A power grid security monitoring system based on threat alert semantic knowledge mining and transfer learning

Abstract

The present invention discloses a power grid security monitoring system based on threat alert semantic knowledge mining and transfer learning, specifically related to the field of power grid security, including a threat alert semantic transfer learning module, a monitoring area division module, a power grid security data collection module, a power grid security data preprocessing module, a power grid security data analysis module, a power grid security assessment index calculation module, a power grid security monitoring and judgment module, a threat alert data processing and analysis module, a threat alert monitoring and judgment module, and a human-computer interaction module. By collecting power stability data and environmental security data of the monitoring sub-areas, a preprocessing model, an analysis model, and a power grid security assessment index calculation model are established, the power stability state proportional coefficient, the environmental security state proportional coefficient, and the power grid security assessment index are calculated, the state of the monitoring sub-areas is judged, and the monitoring sub-areas in the abnormal state are optimized in a timely manner, providing an important guarantee for taking timely measures to avoid the occurrence of security accidents.

Description

Technical Field

[0001] The present invention relates to the technical field of power grid security. More specifically, the present invention relates to a power grid security monitoring system based on threat warning semantic knowledge mining transfer learning. Background Art

[0002] The modern power grid is a highly complex and interconnected system, consisting of multiple links such as power generation, power transmission, power distribution, and power consumption. This complexity makes the power grid prone to chain reactions when suffering from various internal and external threats, such as natural disasters, equipment failures, and human attacks, resulting in large-scale power outages or even system collapses. Therefore, it is particularly important to conduct real-time and comprehensive security monitoring of the power grid.

[0003] Traditional power grid security monitoring systems include a data acquisition module, a data analysis module, and a human-computer interaction module. Among them, the data acquisition module collects power grid security data through existing tools, including electrical energy indicators such as current, voltage, and frequency; the data analysis module is used to analyze the collected data to obtain electrical energy-related parameters; the human-computer interaction module is used to display and store the analyzed electrical energy-related parameters, and the staff processes the unsafe points of power grid monitoring based on the display results.

[0004] However, in actual use, there are still some disadvantages. For example, first, due to the complexity of the power grid and the diversity of threats, traditional power grid security monitoring systems are no longer able to meet the requirements. Therefore, it is necessary to introduce advanced technical means to improve the intelligent level and response speed of power grid security monitoring. For example, threat warning semantic transfer learning can analyze and learn a large amount of historical data, and then build a more comprehensive and accurate security monitoring model; second, the data collected by traditional power grid security monitoring systems is often incomplete and requires a large amount of calculation and analysis to obtain the analysis results, which takes a lot of time and effort. Therefore, it is necessary to establish a more efficient and accurate model to improve the efficiency of power grid security monitoring. Summary of the Invention

[0005] In order to overcome the above-mentioned defects of the prior art, an embodiment of the present invention provides a power grid security monitoring system based on threat warning semantic knowledge mining transfer learning, and through the following solutions, to solve the problems raised in the above background art.

[0006] To achieve the above object, the present invention provides the following technical solution: A power grid security monitoring system based on threat warning semantic knowledge mining transfer learning, comprising:

[0007] The threat warning semantic transfer learning module is used to transfer a trained security monitoring model in a power system to a target security monitoring system that needs to be applied, and perform fine-tuning to adapt to the new data set;

[0008] The monitoring area division module is used to deploy monitoring devices on the key nodes and important equipment in the power grid security monitoring area after transfer learning, establish a complete power grid security monitoring network, and determine the key nodes and important equipment with monitoring devices deployed as each monitoring sub-area, numbered 1, 2, …, i, …, n in sequence;

[0009] The power grid security data acquisition module is used to acquire various types of data in each sub-monitoring area, including a power stability data acquisition unit and an environmental security data acquisition unit, and transmit the acquired data to the power grid security data preprocessing module;

[0010] The power grid security data preprocessing module is used to preprocess the data acquired in the power stability data acquisition module, including a power stability preprocessing unit and an environmental security preprocessing unit. The voltage stability coefficient, current stability coefficient, and frequency stability coefficient are obtained through the power stability preprocessing unit, and the environmental temperature coefficient and environmental humidity coefficient are obtained through the environmental security preprocessing unit, and the preprocessed data is transmitted to the power grid security data analysis module;

[0011] The power grid security data analysis module is used to analyze the data preprocessed in the power stability data preprocessing module, including a power stability analysis unit and an environmental security analysis unit, obtain the power stability state proportion coefficient and the environmental security state proportion coefficient, and transmit the analyzed data to the power grid security assessment index calculation module;

[0012] The power grid security assessment index calculation module is used to establish a power grid security assessment index calculation model, import the power stability state proportion coefficient and the environmental security state proportion coefficient into the power grid security assessment index calculation model, calculate the power grid security assessment index, and transmit the analyzed data to the power grid security monitoring and judgment module;

[0013] The power grid security monitoring and judgment module is used to judge the data transmitted by the power grid security assessment index calculation module, screen each monitoring sub-area in the monitoring network that is in a normal working state and an abnormal working state, and count the numbers of each monitoring sub-area in the monitoring network that is in a normal working state and an abnormal working state;

[0014] The threat warning data processing and analysis module is used to establish a threat warning data processing and analysis model, including a threat warning data processing model and a threat warning data analysis model, and process and analyze the threat warning data, thereby obtaining a threat warning monitoring efficiency assessment index;

[0015] The threat warning monitoring and judgment module is used to establish a preset value of the threat warning monitoring efficiency evaluation index, judge the threat warning monitoring efficiency evaluation index of a specified monitoring area according to the preset value of the threat warning monitoring efficiency evaluation index, and send a signal to the human-computer interaction module according to the judgment result;

[0016] The human-computer interaction module is used to receive the threat warning and warning instruction transmitted by the threat warning monitoring and judgment module, display the threat warning and warning instruction on the terminal display screen, and perform corresponding warning reminders according to the received threat warning and warning instruction.

[0017] Preferably, the security monitoring model in the threat warning semantic transfer learning module learns rich security knowledge and threat warning semantic information based on a large amount of data through training; the target security monitoring system to be applied faces challenges such as data scarcity or the continuous emergence of new threats and needs to perform transfer learning; during the transfer process, it is necessary to clean, denoise, and standardize the data of the security monitoring model and the target security monitoring system to be applied to ensure the quality and consistency of the data; based on the similarity between the security monitoring model and the target security monitoring system to be applied and the availability of the data, perform model-based transfer, and use the data of the target security monitoring system for fine-tuning. During the fine-tuning process, adjust the parameters or structure of the model to adapt to the new data set of the target security monitoring system; use the test set of the target security monitoring system to evaluate the transferred security monitoring model, and further optimize and adjust the model according to the evaluation results to improve its accuracy; deploy the optimized security monitoring model to the actual security monitoring system for real-time monitoring and identification of potential threats.

[0018] Preferably, the power stability data acquisition unit in the power grid security data acquisition module is used to acquire the power stability data of each sub-monitoring area, including voltage stability data, current stability data, and power stability data. The voltage stability data includes the effective value of voltage, voltage harmonic content, fluctuating voltage, and voltage unbalance degree, which are denoted as Uc, Uh, Uf, and Ue respectively. The current stability data includes the effective value of current, current harmonic content, fluctuating current, and current unbalance degree, which are denoted as Ic, Ih, If, and Ie respectively; the power stability data includes the post-accident steady-state power, fault impact power, and post-accident static stability maximum power, which are denoted as Ps, Pb, and Pm respectively; the environmental security data acquisition unit is used to acquire the environmental security data of each monitoring sub-area, including environmental temperature data and environmental humidity data. The environmental temperature data includes the adiabatic body temperature and the temperature gradient change rate, which are denoted as Tc and Tl respectively. The internal environmental data includes the relative humidity and the non-condensing humidity change rate, which are denoted as Hc and Hl respectively.

[0019] Preferably, the power stability preprocessing unit in the power grid security data preprocessing module is used to establish a power stability preprocessing model, import the data transmitted by the power grid security data acquisition module into the power stability preprocessing model, and obtain the voltage stability coefficient, current stability coefficient, and frequency stability coefficient. The specific calculation formula for the voltage stability coefficient is:

[0020]

[0021] Where Ur i represents the voltage stability coefficient of the i-th monitoring sub-region, Uc i represents the effective voltage value of the i-th monitoring sub-region, Uc 允 represents the maximum allowable value of the effective voltage value for power grid security monitoring, Ue i represents the voltage unbalance degree of the i-th monitoring sub-region, Uh i represents the voltage harmonic content of the i-th monitoring sub-region, Uf i represents the voltage fluctuation of the i-th monitoring sub-region;

[0022] The specific calculation formula for the current stability coefficient is:

[0023]

[0024] Where Ir i represents the current stability coefficient of the i-th monitoring sub-region, Ic i represents the effective current value of the i-th monitoring sub-region, Ic 允 represents the maximum allowable value of the effective current value for power grid security monitoring, Ie i represents the current unbalance degree of the i-th monitoring sub-region, Ih i represents the current harmonic content of the i-th monitoring sub-region, If i represents the current fluctuation of the i-th monitoring sub-region;

[0025] The specific calculation formula for the frequency stability coefficient is:

[0026]

[0027] Where Pr i represents the frequency stability coefficient of the i-th monitoring sub-region, Ps i represents the post-accident steady-state power of the i-th monitoring sub-region, Ps 标 represents the standard post-accident steady-state power for power grid security monitoring, Pb i represents the fault impact power of the i-th monitoring sub-region, Pb max represents the maximum fault impact power, Pm i represents the maximum post-accident static stability power of the i-th monitoring sub-region, Pm 标Represents the standard post-accident static stability maximum power of power grid security monitoring.

[0028] Preferably, the environmental security preprocessing unit in the power grid security data preprocessing module is used to establish an environmental security preprocessing model, import the data transmitted in the power grid security data acquisition module into the environmental security preprocessing model, and obtain the environmental temperature stability coefficient and the environmental humidity stability coefficient. The specific analysis formula for the environmental temperature stability coefficient is:

[0029]

[0030] Where Tr i Represents the environmental temperature stability coefficient of the i-th monitoring sub-region, Tc i Represents the adiabatic body temperature of the i-th monitoring sub-region, Tl i Represents the temperature gradient change rate of the i-th monitoring sub-region;

[0031] The specific analysis formula for the environmental humidity stability coefficient is:

[0032]

[0033] Where Hr i Represents the environmental humidity stability coefficient of the i-th monitoring sub-region, Hc i Represents the relative humidity of the i-th monitoring sub-region, Hl i Represents the non-condensing humidity change rate of the i-th monitoring sub-region.

[0034] Preferably, the power stability analysis unit in the power grid security data analysis module is used to establish a power stability analysis model, import the data transmitted in the power stability preprocessing unit into the power stability analysis model, and obtain the power stability state proportionality coefficient. The specific calculation formula is: φ i = Ur i × Ir i × Pr i , where φ i Represents the power stability state proportionality coefficient of the i-th sub-region, Ur i Represents the voltage stability coefficient of the i-th monitoring sub-region, Ir i Represents the current stability coefficient of the i-th monitoring sub-region, Pr i Represents the frequency stability coefficient of the i-th monitoring sub-region; The environmental security analysis unit is used to establish an environmental security analysis model, import the data transmitted in the environmental security preprocessing unit into the environmental security analysis model, and obtain the environmental security state proportionality coefficient. The specific calculation formula is: ψ i = Tr i × Hr i , where ψ iDenote the environmental safety status proportion coefficient of the \(i\)-th sub-region, \(Tr\). i Denote the environmental temperature stability coefficient of the \(i\)-th monitoring sub-region, \(Hr\). i Denote the environmental humidity stability coefficient of the \(i\)-th monitoring sub-region.

[0035] Preferably, the power grid security assessment index is obtained by establishing a power grid security assessment index calculation model, and the specific analysis formula of the power grid security assessment index is:

[0036] Υ i = log2(1 + φ i ) + log3(1 + ψ i ),

[0037] where \(Y\) i denotes the power grid security assessment index of the \(i\)-th monitoring sub-region, \(φ\) i denotes the power stability status proportion coefficient of the \(i\)-th sub-region, \(ψ\) i denotes the environmental safety status proportion coefficient of the \(i\)-th sub-region.

[0038] Preferably, the power grid security monitoring and judgment module compares the power grid security assessment index corresponding to each monitoring sub-region in the monitoring network with the standard power grid security assessment index corresponding to each preset monitoring sub-region. If the power grid security assessment index corresponding to a certain monitoring sub-region is greater than or equal to the standard power grid security assessment index in the preset normal working state, it indicates that the monitoring sub-region is in the normal working state. Otherwise, it indicates that the monitoring sub-region is in the abnormal working state. Then the system monitors that there is an abnormal situation affecting the normal operation of the power grid, and screens and counts each monitoring sub-region in the abnormal working state and each monitoring sub-region in the normal working state according to the working state of the monitoring sub-region, maintains data collection and analysis for each monitoring sub-region in the normal working state, records each monitoring sub-region in the abnormal working state as each designated monitoring sub-region, and numbers them sequentially as 1, 2, …, \(j\), …, \(k\).

[0039] Preferably, based on each monitoring sub-region in the abnormal state, the threat warning data processing model collects historical threat warning data through the power grid security monitoring system, cleans the collected data, and performs defense measures on the clustering analysis results of the obtained data through the clustering analysis method, and obtains the threat warning recognition rate, emergency response speed, and threat warning protection rate of each designated monitoring sub-region, which are denoted as \(Er\), \(Et\), and \(Ed\) respectively; the threat warning data analysis model is used to analyze the results after system identification and protection to obtain the threat warning monitoring efficiency evaluation index, and its specific analysis formula is:

[0040]

[0041] Among them, η represents the threat monitoring efficiency evaluation index of the specified monitoring area, Er j represents the threat alarm recognition rate of the j-th specified monitoring sub-area, Ed j represents the threat alarm protection rate of the j-th specified monitoring sub-area, Et j represents the emergency response speed of the j-th specified monitoring sub-area.

[0042] Preferably, the threat alarm monitoring and judgment module compares the threat alarm monitoring efficiency evaluation index of the specified monitoring area with the preset value η of the threat alarm monitoring efficiency evaluation index D If η D ≤η, it indicates that the ability of the specified monitoring area to handle threat alarms is good. If η D >η, it indicates that the ability of the specified monitoring area to handle threat alarms is poor, and then a threat alarm warning instruction is sent to the human-computer interaction module.

[0043] The technical effects and advantages of the present invention:

[0044] 1. The present invention performs transfer learning on the target power grid security monitoring model through the threat alarm semantic transfer learning module, providing a more refined mathematical model for subsequent power grid security monitoring; through the monitoring area division module, the monitoring area is divided into multiple monitoring sub-areas, enabling refined monitoring of power grid security. The data of each monitoring sub-area is processed independently, enabling the system to more accurately capture subtle changes during monitoring; through the power grid security data collection module, the power stability data and environmental security data of each monitoring sub-area are collected. These data are important bases for evaluating power grid security and provide data support for subsequent data analysis and warning; through the power grid security data preprocessing module, the data in the power grid security data collection module is preprocessed, and a preprocessing model is established to simplify the complex and changeable data, reducing the system operation load and improving the system operation efficiency; through the power grid security data analysis module, the data in the power grid security data preprocessing module is further analyzed, and an analysis model is established. By calculating, the power stability state proportion coefficient and the environmental security state proportion coefficient are obtained, which is of great significance for evaluating the security and stability of power grid operation;

[0045] 2. The power grid security assessment index calculation model is established through the power grid security assessment index calculation module, which can comprehensively reflect the overall operation status of the power grid. The overall operation status of the power grid is judged through the power grid security monitoring and judgment module, and the monitored sub-areas in abnormal status and the monitored sub-areas in normal status are judged. The monitored sub-areas in abnormal status are further analyzed. Through the threat warning data processing and analysis module and the threat warning monitoring and judgment module, the specified monitored sub-areas in abnormal status are further processed, analyzed and judged, and the monitored sub-areas in abnormal status are optimized in time, which provides an important guarantee for taking timely measures to avoid the occurrence of security accidents. Brief Description of the Drawings

[0046] Figure 1 It is a schematic diagram of the overall structure of the present invention. Detailed Embodiment

[0047] Next, the technical solutions in the embodiments of the present invention will be clearly and completely described in conjunction with the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts shall fall within the protection scope of the present invention.

[0048] As shown in the attached Figure 1 A power grid security monitoring system based on threat warning semantic knowledge mining and transfer learning is shown in the figure, including a threat warning semantic transfer learning module, a monitoring area division module, a power grid security data acquisition module, a power grid security data preprocessing module, a power grid security data analysis module, a power grid security assessment index calculation module, a power grid security monitoring and judgment module, a threat warning data processing and analysis module, a threat warning monitoring and judgment module, and a human-computer interaction module.

[0049] The output end of the threat warning semantic transfer learning module is connected to the input end of the monitoring area division module through telecommunications. The output end of the monitoring area division module is connected to the input end of the power grid security data collection module through telecommunications. The output end of the power grid security data collection module is connected to the input end of the power grid security data preprocessing module through telecommunications. The output end of the power grid security data preprocessing module is connected to the input end of the power grid security data analysis module through telecommunications. The output end of the power grid security data analysis module is connected to the input end of the power grid security assessment index calculation module through telecommunications. The output end of the power grid security assessment index calculation module is connected to the input end of the power grid security data collection module through telecommunications. The output end of the power grid security data collection module is connected to the input end of the power grid security monitoring and judgment module through telecommunications. The output end of the power grid security monitoring and judgment module is connected to the input end of the threat warning data processing and analysis module through telecommunications. The output end of the threat warning data processing and analysis module is connected to the input end of the threat warning monitoring and judgment module through telecommunications. The output end of the threat warning monitoring and judgment module is connected to the input end of the human-machine interaction module through telecommunications.

[0050] The threat warning semantic transfer learning module is used to transfer a trained security monitoring model in a power system to a target security monitoring system that needs to be applied, and perform fine-tuning to adapt to the new data set.

[0051] In this embodiment, it should be specifically noted that the security monitoring model learns rich security knowledge and threat warning semantic information through training. The target security monitoring system that needs to be applied faces challenges such as scarce data or emerging new threats, and transfer learning is required. During the transfer process, it is necessary to clean, denoise, and standardize the data of the security monitoring model and the target security monitoring system that needs to be applied to ensure the quality and consistency of the data. Based on the similarity between the security monitoring model and the target security monitoring system that needs to be applied and the availability of the data, perform model-based transfer, and use the data of the target security monitoring system for fine-tuning. During the fine-tuning process, adjust the parameters or structure of the model to adapt to the new data set of the target security monitoring system. Use the test set of the target security monitoring system to evaluate the transferred security monitoring model, and further optimize and adjust the model according to the evaluation results to improve its accuracy. Deploy the optimized security monitoring model to the actual security monitoring system for real-time monitoring and identification of potential threats.

[0052] The monitoring area division module is used to deploy monitoring devices on the key nodes and important equipment in the power grid security monitoring area after transfer learning, establish a complete power grid security monitoring network, and determine the key nodes and important equipment with monitoring devices as each monitoring sub-area, numbered 1, 2,..., i,..., n in sequence.

[0053] In this embodiment, it should be specifically noted that the monitoring device can collect power grid operation data in real time and accurately. Common monitoring devices include sensors, data collectors, and monitoring instruments. The key nodes and important equipment in the power grid security monitoring area after migration learning usually include high-voltage substations, the intersection points of transmission lines, and important load access points. These nodes and equipment undertake important transmission and distribution tasks in the power grid. Once a fault or anomaly occurs, it will have a greater impact on the overall operation of the power grid.

[0054] The power grid security data acquisition module is used to collect various types of data in each sub-monitoring area, including the power stability data acquisition unit and the environmental security data acquisition unit, and transmit the collected data to the power grid security data preprocessing module.

[0055] In this embodiment, it should be specifically noted that the power stability data acquisition unit is used to collect the power stability data in each sub-monitoring area, including voltage stability data, current stability data, and power stability data. The voltage stability data includes the effective value of voltage, voltage harmonic content, fluctuating voltage, and voltage unbalance degree, which are denoted as Uc, Uh, Uf, and Ue respectively. The current stability data includes the effective value of current, current harmonic content, fluctuating current, and current unbalance degree, which are denoted as Ic, Ih, If, and Ie respectively. The power stability data includes the post-accident steady-state power, fault impact power, and post-accident static stability maximum power, which are denoted as Ps, Pb, and Pm respectively.

[0056] Furthermore, the voltage harmonic content and current harmonic content are obtained by sampling and digitizing the current signal through fast Fourier transform to obtain the frequency and amplitude of each harmonic, and calculating the percentage of its fundamental wave amplitude to obtain the harmonic content. The fluctuating voltage and fluctuating current are measured in each monitoring area through a voltage fluctuation recorder and a current fluctuation recorder respectively, and the fluctuating voltage and fluctuating current are obtained through square demodulation detection. The voltage unbalance degree and current unbalance degree are measured by a power quality analyzer. The post-accident steady-state power is the power value under the new steady state reached by the system after a fault or disturbance. The fault impact power is the power peak determined by the magnitude of the transient energy accumulated during the fault. The post-accident static stability maximum power is the maximum power that the system can transmit under the condition of maintaining static stability after a fault.

[0057] In this embodiment, it should be specifically noted that the environmental security data acquisition unit is used to collect the environmental security data in each monitoring sub-area, including environmental temperature data and environmental humidity data. The environmental temperature data includes the adiabatic body temperature and the temperature gradient change rate, which are denoted as Tc and Tl respectively. The internal environmental data includes the relative humidity and the non-condensation humidity change rate, which are denoted as Hc and Hl respectively.

[0058] The power grid security data preprocessing module is used to preprocess the data collected by the power grid security data acquisition module, including a power stability preprocessing unit and an environmental security preprocessing unit. The voltage stability coefficient, current stability coefficient, and frequency stability coefficient are obtained through the power stability preprocessing unit, and the environmental temperature coefficient and environmental humidity coefficient are obtained through the environmental security preprocessing unit. The preprocessed data is then transmitted to the power grid security data analysis module.

[0059] In this embodiment, it should be specifically noted that the power stability preprocessing unit is used to establish a power stability preprocessing model, import the data transmitted by the power grid security data acquisition module into the power stability preprocessing model, and obtain the voltage stability coefficient, current stability coefficient, and frequency stability coefficient. The specific calculation formula for the voltage stability coefficient is:

[0060]

[0061] Where Ur i represents the voltage stability coefficient of the i-th monitoring sub-region, Uc i represents the effective voltage value of the i-th monitoring sub-region, Uc 允 represents the maximum allowable value of the effective voltage value of the power grid security monitoring, Ue i represents the voltage unbalance degree of the i-th monitoring sub-region, Uh i represents the voltage harmonic content of the i-th monitoring sub-region, Uf i represents the voltage fluctuation of the i-th monitoring sub-region;

[0062] The specific calculation formula for the current stability coefficient is:

[0063]

[0064] Where Ir i represents the current stability coefficient of the i-th monitoring sub-region, Ic i represents the effective current value of the i-th monitoring sub-region, Ic 允 represents the maximum allowable value of the effective current value of the power grid security monitoring, Ie i represents the current unbalance degree of the i-th monitoring sub-region, Ih i represents the current harmonic content of the i-th monitoring sub-region, If i represents the current fluctuation of the i-th monitoring sub-region;

[0065] The specific calculation formula for the frequency stability coefficient is:

[0066]

[0067] Where Pr i represents the frequency stability coefficient of the i-th monitoring sub-region, Ps iDenote the post-accident steady-state power of the $i$-th monitoring sub-region as $P_s$. 标 Denote the standard post-accident steady-state power for grid security monitoring as $P_b$. i Denote the fault impact power of the $i$-th monitoring sub-region as $P_b$. max Denote the maximum fault impact power as $P_m$. i Denote the maximum post-accident static stability power of the $i$-th monitoring sub-region as $P_m$. 标 Denote the standard maximum post-accident static stability power for grid security monitoring.

[0068] In this embodiment, it should be specifically noted that the environmental safety preprocessing unit is used to establish an environmental safety preprocessing model, import the data transmitted in the grid security data acquisition module into the environmental safety preprocessing model, and obtain the environmental temperature stability coefficient and the environmental humidity stability coefficient. The specific analysis formula for the environmental temperature stability coefficient is:

[0069]

[0070] where $T_r$ i Denote the environmental temperature stability coefficient of the $i$-th monitoring sub-region as $T_c$ i Denote the adiabatic body temperature of the $i$-th monitoring sub-region as $T_l$ i Denote the temperature gradient change rate of the $i$-th monitoring sub-region;

[0071] The specific analysis formula for the environmental humidity stability coefficient is:

[0072]

[0073] where $H_r$ i Denote the environmental humidity stability coefficient of the $i$-th monitoring sub-region as $H_c$ i Denote the relative humidity of the $i$-th monitoring sub-region as $H_l$ i Denote the non-condensing humidity change rate of the $i$-th monitoring sub-region.

[0074] The grid security data analysis module is used to analyze the data preprocessed in the power stability data preprocessing module, including the power stability analysis unit and the environmental safety analysis unit, obtain the power stability state proportion coefficient and the environmental safety state proportion coefficient, and transmit the analyzed data to the grid security assessment index calculation module.

[0075] In this embodiment, it should be specifically noted that the power stability analysis unit is used to establish a power stability analysis model, import the data transmitted in the power stability preprocessing unit into the power stability analysis model, and obtain the power stability state proportion coefficient. The specific calculation formula is: $\varphi$ i $=U_r$ i $\times I_r$ i $\times P_r$ i, where φ i represents the power stability state proportion coefficient of the i-th sub-region, Ur i represents the voltage stability coefficient of the i-th monitoring sub-region, Ir i represents the current stability coefficient of the i-th monitoring sub-region, Pr i represents the frequency stability coefficient of the i-th monitoring sub-region.

[0076] In this embodiment, it should be specifically noted that the environmental safety analysis unit is used to establish an environmental safety analysis model, import the data transmitted in the environmental safety preprocessing unit into the environmental safety analysis model, and obtain the environmental safety state proportion coefficient. The specific calculation formula is: ψ i = Tr i × Hr i , where ψ i represents the environmental safety state proportion coefficient of the i-th sub-region, Tr i represents the environmental temperature stability coefficient of the i-th monitoring sub-region, Hr i represents the environmental humidity stability coefficient of the i-th monitoring sub-region.

[0077] The power grid safety assessment index calculation module is used to establish a power grid safety assessment index calculation model, import the power stability state proportion coefficient and the environmental safety state proportion coefficient into the power grid safety assessment index calculation model, calculate the power grid safety assessment index, and transmit the analysis data to the power grid safety monitoring and judgment module.

[0078] In this embodiment, it should be specifically noted that the power grid safety assessment index is obtained by establishing a power grid safety assessment index calculation model. The specific analysis formula of the power grid safety assessment index is:

[0079] Υ i = log2(1 + φ i ) + log3(1 + ψ i )

[0080] where Y i represents the power grid safety assessment index of the i-th monitoring sub-region, φ i represents the power stability state proportion coefficient of the i-th sub-region, ψ i represents the environmental safety state proportion coefficient of the i-th sub-region.

[0081] The power grid safety monitoring and judgment module is used to judge the data transmitted by the power grid safety assessment index calculation module, screen the monitoring sub-regions in the monitoring network that are in normal working state and abnormal working state, and count the numbers of the monitoring sub-regions in the monitoring network that are in normal working state and abnormal working state.

[0082] In this embodiment, it should be specifically noted that the power grid security assessment indices corresponding to the monitoring sub-regions in the monitoring network are compared with the standard power grid security assessment indices corresponding to the preset monitoring sub-regions. If the power grid security assessment index corresponding to a certain monitoring sub-region is greater than or equal to the standard power grid security assessment index in the preset normal working state, it indicates that the monitoring sub-region is in a normal working state. On the contrary, it indicates that the monitoring sub-region is in an abnormal working state. Then the system monitors that there are abnormal situations affecting the normal operation of the power grid, such as equipment failures, network attacks, and abnormal power loads, and filters and counts the monitoring sub-regions in the abnormal working state and the monitoring sub-regions in the normal working state according to the working state of the monitoring sub-regions. The data collection and analysis of the monitoring sub-regions in the normal working state are maintained, and the monitoring sub-regions in the abnormal working state are recorded as the designated monitoring sub-regions, which are numbered 1, 2, …, j, …, k in sequence.

[0083] The threat warning data processing and analysis module is used to establish a threat warning data processing and analysis model, including a threat warning data processing model and a threat warning data analysis model, to process and analyze the threat warning data, and thus obtain a threat warning monitoring efficiency evaluation index.

[0084] In this embodiment, it should be specifically noted that the threat warning data processing model is based on the monitoring sub-regions in the abnormal state. Historical threat warning data is collected through the power grid security monitoring system, including threat warning types, threat warning levels, threatened device information, and fault descriptions. The collected data is cleaned to remove duplicate data, invalid data, and noise data. The warnings with similar characteristics are normalized, and the obtained data is organized into groups through a clustering analysis method to identify common threat patterns, abnormal behaviors, and potential risks. Defense measures are taken for the clustering analysis results through transfer learning, and the threat warning recognition rate, emergency response speed, and threat warning protection rate of each designated monitoring sub-region are obtained, which are denoted as Er, Et, and Ed respectively. The threat warning data analysis model is used to analyze the results after system identification and protection to obtain a threat warning monitoring efficiency evaluation index. The specific analysis formula is as follows:

[0085]

[0086] where η represents the threat monitoring efficiency evaluation index of the designated monitoring area, Er j represents the threat warning recognition rate of the jth designated monitoring sub-region, Ed j represents the threat warning protection rate of the jth designated monitoring sub-region, Et j represents the emergency response speed of the jth designated monitoring sub-region.

[0087] The threat alarm monitoring and judgment module is used to establish a preset value of the threat alarm monitoring efficiency evaluation index, judge the threat alarm monitoring efficiency evaluation index of the specified monitoring area according to the preset value of the threat alarm monitoring efficiency evaluation index, and send a signal to the human-computer interaction module according to the judgment result.

[0088] In this embodiment, it should be specifically noted that the threat alarm monitoring efficiency evaluation index of the specified monitoring area is compared with the preset value η of the threat alarm monitoring efficiency evaluation index D If η D ≤η, it means that the threat alarm monitoring efficiency evaluation index of the specified monitoring area is greater than or equal to the preset value of the threat alarm monitoring efficiency evaluation index, indicating that the threat prevention and alarm ability of the specified monitoring area is good. If η D >η, it means that the threat alarm monitoring efficiency evaluation index of the specified monitoring area is less than the preset value of the threat alarm monitoring efficiency evaluation index, indicating that the threat prevention and alarm ability of the specified monitoring area is poor, and then a threat alarm warning instruction is sent to the human-computer interaction module.

[0089] The human-computer interaction module is used to receive the threat alarm warning instruction transmitted by the threat alarm monitoring and judgment module, display the threat alarm warning instruction on the terminal display screen, and perform corresponding warning reminders according to the received threat alarm warning instruction.

[0090] In this embodiment, it should be specifically noted that the model of the specified monitoring area with poor threat prevention and alarm ability is optimized through the threat alarm warning instruction, and a monitoring report is generated based on the power stable state proportion coefficient, the environmental safety state proportion coefficient, the power grid safety evaluation index, and the threat alarm monitoring efficiency evaluation index. The monitoring report is presented to the operation and maintenance personnel, and the operation and maintenance personnel improve and optimize the analysis method and results according to the monitoring report, so as to improve the safety and stability of the power grid security monitoring area and ensure the safe operation of the power grid.

[0091] The present invention performs transfer learning on the target power grid security monitoring model through a threat warning semantic transfer learning module, providing a more refined mathematical model for subsequent power grid security monitoring; divides the monitoring area into multiple monitoring sub-areas through a monitoring area division module, enabling refined monitoring of power grid security. The data of each monitoring sub-area is processed independently, allowing the system to more accurately capture subtle changes during monitoring; collects power stability data and environmental security data of each monitoring sub-area through a power grid security data collection module. These data are important bases for evaluating power grid security and provide data support for subsequent data analysis and warning; preprocesses the data in the power grid security data collection module through a power grid security data preprocessing module, establishes a preprocessing model, simplifies the complex and variable data, reduces the system operation load, and improves the system operation efficiency; further analyzes the data in the power grid security data preprocessing module through a power grid security data analysis module, establishes an analysis model, and calculates the power stability state proportion coefficient and the environmental security state proportion coefficient, which is of great significance for evaluating the security and stability of power grid operation; establishes a power grid security evaluation index calculation model through a power grid security evaluation index calculation module, which can comprehensively reflect the overall operation state of the power grid; judges the overall operation state of the power grid through a power grid security monitoring judgment module, determines the monitoring sub-areas in abnormal state and the monitoring sub-areas in normal state, and further analyzes the monitoring sub-areas in abnormal state; further processes, analyzes, and judges the specified monitoring sub-areas in abnormal state through a threat warning data processing and analysis module and a threat warning monitoring judgment module, and timely optimizes the monitoring sub-areas in abnormal state, providing an important guarantee for taking timely measures to avoid the occurrence of safety accidents.

[0092] Secondly: In the attached drawings of the disclosed embodiments of the present invention, only the structures related to the disclosed embodiments are involved. For other structures, reference can be made to the general design. Without conflict, the same embodiment and different embodiments of the present invention can be combined with each other;

[0093] Finally: The above are only the preferred embodiments of the present invention and are not used to limit the present invention. Any modifications, equivalent replacements, improvements, etc. made within the spirit and principle of the present invention shall be included within the protection scope of the present invention.

Claims

1. A power grid security monitoring system based on threat warning semantic knowledge mining and transfer learning, characterized in that: include: Threat alert semantic transfer learning module: used to migrate a security monitoring model trained in a power system to the target security monitoring system where it needs to be applied, and fine-tune it to adapt to the new data set; Monitoring area division module: used to deploy monitoring equipment on key nodes and important equipment in the power grid security monitoring area after transfer learning, establish a complete power grid security monitoring network, and determine the key nodes and important equipment where monitoring equipment are deployed as each monitoring sub-area, numbered 1, 2, ..., i, ..., n in sequence; Grid security data acquisition module: used to collect various types of data from each sub-monitoring area, including power stability data acquisition unit and environmental safety data acquisition unit, and transmit the collected data to the grid security data preprocessing module; A power grid safety data preprocessing module is used to preprocess the data collected in the power grid safety data acquisition module. The power grid safety data preprocessing module includes a power stability preprocessing unit and an environmental safety preprocessing unit. The power stability preprocessing unit is used to obtain a voltage stability coefficient, a current stability coefficient, and a frequency stability coefficient. The environmental temperature coefficient and the environmental humidity coefficient are obtained by the environmental safety preprocessing unit, and the obtained data are transmitted to the power grid safety data analysis module. The power grid security data analysis module is used to analyze the data preprocessed in the power grid security data preprocessing module, including the power stability analysis unit and the environmental safety analysis unit, obtain the power stability state proportional coefficient and the environmental safety state proportional coefficient, and transmit the analyzed data to the power grid security assessment index calculation module; The power stability state proportional coefficient is calculated by the power stability analysis unit to establish a power stability analysis model, and the data transmitted in the power grid security data preprocessing module is imported into the power stability analysis model. The specific calculation formula is: i =Ur i ×Ir i ×Pr i , where φ i represents the power stability proportionality coefficient of the ith sub-region, Ur i represents the voltage stability coefficient of the ith monitoring sub-area, Ir i represents the current stability coefficient of the ith monitoring sub-area, Pr i represents the frequency stability coefficient of the ith monitoring sub-area; The specific calculation formula of the voltage stability coefficient is: Among them Ur i represents the voltage stability coefficient of the ith monitoring sub-area, Uc i Indicates the effective value of the voltage in the ith monitoring sub-area, Uc 允 Indicates the maximum allowable value of the effective value of the power grid safety monitoring voltage, Ue i Indicates the voltage unbalance degree of the ith monitoring sub-area, Uh i represents the voltage harmonic content of the ith monitoring sub-area, Uf i represents the voltage fluctuation of the ith monitoring sub-area; The specific calculation formula of the current stability coefficient is: Among them, Ir i represents the current stability coefficient of the ith monitoring sub-area, Ic i represents the effective value of the current in the ith monitoring sub-area, Ic 允 Indicates the maximum allowable value of the effective value of the power grid safety monitoring current, Ie i Indicates the current unbalance degree of the i-th monitoring sub-area, Ih i Indicates the current harmonic content of the i-th monitoring sub-area, If i represents the current fluctuation of the i-th monitoring sub-area; The specific calculation formula of the frequency stability coefficient is: Among them, Pr i represents the frequency stability coefficient of the ith monitoring sub-area, Ps i represents the post-accident steady-state power of the ith monitoring sub-area, Ps 标 represents the standard post-accident steady-state power for power grid security monitoring, Pb i represents the fault impulse power of the ith monitoring sub-area, Pb max Indicates the maximum fault impact power, Pm i Pm represents the maximum static stable power after the accident in the ith monitoring sub-area. 标 It indicates the maximum static stable power after the standard accident of power grid security monitoring; Power grid security assessment index calculation module: used to establish a power grid security assessment index calculation model, import the power stability state proportional coefficient and the environmental safety state proportional coefficient into the power grid security assessment index calculation model, calculate the power grid security assessment index, and transmit the analysis data to the power grid security monitoring and judgment module; Power grid security monitoring and judgment module: used to judge the data transmitted by the power grid security assessment index calculation module, screen the monitoring sub-areas in normal working state and abnormal working state in the monitoring network, and count the numbers of the monitoring sub-areas in normal working state and abnormal working state in the monitoring network; Threat alarm data processing and analysis module: used to establish threat alarm data processing and analysis models, including threat alarm data processing model and threat alarm data analysis model, and process and analyze threat alarm data to obtain threat alarm monitoring efficiency evaluation index; Threat alarm monitoring and judgment module: used to establish a preset value of the threat alarm monitoring efficiency evaluation index, judge the threat alarm monitoring efficiency evaluation index of the specified monitoring area according to the preset value of the threat alarm monitoring efficiency evaluation index, and send a signal to the human-computer interaction module according to the judgment result; Human-computer interaction module: used to receive the threat alarm warning instructions transmitted by the threat alarm monitoring and judgment module, and display the threat alarm warning instructions on the terminal display screen, and at the same time make corresponding warning reminders according to the received threat alarm warning instructions.

2. A power grid security monitoring system based on threat warning semantic knowledge mining and transfer learning according to claim 1, characterized in that: The power stability data acquisition unit in the power grid safety data acquisition module is used to collect power stability data of each sub-monitoring area, including voltage stability data, current stability data, and power stability data. The voltage stability data includes voltage effective value, voltage harmonic content, fluctuating voltage and voltage imbalance, which are respectively recorded as Uc, Uh, Uf and Ue. The current stability data includes current effective value, current harmonic content, fluctuating current and current imbalance, which are respectively recorded as Ic, Ih, If and Ie; the power stability data includes steady-state power after the accident, fault impact power and static stable maximum power after the accident, which are respectively recorded as Ps, Pb and Pm; the environmental safety data acquisition unit is used to collect environmental safety data of each monitoring sub-area, including ambient temperature data and ambient humidity data. The ambient temperature data includes insulation temperature and temperature gradient change rate, which are respectively recorded as Tc and Tl; the internal environment data includes relative humidity and non-condensing humidity change rate, which are respectively recorded as Hc and Hl.

3. A power grid security monitoring system based on threat warning semantic knowledge mining and transfer learning according to claim 1, characterized in that: The environmental safety preprocessing unit in the power grid safety data preprocessing module is used to establish an environmental safety preprocessing model, import the data transmitted in the power grid safety data acquisition module into the environmental safety preprocessing model, and obtain the environmental temperature stability coefficient and the environmental humidity stability coefficient. The specific analysis formula of the environmental temperature stability coefficient is: Medium Tr i represents the ambient temperature stability coefficient of the ith monitoring sub-area, Tc i represents the adiabatic temperature of the ith monitoring sub-area, Tl i represents the temperature gradient change rate of the i-th monitoring sub-area; The specific analysis formula of the environmental humidity stability coefficient is: Among them, Hr i represents the environmental humidity stability coefficient of the ith monitoring sub-area, Hc i represents the relative humidity of the ith monitoring sub-area, Hl i Represents the non-condensing humidity change rate of the i-th monitoring sub-area.

4. A power grid security monitoring system based on threat warning semantic knowledge mining and transfer learning according to claim 1, characterized in that: The environmental safety analysis unit is used to establish an environmental safety analysis model, import the data transmitted in the power stability preprocessing module into the environmental safety analysis model, and obtain the environmental safety state proportional coefficient. The specific calculation formula is: i =Tr i ×Hr i , where ψ i represents the environmental safety state proportional coefficient of the i-th sub-area, Tr i represents the ambient temperature stability coefficient of the ith monitoring sub-area, Hr i Represents the environmental humidity stability coefficient of the i-th monitoring sub-area.

5. The power grid security monitoring system based on threat warning semantic knowledge mining and transfer learning according to claim 1 is characterized by: The power grid security assessment index calculation module obtains the power grid security assessment index by establishing a power grid security assessment index calculation model. The specific analysis formula of the power grid security assessment index is: Y i =log2(1+φ i )+log3(1+ψ i ), where Y i represents the power grid security assessment index of the ith monitoring sub-area, φ i represents the power stability proportionality coefficient of the ith sub-region, ψ i Represents the environmental safety status proportional coefficient of the i-th sub-area.

6. A power grid security monitoring system based on threat warning semantic knowledge mining and transfer learning according to claim 1, characterized in that: The power grid security monitoring and judgment module compares the power grid security assessment index corresponding to each monitoring sub-area in the monitoring network with the preset standard power grid security assessment index corresponding to each monitoring sub-area. If the power grid security assessment index corresponding to a monitoring sub-area is greater than or equal to the preset standard power grid security assessment index of normal working status, it indicates that the monitoring sub-area is in normal working status. Otherwise, it indicates that the monitoring sub-area is in abnormal working status. The monitoring sub-areas in abnormal working status and the monitoring sub-areas in normal working status are screened and counted according to the working status of the monitoring sub-areas. The data collection and analysis of the monitoring sub-areas in normal working status are maintained. The monitoring sub-areas in abnormal working status are recorded as designated monitoring sub-areas, and they are numbered 1, 2, ..., j, ..., k in sequence.

7. A power grid security monitoring system based on threat warning semantic knowledge mining and transfer learning according to claim 1, characterized in that: The threat alarm data processing model is based on each monitoring sub-area in an abnormal state, collects historical threat alarm data through the power grid security monitoring system, cleans the collected data, and uses the obtained data for defense measures through a cluster analysis method, and obtains the threat alarm recognition rate, emergency response speed and threat alarm protection rate of each designated monitoring sub-area, which are recorded as Er, Et and Ed respectively; the threat alarm data analysis model is used to analyze the results after system identification and protection, and obtain the threat alarm monitoring efficiency evaluation index, and its specific analysis formula is: Where η represents the threat monitoring efficiency evaluation index of the specified monitoring area, Er j represents the threat alarm recognition rate of the jth specified monitoring sub-area, Ed j represents the threat alarm protection rate of the jth designated monitoring sub-area, Et j Represents the emergency response speed of the jth designated monitoring sub-area.

8. The power grid security monitoring system based on threat warning semantic knowledge mining and transfer learning according to claim 1 is characterized by: The threat alarm monitoring judgment module compares the threat alarm monitoring efficiency evaluation index of the specified monitoring area with the preset value η of the threat alarm monitoring efficiency evaluation index. D For comparison, if η D ≤η, indicating that the designated monitoring area has good threat warning processing capabilities. D >η, indicating that the designated monitoring area has poor threat warning processing capability.

Images