishment of the grid congestion risk index to calibrate the preprocessed data is specifically as follows: the grid congestion risk label under the preprocessed data is defined as the line power overload index R, that is: in: K refers to the total number of power grid lines, ρ k is the power overload effect of line k; p k is the transmission power of line k, is the rated maximum transmission power of line k, μ k Indicates the importance factor of line k; Dimensionality reduction module: used to reduce the dimension of the blocking risk assessment sample set using feature extraction technology; Evaluation module: used to build a power grid congestion risk assessment neural network model for the congestion risk assessment sample set after dimensionality reduction processing, and train the power grid congestion risk assessment neural network model, and use the trained power grid congestion risk assessment neural network model to evaluate the power grid congestion risk, wherein the power grid congestion risk assessment neural network model adopts a composite model including a long short-term memory neural network and a back propagation neural network.