Smart grid global data monitoring and management system and method based on digital twin

Abstract

The present invention discloses a smart grid global data monitoring and management system and method based on digital twins, which relates to the field of power grid data technology; it includes a user data classification module, a historical abnormal event capture module, a characteristic response model construction module, a priority sorting module and a response encryption module; the user data classification module is used to classify user data based on power grid operation data; the historical abnormal event capture module is used to capture historical abnormal events that trigger responses in a digital twin model, and the characteristic response model construction module is used to analyze the characteristic response model of each type of user data corresponding to each historical abnormal event; the priority sorting module is used to prioritize user data of corresponding types using characteristic response frequencies; the response encryption module is used to encrypt user data that meets the characteristic response model in priority order based on the priority of the user data.

Description

Technical Field

[0001] The present invention relates to the field of power grid data technology, and in particular to a smart grid global data monitoring and management system and method based on digital twins. Background Art

[0002] In smart grids, data comes from a wide range of sources, including various sensors, smart meters, and monitoring systems, as well as various electricity consumption data from users, such as electricity consumption and power consumption. Smart grids involve a large amount of sensitive information, such as users' electricity consumption habits and grid operation parameters. Once leaked, these data may be used for malicious purposes, such as power theft or attacks on the grid. For example, hackers can obtain information about users' electricity consumption patterns and infer whether users are at home, thereby committing theft. As data is frequently transmitted between digital twin models and physical grids and between different systems, it is crucial to ensure data encryption and access control. For example, the symmetric encryption algorithms and asymmetric encryption algorithms updated in existing technologies have achieved a certain degree of security; but at the same time, due to the increased demand for encryption, a large amount of electricity consumption data in the entire domain will bring a lot of workload to the grid system, which is prone to data transmission and execution delays. Therefore, how to effectively capture high-demand encrypted data and reduce encryption load in a large amount of grid data is worth discussing. Summary of the invention

[0003] The purpose of the present invention is to provide a smart grid global data monitoring and management system and method based on digital twins to solve the problems raised in the prior art.

[0004] To achieve the above object, the present invention provides the following technical solution: a method for monitoring and managing global data of a smart grid based on digital twins, the method comprising the following:

[0005] Step S100: extracting user data stored and recorded in a digital twin model built with the smart grid as the main body, and classifying the user data based on the grid operation data;

[0006] Step S200: capturing historical abnormal events that trigger responses in the digital twin model. Historical abnormal events refer to events that pose a security threat to user electricity consumption data. Based on historical abnormal events, a characteristic response model corresponding to each type of user data and each historical abnormal event is analyzed;

[0007] Step S300: Analyze the characteristic response frequency of user data corresponding to historical abnormal events, and prioritize the corresponding types of user data using the characteristic response frequency;

[0008] Step S400: real-time monitoring of user data transmitted during power grid operation and real-time triggering of a characteristic response model for judgment as user data increases, and encrypting user data that meets the characteristic response model in order of priority based on the priority of the user data.

[0009] Furthermore, step S100 includes the following specific steps:

[0010] Step S110: The power grid operation data refers to the operation parameters recorded by various devices installed in the power grid; taking each operation parameter as a search item, traverse and search all user data transmitted to the power grid control center, and preliminarily classify the user data of the same mathematical dimension into the same type of user data;

[0011] Step S120: Based on each search item, matching operation parameter records of similar user data under the same digital twin model storage power consumption event to form a matching data group A, A=(a1,a2), where a1 represents the search item, and a2 represents the target item composed of similar user data;

[0012] When a2 records the type of user data of the same type but is not unique, a2={a 21 、a 22 ,......,a 2n}, a 21 、a 22 ,......,a 2n represents the 1st, 2nd, ..., nth types of user data recorded in the user event under the corresponding search item a1; then the matching data group is updated to any free combination of all types of user data recorded in a2 as the target item in the matching data group A, and several matching data groups containing the same search item but different target items are generated as the target data group;

[0013] Step S130: Based on each target data group, a linear regression model of the search item and the target item is constructed. When the number of target items in the linear regression model is one, user data of the target item corresponding to the regression coefficient greater than a preset threshold is extracted as one type of target user data; when the number of target items in the linear regression model is greater than one, all types of user data in the same linear regression model when the regression coefficient is greater than the preset threshold and the similarity of the regression coefficients corresponding to the target items is greater than the similarity threshold are extracted as one type of target user data.

[0014] Through the correlation analysis of user data based on power grid operation data, user data classification can be maximized and the analysis of the impact correlation between data can be enhanced; in the data monitoring process, the data monitoring efficiency can be maximized on the basis of ensuring the data scope.

[0015] Furthermore, step S200 includes the following specific steps:

[0016] Step S210: Taking the interval period between two adjacent abnormal events with the same abnormal object recorded in the digital twin model as the unit monitoring period, extracting the target user data of each type under all the unit monitoring periods in the historical records, and drawing a line graph of the target user data of each type in each unit monitoring period in units of days, wherein the line graph is composed of days as the horizontal axis and the mathematical unit of the target user data as the vertical axis;

[0017] Step S220: Based on the line graph, use the formula:

[0018] g1={m / [(m-1)(m-2)]}×{∑[(x j -x0) / d] 3};

[0019] Calculate the deviation g1 of each type of target user data in each unit monitoring period, where m represents the number of target user data obtained in the line graph, x j represents the jth target user data in a unit monitoring period in the line graph, x0 represents the average value of all user data recorded with the same type of target data, d represents the standard deviation of all user data recorded with the same type of target data, when the number of the same type of user data types contained in the same type of target user data is not 1, the deviation corresponding to each type of user data is calculated respectively and the average value is calculated as the deviation of the target user data of this type;

[0020] Step S230: Filter the target user data whose absolute value of deviation is less than or equal to the deviation threshold as the analysis object, and extract the deviation g1 calculated after generating the line graph of the same type of analysis object in all unit monitoring periods, and calculate the dispersion coefficient p of the deviation of the corresponding type of target user data using the formula: p={[∑(g1-g0) 2 ] / w} 1 / 2 , where g0 represents the average deviation of all unit monitoring periods of historical records, and w represents the total number of unit monitoring periods of historical records;

[0021] Step S240: When the target user data whose dispersion coefficient is less than the dispersion coefficient threshold and the monitored abnormal object of the historical abnormal event is the corresponding type, the target user data is used as the key user data; the deviation g1 of the key user data records in all unit monitoring periods is obtained, and the characteristic response model Y of the corresponding type of key user data is constructed, Y∈[g1min,g1max], where g1min represents the minimum deviation and g1max represents the maximum deviation.

[0022] The analysis feature response model can effectively discover user electricity consumption data that poses certain security risks among numerous power grid data, and these data that cause security risks are often captured and analyzed by intruders based on the patterns generated by the data themselves, resulting in a certain degree of risk infringement; therefore, this application starts from the perspective of the intruder and implements a first-level data security monitoring from the perspective of electricity consumption data that is easy to capture patterns, thereby improving the efficiency and security of global data security monitoring.

[0023] Furthermore, prioritizing corresponding types of user data by using the response frequency includes the following specific steps:

[0024] Extract historical abnormal events when the monitored abnormal object is key user data, obtain the number of historical abnormal events C1 and the maximum span period L1 of the corresponding records corresponding to the same key user data, and use the formula: Z=C1 / L1;

[0025] Calculate the response frequency Z corresponding to each key user data;

[0026] Based on the numerical value of the response frequency Z, the key user data of each type corresponding to the response frequency is sorted in order from large to small to generate a response priority.

[0027] Further, step S400 includes the following:

[0028] When the user data increases to the point where it can be drawn into a line graph, the line graph contains at least f matching data groups, where f>2; the deviation of the corresponding type of user data is calculated based on the line graph and substituted into the corresponding feature response model; if the calculated real-time deviation g satisfies g∈[g1min,g1max], the response is triggered to obtain all current user data that satisfies the corresponding feature response model, and each user data is sequentially encrypted based on priority sorting.

[0029] Selecting the target type of encrypted data can effectively save the time and efficiency analysis of capturing data that is more in need of security protection in the massive global data of the power grid; reduce the pressure of data encryption and conduct real-time monitoring during the dynamic data flow, adaptively adjust the encryption requirements of different power grid data generated at different stages, realize intelligent global data monitoring of the power grid, and when it is necessary to encrypt the electricity consumption data of multiple types of users within a period, the reasonable and precise calculation of the encryption priority greatly improves the rationality of the encryption algorithm deployment and the optimization of the data security level.

[0030] The smart grid global data monitoring and management system based on digital twins includes a user data classification module, a historical abnormal event capture module, a feature response model construction module, a priority sorting module, and a response encryption module;

[0031] The user data classification module is used to extract the user data stored and recorded in the digital twin model built with the smart grid as the main body, and classify the user data based on the grid operation data;

[0032] The historical abnormal event capture module is used to capture historical abnormal events that trigger responses in the digital twin model.

[0033] The feature response model building module is used to analyze the feature response model of each type of user data corresponding to each historical abnormal event;

[0034] The priority sorting module is used to prioritize the corresponding types of user data using the characteristic response frequency;

[0035] The response encryption module monitors the user data transmitted during the operation of the power grid in real time and triggers the characteristic response model in real time to make judgments as the user data increases, and encrypts the user data that meets the characteristic response model in priority order based on the priority of the user data.

[0036] Further, the user data classification module includes a user data preliminary division unit, a matching data group generation unit, a target data group determination unit and a linear regression model analysis unit;

[0037] The user data preliminary division unit is used to preliminary divide the user data of the same mathematical dimension into the same type of user data;

[0038] The matching data group generating unit is used to match the operation parameter records of the same type of user data under the power consumption event stored in the same digital twin model based on each search item to form a matching data group;

[0039] The target data group determination unit is used to generate a plurality of matching data groups containing the same search term but different target terms as the target data group;

[0040] The linear regression model analysis unit is used to construct a linear regression model of the search terms and the target terms. When the number of target terms in the linear regression model is one, user data of the target terms corresponding to the regression coefficients greater than a preset threshold is extracted as one type of target user data; when the number of target terms in the linear regression model is greater than one, all types of user data in the same linear regression model when the regression coefficients are greater than the preset threshold and the similarity of the regression coefficients corresponding to the target terms is greater than the similarity threshold are extracted as one type of target user data.

[0041] Furthermore, the characteristic response model construction module includes a line graph drawing unit, a deviation calculation unit, a discrete coefficient analysis unit and a characteristic response model generation unit;

[0042] The line graph drawing unit is used to extract the target user data of each type in all unit monitoring periods of the historical records, and draw a line graph recording the target user data of each type in each unit monitoring period in units of days;

[0043] The deviation calculation unit is used to calculate the deviation of each type of target user data within each unit monitoring period;

[0044] The dispersion coefficient analysis unit is used to calculate the dispersion coefficient of the deviation degree of the corresponding type of target user data based on the deviation degree;

[0045] The characteristic response model generation unit is used to obtain the deviation of key user data records within all unit monitoring periods and construct a characteristic response model of key user data of corresponding types.

[0046] Further, the priority ranking module includes a response frequency calculation unit and a ranking generation unit;

[0047] The response frequency calculation unit is used to calculate the response frequency corresponding to each key user data;

[0048] The sorting generation unit is used to sort the key user data of each type corresponding to the response frequency in order from large to small based on the numerical value of the response frequency to generate a response priority.

[0049] Compared with the prior art, the present invention has the following beneficial effects:

[0050] 1. The present invention classifies user data based on grid operation characteristics instead of simply dividing based on dimensions. Instead, it analyzes the correlation of user power consumption data from the source of data generation and the influencing party, so as to maximize the data monitoring efficiency on the basis of ensuring the data range during the data monitoring process;

[0051] 2. Secondly, this application constructs a characteristic response model to achieve a higher level of data security monitoring from the perspective of electricity consumption data that is easy to capture regularities, thereby improving the efficiency and security of global data security monitoring.

[0052] 3. This application encrypts user data that meets the response model, reduces the pressure of data encryption, and conducts real-time monitoring during dynamic data flow. It adaptively adjusts the encryption requirements of different power grid data generated at different stages, realizes intelligent power grid full-domain data monitoring, and when it is necessary to encrypt the electricity consumption data of multiple types of users within a period, the encryption priority is reasonably and accurately calculated, which greatly improves the rationality of the encryption algorithm deployment and the optimization of the data security level. BRIEF DESCRIPTION OF THE DRAWINGS

[0053] Figure 1 It is a structural schematic diagram of the smart grid global data monitoring and management system based on digital twins of the present invention. DETAILED DESCRIPTION

[0054] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions 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 are within the scope of protection of the present invention.

[0055] Example: Figure 1 As shown, the present invention provides a smart grid global data monitoring and management system and method technical solution based on digital twins, and a smart grid global data monitoring and management method based on digital twins, the method includes the following:

[0056] Step S100: extracting user data stored and recorded in a digital twin model built with the smart grid as the main body, and classifying the user data based on the grid operation data;

[0057] Step S200: capturing historical abnormal events that trigger responses in the digital twin model. Historical abnormal events refer to events that pose a security threat to user electricity consumption data. Based on historical abnormal events, a characteristic response model corresponding to each type of user data and each historical abnormal event is analyzed;

[0058] Step S300: Analyze the characteristic response frequency of user data corresponding to historical abnormal events, and prioritize the corresponding types of user data using the characteristic response frequency;

[0059] Step S400: real-time monitoring of user data transmitted during power grid operation and real-time triggering of a characteristic response model for judgment as user data increases, and encrypting user data that meets the characteristic response model in order of priority based on the priority of the user data.

[0060] Step S100 includes the following specific steps:

[0061] Step S110: The power grid operation data refers to the operation parameters recorded by various devices installed in the power grid; taking each operation parameter as a search item, traverse and search all user data transmitted to the power grid control center, and preliminarily classify the user data of the same mathematical dimension into the same type of user data;

[0062] Step S120: Based on each search item, matching operation parameter records of similar user data under the same digital twin model storage power consumption event to form a matching data group A, A=(a1,a2), where a1 represents the search item, and a2 represents the target item composed of similar user data;

[0063] When a2 records the type of user data of the same type but is not unique, a2={a21 、a 22 ,......,a 2n}, a 21 、a 22 ,......,a 2n represents the 1st, 2nd, ..., nth types of user data recorded in the user event under the corresponding search item a1; then the matching data group is updated to any free combination of all types of user data recorded in a2 as the target item in the matching data group A, and several matching data groups containing the same search item but different target items are generated as the target data group;

[0064] As shown in the examples:

[0065] When the search item is the device output power and the target item is the user's power consumption, power consumption and power consumption time, the matching data group is updated as follows:

[0066] A=(output power, user power consumption), A=(output power, power consumption), A=(output power, power consumption time);

[0067] A=[output power, (user power consumption, power consumption)], A=[output power, (power consumption time, user power consumption)], A=[(output power, (power consumption time, power consumption)];

[0068] The above is taken as the new target data set;

[0069] Step S130: Based on each target data group, a linear regression model of the search item and the target item is constructed. When the number of target items in the linear regression model is one, user data of the target item corresponding to the regression coefficient greater than a preset threshold is extracted as one type of target user data; when the number of target items in the linear regression model is greater than one, all types of user data in the same linear regression model when the regression coefficient is greater than the preset threshold and the similarity of the regression coefficients corresponding to the target items is greater than the similarity threshold are extracted as one type of target user data.

[0070] Through the correlation analysis of user data based on power grid operation data, user data classification can be maximized and the analysis of the impact correlation between data can be enhanced; in the data monitoring process, the data monitoring efficiency can be maximized on the basis of ensuring the data scope.

[0071] Step S200 includes the following specific steps:

[0072] Step S210: Taking the interval period between two adjacent abnormal events with the same abnormal object recorded in the digital twin model as the unit monitoring period, extracting the target user data of each type under all the unit monitoring periods in the historical records, and drawing a line graph of the target user data of each type in each unit monitoring period in units of days, wherein the line graph is composed of days as the horizontal axis and the mathematical unit of the target user data as the vertical axis;

[0073] Step S220: Based on the line graph, use the formula:

[0074] g1={m / [(m-1)(m-2)]}×{∑[(x j -x0) / d] 3};

[0075] Calculate the deviation g1 of each type of target user data in each unit monitoring period, where m represents the number of target user data obtained in the line graph, x j represents the jth target user data in the unit monitoring period in the line graph, x0 represents the average value of all user data recorded by the same type of target data, d represents the standard deviation of all user data recorded by the same type of target data, when the number of the same type of user data types contained in the same type of target user data is not 1, the deviation corresponding to each type of user data is calculated respectively and the average value is calculated as the deviation of the target user data of this type; ∑[(x j -x0) / d] 3 It refers to the [(x j -x0) / d] 3 sum;

[0076] As shown in the examples:

[0077] If there is an obvious regularity in the electricity usage time data of a certain user, then the deviation calculated corresponding to the electricity usage time in the unit monitoring cycle will be less than the deviation threshold, indicating that the user's electricity usage time in this cycle presents a regular distribution;

[0078] And zoom in to all monitoring periods to analyze the degree of dispersion. When the degree of dispersion is small, it means that the regular distribution still maintains with the change of time period;

[0079] In addition, in historical abnormal events, there are safety issues caused by the obvious characteristics of electricity usage time. For example, electricity usage time can be used to infer whether the user is at home, whether the user is using certain electrical appliances, etc., and some illegal acts can be committed based on this information; then this type of data can be used as the key user data to be analyzed in this application.

[0080] Step S230: Filter the target user data whose absolute value of deviation is less than or equal to the deviation threshold as the analysis object, and extract the deviation g1 calculated after generating the line graph of the same type of analysis object in all unit monitoring periods, and calculate the dispersion coefficient p of the deviation of the corresponding type of target user data using the formula: p={[∑(g1-g0) 2 ] / w} 1 / 2 , where g0 represents the average deviation of all unit monitoring periods of historical records, and w represents the total number of unit monitoring periods of historical records;

[0081] Step S240: When the target user data whose dispersion coefficient is less than the dispersion coefficient threshold and the monitored abnormal object of the historical abnormal event is the corresponding type, the target user data is used as the key user data; the deviation g1 of the key user data records in all unit monitoring periods is obtained, and the characteristic response model Y of the corresponding type of key user data is constructed, Y∈[g1min,g1max], where g1min represents the minimum deviation and g1max represents the maximum deviation.

[0082] The analysis feature response model can effectively discover user electricity consumption data that poses certain security risks among numerous power grid data, and these data that cause security risks are often captured and analyzed by intruders based on the patterns generated by the data themselves, resulting in a certain degree of risk infringement; therefore, this application starts from the perspective of the intruder and implements a first-level data security monitoring from the perspective of electricity consumption data that is easy to capture patterns, thereby improving the efficiency and security of global data security monitoring.

[0083] Using response frequency to prioritize corresponding types of user data includes the following specific steps:

[0084] Extract historical abnormal events when the monitored abnormal object is key user data, obtain the number of historical abnormal events C1 and the maximum span period L1 of the corresponding records corresponding to the same key user data, and use the formula: Z=C1 / L1;

[0085] Calculate the response frequency Z corresponding to each key user data;

[0086] Based on the numerical value of the response frequency Z, the key user data of each type corresponding to the response frequency is sorted in order from large to small to generate a response priority.

[0087] Step S400 includes the following:

[0088] When the user data increases to the point where it can be drawn into a line graph, the line graph contains at least f matching data groups, where f>2; the deviation of the corresponding type of user data is calculated based on the line graph and substituted into the corresponding feature response model; if the calculated real-time deviation g satisfies g∈[g1min,g1max], the response is triggered to obtain all current user data that satisfies the corresponding feature response model, and each user data is sequentially encrypted based on priority sorting.

[0089] As shown in the embodiment: if the deviations calculated for "power consumption" and "power consumption" in the line graph both satisfy the corresponding characteristic response model, and the priority of "power consumption" is greater than that of "power consumption", the power consumption data output by the user is encrypted first, and then the power consumption data is encrypted, and other power consumption data such as power consumption equipment may not be encrypted. In this application, the encryption method can be based on system adaptive selection or matching.

[0090] Selecting the target type of encrypted data can effectively save the time and efficiency analysis of capturing data that is more in need of security protection in the massive global data of the power grid; reduce the pressure of data encryption and conduct real-time monitoring during the dynamic data flow, adaptively adjust the encryption requirements of different power grid data generated at different stages, realize intelligent global data monitoring of the power grid, and when it is necessary to encrypt the electricity consumption data of multiple types of users within a period, the reasonable and precise calculation of the encryption priority greatly improves the rationality of the encryption algorithm deployment and the optimization of the data security level.

[0091] The smart grid global data monitoring and management system based on digital twins includes a user data classification module, a historical abnormal event capture module, a feature response model construction module, a priority sorting module, and a response encryption module;

[0092] The user data classification module is used to extract the user data stored and recorded in the digital twin model built with the smart grid as the main body, and classify the user data based on the grid operation data;

[0093] The historical abnormal event capture module is used to capture historical abnormal events that trigger responses in the digital twin model.

[0094] The feature response model building module is used to analyze the feature response model of each type of user data corresponding to each historical abnormal event;

[0095] The priority sorting module is used to prioritize the corresponding types of user data using the characteristic response frequency;

[0096] The response encryption module monitors the user data transmitted during the operation of the power grid in real time and triggers the characteristic response model in real time to make judgments as the user data increases, and encrypts the user data that meets the characteristic response model in priority order based on the priority of the user data.

[0097] The user data classification module includes a user data preliminary division unit, a matching data group generation unit, a target data group determination unit and a linear regression model analysis unit;

[0098] The user data preliminary division unit is used to preliminary divide the user data of the same mathematical dimension into the same type of user data;

[0099] The matching data group generating unit is used to match the operation parameter records of the same type of user data under the power consumption event stored in the same digital twin model based on each search item to form a matching data group;

[0100] The target data group determination unit is used to generate a plurality of matching data groups containing the same search term but different target terms as the target data group;

[0101] The linear regression model analysis unit is used to construct a linear regression model of the search terms and the target terms. When the number of target terms in the linear regression model is one, user data of the target terms corresponding to the regression coefficients greater than a preset threshold is extracted as one type of target user data; when the number of target terms in the linear regression model is greater than one, all types of user data in the same linear regression model when the regression coefficients are greater than the preset threshold and the similarity of the regression coefficients corresponding to the target terms is greater than the similarity threshold are extracted as one type of target user data.

[0102] The characteristic response model construction module includes a line graph drawing unit, a deviation calculation unit, a discrete coefficient analysis unit and a characteristic response model generation unit;

[0103] The line graph drawing unit is used to extract the target user data of each type in all unit monitoring periods of the historical records, and draw a line graph recording the target user data of each type in each unit monitoring period in units of days;

[0104] The deviation calculation unit is used to calculate the deviation of each type of target user data within each unit monitoring period;

[0105] The dispersion coefficient analysis unit is used to calculate the dispersion coefficient of the deviation degree of the corresponding type of target user data based on the deviation degree;

[0106] The characteristic response model generation unit is used to obtain the deviation of key user data records within all unit monitoring periods and construct a characteristic response model of key user data of corresponding types.

[0107] The priority ranking module includes a response frequency calculation unit and a ranking generation unit;

[0108] The response frequency calculation unit is used to calculate the response frequency corresponding to each key user data;

[0109] The sorting generation unit is used to sort the key user data of each type corresponding to the response frequency in order from large to small based on the numerical value of the response frequency to generate a response priority.

[0110] It will be apparent to those skilled in the art that the invention is not limited to the details of the exemplary embodiments described above and that the invention can be implemented in other specific forms without departing from the spirit or essential features of the invention. Therefore, the embodiments should be considered exemplary and non-limiting in all respects, and the scope of the invention is defined by the appended claims rather than the foregoing description, and it is intended that all variations falling within the meaning and scope of the equivalent elements of the claims be included in the invention. Any reference numeral in a claim should not be considered as limiting the claim to which it relates.

Claims

1. A smart grid global data monitoring and management method based on digital twins, characterized by: The method comprises the following: Step S100: extracting user data stored and recorded in a digital twin model built with the smart grid as the main body, and classifying the user data based on the grid operation data; The step S100 includes the following specific steps: Step S110: The power grid operation data refers to the operation parameters recorded by various devices installed in the power grid; taking each operation parameter as a search item, traverse and search all user data transmitted to the power grid control center, and preliminarily classify user data of the same mathematical dimension into the same type of user data; Step S120: Based on each search item, matching operation parameter records of similar user data under the same digital twin model storage power consumption event to form a matching data group A, A=(a1,a2), where a1 represents the search item, and a2 represents the target item composed of similar user data; When a2 records the type of user data of the same type but is not unique, a2={a 21 、a 22 ,......,a 2n }, a 21 、a 22 ,......,a 2n represents the 1st, 2nd, ..., nth types of user data recorded in the user event under the corresponding search item a1; then the matching data group is updated to any free combination of all types of user data recorded in a2 as the target item in the matching data group A, and several matching data groups containing the same search item but different target items are generated as the target data group; Step S130: Based on each target data group, a linear regression model of the search item and the target item is constructed. When the number of target items in the linear regression model is one, user data of the target item corresponding to the regression coefficient greater than a preset threshold is extracted as one type of target user data; when the number of target items in the linear regression model is greater than one, all types of user data in the same linear regression model whose regression coefficient is greater than the preset threshold and whose similarity of the regression coefficients corresponding to the target items is greater than the similarity threshold are extracted as one type of target user data; Step S200: capturing historical abnormal events that trigger responses in the digital twin model, where the historical abnormal events refer to events that pose a security threat to the user's electricity consumption data, and analyzing the characteristic response model of each type of user data corresponding to each historical abnormal event based on the historical abnormal events; The step S200 includes the following specific steps: Step S210: Taking the interval period between two adjacent abnormal events with the same abnormal object recorded in the digital twin model as the unit monitoring period, extracting the target user data of each type under all the unit monitoring periods in the historical records, and drawing a line graph of the target user data of each type in each unit monitoring period in units of days, wherein the line graph is composed of days as the horizontal axis and the mathematical unit of the target user data as the vertical axis; Step S220: Based on the line graph, use the formula: g1={m / [(m-1)(m-2)]}×{∑[(x j -x0) / d] 3 }; Calculate the deviation g1 of each type of target user data in each unit monitoring period, where m represents the number of target user data obtained in the line graph, x j represents the jth target user data in a unit monitoring period in the line graph, x0 represents the average value of all user data recorded with the same type of target data, d represents the standard deviation of all user data recorded with the same type of target data, when the number of the same type of user data types contained in the same type of target user data is not 1, the deviation corresponding to each type of user data is calculated respectively and the average value is calculated as the deviation of the target user data of this type; Step S230: Filter the target user data whose absolute value of deviation is less than or equal to the deviation threshold as the analysis object, and extract the deviation g1 calculated after generating the line graph of the same type of analysis object in all unit monitoring periods, and calculate the dispersion coefficient p of the deviation of the corresponding type of target user data using the formula: p={[∑(g1-g0) 2 ] / w} 1 / 2 , where g0 represents the average deviation of all unit monitoring periods of historical records, and w represents the total number of unit monitoring periods of historical records; Step S240: When the target user data whose dispersion coefficient is less than the dispersion coefficient threshold and the monitored abnormal object of the historical abnormal event is the corresponding type is selected, the target user data is used as the key user data; the deviation g1 of the key user data records in all unit monitoring periods is obtained, and the characteristic response model Y of the corresponding type of key user data is constructed, Y∈[g1min,g1max], where g1min represents the minimum deviation and g1max represents the maximum deviation; Step S300: Analyze the characteristic response frequency of user data corresponding to historical abnormal events, and prioritize the corresponding types of user data using the characteristic response frequency; Step S400: real-time monitoring of user data transmitted during power grid operation and real-time triggering of a characteristic response model for judgment as user data increases, and encrypting user data that meets the characteristic response model in order of priority based on the priority of the user data.

2. The method for monitoring and managing global data of a smart grid based on digital twins according to claim 1 is characterized in that: The prioritization of corresponding types of user data by using characteristic response frequency comprises the following specific steps: Extract historical abnormal events when the monitored abnormal object is key user data, obtain the number of historical abnormal events C1 and the maximum span period L1 of the corresponding records corresponding to the same key user data, and use the formula: Z=C1 / L1; Calculate the response frequency Z corresponding to each key user data; Based on the numerical value of the response frequency Z, the key user data of each type corresponding to the response frequency is sorted in order from large to small to generate a response priority.

3. The method for monitoring and managing global data of a smart grid based on digital twins according to claim 2 is characterized in that: The step S400 includes the following: When the user data increases to the point where it can be drawn into a line graph, the line graph contains at least f matching data groups, where f>2; the deviation of the corresponding type of user data is calculated based on the line graph and substituted into the corresponding feature response model; if the calculated real-time deviation g satisfies g∈[g1min,g1max], a response is triggered to obtain all current user data that satisfies the corresponding feature response model, and each user data is sequentially encrypted based on priority sorting.

4. A smart grid global data monitoring and management system based on digital twins, such as implementing the smart grid global data monitoring and management method based on digital twins as described in any one of claims 1 to 3, characterized in that: The system includes a user data classification module, a historical abnormal event capture module, a feature response model construction module, a priority sorting module and a response encryption module; The user data classification module is used to extract user data stored and recorded in the digital twin model built with the smart grid as the main body, and classify the user data based on the grid operation data; The historical abnormal event capture module is used to capture the historical abnormal events that trigger responses in the digital twin model. The feature response model building module is used to analyze the feature response model of each type of user data corresponding to each historical abnormal event; The priority sorting module is used to prioritize the corresponding types of user data using the characteristic response frequency; The response encryption module monitors the user data transmitted during the operation of the power grid in real time and triggers the characteristic response model to make judgments in real time as the user data increases, and encrypts the user data that meets the characteristic response model in priority order based on the priority of the user data.

5. The digital twin-based smart grid global data monitoring and management system according to claim 4 is characterized by: The user data classification module includes a user data preliminary division unit, a matching data group generation unit, a target data group determination unit and a linear regression model analysis unit; The user data preliminary division unit is used to preliminary divide the user data of the same mathematical dimension into the same type of user data; The matching data group generating unit is used to match the operating parameter records of the same type of user data under the power usage event stored in the same digital twin model based on each search item to form a matching data group; The target data group determination unit is used to generate a plurality of matching data groups containing the same search term but different target terms as the target data group; The linear regression model analysis unit is used to construct a linear regression model of the search terms and the target terms. When the number of the target terms in the linear regression model is one, user data of the target terms corresponding to the regression coefficients greater than a preset threshold is extracted as a type of target user data. When the number of target items in the linear regression model is greater than one, all types of user data in the same linear regression model whose regression coefficient is greater than a preset threshold and whose similarity of the regression coefficients corresponding to the target items is greater than the similarity threshold are extracted as one type of target user data.

6. The digital twin-based smart grid global data monitoring and management system according to claim 5 is characterized by: The characteristic response model construction module includes a line graph drawing unit, a deviation calculation unit, a discrete coefficient analysis unit and a characteristic response model generation unit; The line graph drawing unit is used to extract the target user data of each type in all unit monitoring periods of the historical records, and draw a line graph recording the target user data of each type in each unit monitoring period in units of days; The deviation calculation unit is used to calculate the deviation of each type of target user data in each unit monitoring period; The dispersion coefficient analysis unit is used to calculate the dispersion coefficient of the deviation degree of the corresponding type of target user data based on the deviation degree; The characteristic response model generating unit is used to obtain the deviation of the key user data records in all unit monitoring periods, and to construct a characteristic response model of the key user data of the corresponding type.

7. The digital twin-based smart grid global data monitoring and management system according to claim 6 is characterized by: The priority ranking module includes a response frequency calculation unit and a ranking generation unit; The response frequency calculation unit is used to calculate the response frequency corresponding to each key user data; The ranking generating unit is used to generate a response priority by ranking the key user data of each type corresponding to the response frequency in a descending order based on the numerical value of the response frequency.

Images