Electric power remote safety protection system

Abstract

The invention provides an electric power remote safety protection system, and belongs to the technical field of electric power protection, and the system comprises a monitoring module which is used for monitoring the data information of electric power terminal equipment in a power grid, and obtaining a terminal data set; the extraction module is used for performing feature extraction on the terminal data set; the analysis module is used for acquiring the detection instruction book from the test management platform and analyzing the detection instruction book to generate a detection object and a compliance index associated with the detection object; the establishing module is used for establishing a feature evaluation model based on a preset algorithm and in combination with the compliance index; and the analysis module is used for carrying out situation analysis on the extracted features according to the feature evaluation model, and judging whether an abnormal behavior warning is sent out or not according to a situation analysis result. The problem that a large amount of time and manpower are consumed because an existing protection system needs a person to judge accident reasons on site is solved.

Description

Technical Field

[0001] The present invention relates to the technical field of power protection, and particularly to a power remote security protection system. Background Art

[0002] At present, during the operation of the power grid, accidents always occur, such as short circuits in transmission lines, open circuits in transmission lines, transmission lines being struck by lightning, equipment tripping due to overloading, damage to power equipment caused by external forces, and equipment power outages caused by illegal operations. When a fault occurs in the power grid, whether in a substation, on a line, or at an equipment end, it will first cause a large-scale power outage, affecting the production of other industries and people's lives. Therefore, it has become a very necessary and urgent task to quickly determine the fault location and quickly direct the corresponding personnel to arrive at the scene with appropriate repair tools for fault repair and restore power supply. However, the existing protection systems require personnel to go to the scene to judge the cause of the accident, consuming a large amount of time and manpower.

[0003] Therefore, the present invention proposes a power remote security protection system. Summary of the Invention

[0004] The present invention provides a power remote security protection system. By monitoring the data information of power terminal equipment in the power grid, a terminal data set is obtained. Feature extraction is performed on the terminal data set. A test instruction manual is obtained from the test management platform, and the test instruction manual is parsed to generate test objects and compliance indicators associated with the test objects. Based on a preset algorithm and in combination with the compliance indicators, a feature evaluation model is established. According to the feature evaluation model, a situation analysis is performed on the extracted features, and based on the situation analysis result, it is determined whether to issue a warning for abnormal behavior, solving the problem in the background art that the existing protection systems require personnel to go to the scene to judge the cause of the accident, consuming a large amount of time and manpower.

[0005] The present invention proposes a power remote security protection system, which includes:

[0006] A monitoring module: used for monitoring the data information of power terminal equipment in the power grid to obtain a terminal data set;

[0007] An extraction module: used for performing feature extraction on the terminal data set;

[0008] An analysis module: obtains a test instruction manual from the test management platform and parses the test instruction manual to generate test objects and compliance indicators associated with the test objects;

[0009] A building module: used for establishing a feature evaluation model based on the compliance indicators;

[0010] Analysis module: Perform situation analysis on the extracted features according to the feature evaluation model, and determine whether to issue an abnormal behavior warning based on the situation analysis result.

[0011] Preferably, the monitoring module includes:

[0012] The first acquisition unit: Acquire the device attributes and device types of power terminal devices in the power grid;

[0013] The setting unit: Set the data acquisition period, data acquisition rate, and data items based on the device attributes and device types;

[0014] The acquisition unit: Acquire the data information of each data item in the power terminal device according to the data acquisition period and the data acquisition rate;

[0015] The integration unit: Integrate and process the data information of each data item of the power terminal device to obtain a terminal data set.

[0016] Preferably, the extraction module includes:

[0017] The screening unit: Screen the eigenvalues in the obtained terminal data set to obtain a basic eigenvalue set;

[0018] The first classification unit: Classify the eigenvalues under the same data item in the basic eigenvalue set to obtain a classification result under the same data item;

[0019] The first determination unit: Determine the eigenvalue distribution of the same data item in the terminal data set based on the classification result;

[0020] The extraction unit: Select the corresponding feature extraction method from the distribution-method mapping table according to the eigenvalue distribution of different data items, and perform feature extraction on all the data under the corresponding data item in the terminal data set.

[0021] Preferably, the first classification unit includes:

[0022] The acquisition subunit: Acquire multiple random eigenvalues in each data item from the basic eigenvalue set by a random sampling method;

[0023] The determination subunit: Determine the edit distance between multiple random eigenvalues under the same data item to obtain a distance similarity array for each random eigenvalue;

[0024] The classification subunit: Classify the eigenvalues under the same data item based on the distance similarity array to obtain a classification result.

[0025] Preferably, the parsing module includes:

[0026] Decryption unit: Obtain the inspection instruction manual from the test management platform through data acquisition permissions, and perform decryption processing on the inspection instruction manual;

[0027] Second acquisition unit: Acquire multiple inspection items of power terminals in the power grid;

[0028] Second determination unit: Determine the inspection object according to the decrypted inspection instruction manual and multiple inspection items of the power terminal;

[0029] Third determination unit: Analyze the inspection instruction manual to obtain the standard working parameters of each inspection object, and determine the compliance indicators associated with each inspection object according to the standard working parameters.

[0030] Preferably, the establishment module includes:

[0031] Construction unit: Obtain the characteristic data of each compliance indicator, and obtain the feedback evaluation of each characteristic data in the same compliance indicator, and construct the corresponding feature vector;

[0032] Establishment unit: Input different feature vectors into the neural network model in turn to establish a feature evaluation model.

[0033] Preferably, the analysis module includes:

[0034] Conversion unit: Convert the extracted features into standardized data according to the feature-conversion mapping table;

[0035] Input unit: Input the extracted features into the feature evaluation model to obtain an initial evaluation result;

[0036] Analysis unit: Perform feature analysis on the standardized data to obtain an optimization factor;

[0037] Judgment unit: Optimize the initial evaluation result based on the optimization factor, and judge whether to issue an abnormal behavior warning.

[0038] Preferably, it further includes:

[0039] Data determination module: Read the original power data collected from the power terminals in the power grid, determine the data type of the original power data, and obtain regenerated data;

[0040] Source matching module: Trace the data source of the regenerated data, and match the regenerated data with the data source one by one according to the source database to obtain classified data;

[0041] Dataset acquisition module: Sort the classified data of the same data source in chronological order, and correspond each time-point dataset with the load in the system one by one to obtain the target load dataset of the same data source;

[0042] Data clustering module: Obtain the operation characteristics of the same data source, and perform clustering processing on the corresponding target load data set based on the operation characteristics to obtain several operation data groups;

[0043] Match the initial value of each parameter in the same operation data group with the value mapping table to obtain the target value of the corresponding operation data group;

[0044] Lock the operation time points involved in the same operation data group, match the load values with the data sets at each operation time point, and draw the array load curve corresponding to the operation data group;

[0045] Curve comparison module: Based on the typical daily load curve, array load curve, and standard load curve under the same operation data group;

[0046] Calculate the first similarity S1 between the typical daily load curve y01 and the array load curve y0s under the same operation array, and the second similarity S2 between the standard load curve y0b and the array load curve y0s under the same operation array;

[0047] S1 = sim(y01, y0s);

[0048] S2 = sim(y0b, y0s);

[0049] Among them, sim() represents the similarity function;

[0050] Perform a first comparison based on the target value r01 of the typical load curve y01 and the standard value r02 of the standard load curve y0b. At the same time, perform a second comparison between the first similarity and the second similarity;

[0051] Determine the upper division boundary and the lower division boundary according to the first comparison result and the second comparison result;

[0052]

[0053]

[0054] Among them, represents the adjustment coefficient; F1 represents the value of the upper division boundary; F2 represents the value of the lower division boundary; F Z represents the set left range value; F Y represents the set right range value;

[0055] Lock the data outside the boundary for the corresponding array load curve based on the upper division boundary and the lower division boundary, and regard it as bad data;

[0056] Determine the current position of the bad data, obtain the values of the first data and the second data that are at the closest distance to the bad data, determine the replacement data corresponding to the bad data, and perform the replacement;

[0057]

[0058] Wherein, J represents the boundary value for determining that the data at the current position is bad data; T1 represents the value of the first data; T2 represents the value of the second data; T0 represents the value of the replacement data.

[0059] Other features and advantages of the present invention will be described in the subsequent description, and part of them will become obvious from the description, or will be understood by implementing the present invention. The objectives and other advantages of the present invention can be achieved and obtained through the structures specifically pointed out in the written description, claims, and drawings.

[0060] The technical solution of the present invention will be further described in detail below through the drawings and embodiments. Description of the Drawings

[0061] The drawings are used to provide a further understanding of the present invention, and constitute a part of the description. They are used together with the embodiments of the present invention to explain the present invention, and do not constitute a limitation to the present invention. In the drawings:

[0062] Figure 1 is a structural diagram of a power remote security protection system in an embodiment of the present invention;

[0063] Figure 2 is another structural diagram of a power remote security protection system in an embodiment of the present invention. Detailed Embodiments

[0064] The following describes the preferred embodiments of the present invention with reference to the drawings. It should be understood that the preferred embodiments described herein are only used to illustrate and explain the present invention, and are not used to limit the present invention.

[0065] Embodiment 1:

[0066] The present invention provides a power remote security protection system, as Figure 1 shown. The system includes:

[0067] Monitoring module: used to monitor the data information of power terminal devices in the power grid to obtain a terminal data set;

[0068] Extraction module: used to extract features from the terminal data set;

[0069] Parsing module: Obtain the detection instruction manual from the test management platform, and parse the detection instruction manual to generate detection objects and compliance indicators associated with the detection objects;

[0070] Establishment module: Used to establish a feature evaluation model based on compliance indicators;

[0071] Analysis module: Perform a situation analysis on the extracted features according to the feature evaluation model, and determine whether to issue an abnormal behavior warning based on the situation analysis result.

[0072] In this embodiment, the power terminal equipment includes: substations, distribution rooms, distribution boxes, electric meters, power load controllers, and on-line monitoring equipment.

[0073] In this embodiment, the data information includes working parameters such as the power generation amount, power generation capacity, and power generation efficiency of the power terminal equipment.

[0074] In this embodiment, the terminal data set refers to the data set formed by summarizing all the data of the terminal equipment.

[0075] In this embodiment, feature extraction refers to extracting features from the data in the data set. For example, the power generation amounts of equipment 1 are 20 kWh, 19 kWh, 21 kWh, etc., so as to determine the power generation characteristics of equipment 1.

[0076] In this embodiment, the detection instruction manual is generated by the test management platform according to the detection items, and the detection items are constructed by the test management platform according to the detection objects and the compliance indicators associated with the detection objects.

[0077] Among them, the detection object refers to the host device, application system, and network security device, and the compliance indicators associated with the detection object are determined according to the compliance indicators stored in the compliance library. Among them, the compliance indicators refer to, for example, how many kilowatts the power generation amount per hour must reach.

[0078] In this embodiment, the feature evaluation model refers to a model for evaluating the features of a data set, which is a model trained based on a large number of data sets containing features and the evaluation results of experts on different features as training samples.

[0079] In this embodiment, the situation analysis refers to analyzing the advantages, disadvantages, strengths, and weaknesses formed by the extracted features. For example, if the current per hour is 220 mA, which can make the efficiency of the entire power generation process the highest, at this time, it is determined that the power generation process is efficient, there is no abnormality, and no abnormal behavior warning is required.

[0080] In this embodiment, the abnormal behavior may be: excessive load of the power load controller, abnormal operation of the electric meter, and instantaneous excessive current.

[0081] The beneficial effects of the above technical solution are as follows: By remotely monitoring the data of power grid power terminal devices, forming a data set, extracting features, constructing a feature evaluation model, and analyzing the situation of the extracted features, the abnormal cause can be quickly determined remotely, the judgment efficiency is improved, and a large amount of time is saved.

[0082] Embodiment 2:

[0083] The present invention provides a power remote security protection system, as Figure 2 shown, the monitoring module includes:

[0084] The first acquisition unit: acquires the device attributes and device types of power terminal devices in the power grid;

[0085] The setting unit: sets the data acquisition period, data acquisition rate, and data items based on the device attributes and device types;

[0086] The acquisition unit: acquires the data information of each data item in the power terminal device according to the data acquisition period and the data acquisition rate;

[0087] The integration unit: integrates and processes the data information of each data item of the power terminal device to obtain a terminal data set.

[0088] In this embodiment, the device attributes include: basic information, parameters, such as brand model, rated voltage, rated current, rated frequency, etc.

[0089] In this embodiment, the device types include: monitoring devices, control devices, communication devices, and power generation devices.

[0090] In this embodiment, the data acquisition period refers to how often data is acquired, for example, it is set to acquire data from the monitoring device every two days.

[0091] In this embodiment, the data acquisition rate refers to the speed of acquiring data, for example, the acquisition speed each time is to complete the acquisition within 1 hour.

[0092] In this embodiment, the data items are, for example, current, voltage, and capacitance.

[0093] In this embodiment, the data information means, for example, the current is 200 amperes.

[0094] In this embodiment, the terminal data set is the data information obtained based on various data items.

[0095] The beneficial effects of the above technical solution are as follows: According to the device attributes and device types of power terminals in the power grid, the data information of each data item of the power terminal device is acquired. The data information of each data item is integrated and processed to obtain a terminal data set, which can integrate different data and make the data more centralized and unified.

[0096] Example 3:

[0097] The present invention provides a power remote security protection system, and an extraction module, including:

[0098] A screening unit: screening the eigenvalues in the obtained terminal data set to obtain a basic eigenvalue set;

[0099] A first classification unit: classifying the eigenvalues under the same data item in the basic eigenvalue set to obtain a classification result under the same data item;

[0100] A first determination unit: determining the eigenvalue distribution of the same data item in the terminal data set based on the classification result;

[0101] An extraction unit: selecting a corresponding feature extraction method from the distribution - method mapping table according to the eigenvalue distribution of different data items, and extracting the features of all data under the corresponding data item in the terminal data set.

[0102] In this embodiment, the data in the terminal data set refers to the data set aggregated from various data of all terminal devices. Among them, for example, if the variance of a feature itself is very small, it means that there is basically no difference in this feature for the samples, then such eigenvalues can be filtered out, and the remaining data together form the basic feature set.

[0103] In this embodiment, the terminal data set includes data item 1, data item 2, and data item 3. Among them, under data item 1, there are data 1, data 2, and data 3. At this time, data 1 and data 2 are relatively close. At this time, data 1 and data 2 are regarded as a classification result, and data 3 is regarded as a classification result. At this time, the classification result under data item 1 is obtained, and the same applies to data item 2 and data item 3.

[0104] In this embodiment, the eigenvalue distribution refers to a situation after classifying the data included in the same data item, that is: data 1 and data 2 are the first category, and there are 2 data in the first category, data 3 is the second category, and there is 1 data in the second category.

[0105] In this embodiment, the distribution - method mapping table includes different distribution situations and extraction methods. Therefore, a feature extraction method consistent with the eigenvalue distribution can be directly obtained, and then feature extraction is carried out. And the feature extraction method may be to extract the data with too large a standard error as the extracted feature.

[0106] The beneficial effects of the above technical solution are as follows: By screening the eigenvalues of the data to obtain the basic eigenvalue set, classifying the data in the set, and determining the feature distribution according to the classification results, the features of the data can be accurately extracted, making the extracted features more representative.

[0107] Embodiment 4:

[0108] The present invention provides a power remote security protection system, and the first classification unit includes:

[0109] An acquisition subunit: acquiring multiple random eigenvalues in each data item from the basic eigenvalue set through a random acquisition method;

[0110] A determination subunit: determining the edit distance between multiple random eigenvalues under the same data item to obtain a distance similarity array for each random eigenvalue;

[0111] A classification subunit: classifying the eigenvalues with similar distances based on the distance similarity array for the eigenvalues under the same data item to obtain a classification result.

[0112] In this embodiment, the random eigenvalues in data item 1 include: a1, a2, and a3. Among them, the distance similarity array for a1 is: [|a1 - a2|, |a1 - a3|]; the distance similarity array for a2 is: [|a2 - a1|, |a2 - a3|], and the distance similarity array for a3 is: [|a3 - a1|, |a3 - a2|].

[0113] The beneficial effects of the above technical solution are as follows: By randomly acquiring multiple random eigenvalues from the basic eigenvalue set, determining the edit distance between the multiple random eigenvalues to obtain a distance similarity array, and classifying the eigenvalues, the eigenvalues can be made more unified and concentrated.

[0114] Embodiment 5:

[0115] The present invention provides a power remote security protection system, and the parsing module includes:

[0116] A decryption unit: obtaining a test instruction book from the test management platform through data acquisition authority and performing decryption processing on the test instruction book;

[0117] A second acquisition unit: acquiring multiple detection items of power terminals in the power grid;

[0118] A second determination unit: determining a detection object according to the decrypted test instruction book and the multiple detection items of the power terminal;

[0119] A third determination unit: parsing the test instruction book to obtain the standard working parameters of each detection object, and determining the compliance indicators associated with each detection object according to the standard working parameters.

[0120] In this embodiment, the decryption process refers to canceling the confidentiality level of the test instruction manual, performing reverse operations on the ciphertext data according to the encryption algorithm and key, removing the noise, transformation, randomness, etc. added during encryption, and restoring the ciphertext data to its original plaintext data state. After the decryption process, the guidance process based on the test instruction manual can be obtained, and the guidance process includes different objects and implementation items of different objects.

[0121] In this embodiment, the multiple detection items include: capacitance detection item, voltage detection item, power generation detection item, etc.

[0122] In this embodiment, the objects that meet the multiple detection items are determined, and the objects that meet the requirements are regarded as the detection objects, such as host devices, application systems, network security devices, etc.

[0123] In this embodiment, for example, the standard working parameters of Detection Object 1 are: Parameter 1, Parameter 2, and Parameter 3. Among them, Parameter 2 and Parameter 3 are the parameters that need to be monitored key points. Therefore, the compliance indicators must be related to Parameter 2 and Parameter 3.

[0124] The beneficial effects of the above technical solution are: by obtaining the test instruction manual from the test management platform, then performing decryption processing on the obtained test instruction manual, obtaining multiple detection items of power terminals in the power grid, determining the detection objects and the compliance indicators associated with the detection objects, not only solving the problems of low efficiency and low accuracy of traditional compliance detection, but also being able to continuously detect the objects to be detected to avoid omissions.

[0125] Embodiment 6:

[0126] The present invention provides a power remote security protection system, and a building module, including:

[0127] A construction unit: obtaining the characteristic data of each compliance indicator, and obtaining the feedback evaluation of each characteristic data in the same compliance indicator, and constructing a corresponding characteristic vector;

[0128] An establishment unit: sequentially inputting different characteristic vectors into the neural network model to establish a characteristic evaluation model.

[0129] In this embodiment, for example, the parameter description of the obtained characteristic data that meets the compliance indicator is 01, and the obtained characteristic data are respectively: 21 mA, 20 mA, 0 mA, 30 mA, 40 mA, etc., and the corresponding feedback evaluations are respectively: qualified, qualified, unqualified, qualified, exceeded the standard evaluations. The constructed characteristic vector is: [21 mA - qualified, 20 mA - qualified, 0 mA - unqualified, 30 mA - qualified, 40 mA - exceeded the standard].

[0130] The beneficial effects of the above technical solution are as follows: By obtaining the characteristic data of each index and the feedback evaluation of the data, it is convenient to construct a reasonable vector and input it into the model for training, which is convenient to construct a characteristic evaluation model.

[0131] Example 7:

[0132] The present invention provides a power remote security protection system, and the analysis module includes:

[0133] Conversion unit: Convert the extracted features into standardized data according to the feature-conversion mapping table;

[0134] Input unit: Input the extracted features into the feature evaluation model to obtain an initial evaluation result;

[0135] Analysis unit: Perform feature analysis on the standardized data to obtain an optimization factor;

[0136] Judgment unit: Optimize the initial evaluation result based on the optimization factor and judge whether to issue a warning for abnormal behavior.

[0137] In this embodiment, the feature-conversion mapping table is a table including different characteristic data and the standardized performance of the data. Furthermore, the corresponding standard result can be obtained through matching with the characteristic data, which is convenient for optimizing the initial evaluation result subsequently.

[0138] In this embodiment, the standardized data can be the normalization processing of the same type of characteristic data.

[0139] In this embodiment, the initial evaluation result is directly obtained based on the model and is the evaluation result for each feature.

[0140] In this embodiment, the purpose of feature analysis is to determine whether the standardized data is within a reasonable standard range, and then determine the fluctuation situation of each standardized data. If the fluctuation situations tend to be consistent and are all within the standard range, the optimization factor is 0. At this time, there is no need to optimize the initial evaluation result.

[0141] In this embodiment, if there is standardized data corresponding to fluctuations outside the standard range, and the evaluation result of the characteristic data consistent with this standardized data is qualified, at this time, the two are inconsistent, and the evaluation result needs to be optimized. For example, adding an unqualified label to the corresponding same characteristic data is the optimized result, that is, finally there are two labels for the same characteristic data, one is qualified and the other is unqualified. At this time, a behavior warning needs to be issued.

[0142] The beneficial effects of the above technical solution are as follows: By standardizing the features and performing feature analysis to determine the optimization factors, and at the same time, analyzing the features based on the model to obtain the evaluation results, and judging whether to optimize the results through the two results, so as to realize the warning of abnormal behavior and facilitate security protection.

[0143] Embodiment 8:

[0144] The present invention provides a power remote security protection system, further comprising:

[0145] Data determination module: Read the original power data collected by the power terminals in the power grid, determine the data type of the original power data, and obtain the regenerated data;

[0146] Source matching module: Trace the data source of the regenerated data, and match the regenerated data with the data source one by one according to the source database to obtain the classified data;

[0147] Dataset acquisition module: Sort the classified data from the same data source in chronological order, and correspond the dataset at each time point with the load in the system one by one to obtain the target load dataset from the same data source;

[0148] Data clustering module: Obtain the operation characteristics of the same data source, and perform clustering processing on the corresponding target load dataset based on the operation characteristics to obtain several operation data groups;

[0149] Match the initial value of each parameter in the same operation data group with the value mapping table to obtain the target value of the corresponding operation data group;

[0150] Lock the operation time points involved in the same operation data group, match the load value with the dataset at each operation time point, and draw the array load curve corresponding to the operation data group;

[0151] Curve comparison module: According to the typical daily load curve, array load curve and standard load curve under the same operation data group;

[0152] Calculate the first similarity S1 between the typical daily load curve y01 and the array load curve y0s under the same operation array, and the second similarity S2 between the standard load curve y0b and the array load curve y0s under the same operation array;

[0153] S1 = sim(y01, y0s);

[0154] S2 = sim(y0b, y0s);

[0155] Wherein, sim() represents the similarity function;

[0156] Perform a first comparison between the target value r01 of the typical load curve y01 and the standard value r02 of the standard load curve y0b. At the same time, perform a second comparison between the first similarity and the second similarity;

[0157] Determine the upper boundary and lower boundary of the division according to the first comparison result and the second comparison result;

[0158]

[0159] Among them, represents the adjustment coefficient; F1 represents the value of the upper boundary of the division; F2 represents the value of the lower boundary of the division; F Z represents the set left range value; F Y represents the set right range value;

[0160] Lock the data outside the boundary for the corresponding array load curve based on the upper boundary and lower boundary of the division, and regard it as bad data;

[0161] Determine the current position of the bad data, obtain the values of the first data and the second data that are at the closest distance to the bad data, determine the replacement data corresponding to the bad data, and perform the replacement;

[0162]

[0163] Among them, J represents the boundary value for determining that the data at the current position is bad data; T1 represents the value of the first data; T2 represents the value of the second data; T0 represents the value of the replacement data.

[0164] In this embodiment, the acquisition parameters for the same data source at the same time point are the same, only the parameter values are different. Therefore, the obtained target load data set includes the acquisition data at different time points under the same data source.

[0165] In this embodiment, the operating characteristics are determined based on the working attributes of the device itself. Since the device has different working modes and the operating characteristics under different working modes are different, the classification data can be clustered according to the working mode, and then several operating data groups can be obtained, that is, each clustering result corresponds to an operating data group.

[0166] In this embodiment, bad data refers to the data with acquisition or reading errors during the acquisition or reading process and needs to be replaced.

[0167] In this embodiment, the original power data refers to the data that has not been processed, that is, the data directly collected based on the power terminal.

[0168] In this embodiment, the data types include: current type, voltage type, power type, etc. The data is classified according to the source. For example, the determined current data includes: Data 01, Data 02, the voltage data includes: Data a01, Data a02, and the power data includes: g01, g02. At this time, the source of Data 01 is device @1, the source of Data 02 is device @2, the source of Data a01 is device @1, the source of Data a02 is device #1, the source of Data g01 is device #2, and the source of Data g02 is device #1. Therefore, the data for device @1 includes: Data 01, Data a01; the data for device @2 includes: Data 02; the data for device #1 includes: Data a02, g02; and the data for device #2 includes: g01. At this time, the data classification is completed.

[0169] In this embodiment, since the original power data is collected based on power terminals and different data collection processes are related to corresponding collection terminals, it is possible to directly trace the source of the regenerated data, which is the relevant terminal device.

[0170] In this embodiment, each classified data corresponds to a device terminal. Therefore, the data feature is the analysis result including multiple data of the device at the same moment, that is, there are multiple different parameter data collected at the same moment.

[0171] In this embodiment, the value mapping table contains combinations of initial values of different parameters in the same array and values that match the target values for these combinations. For example, parameter 1 = r1, parameter 2 = r2, and the target value matching the combined parameters 1 = r1, 2 = r2 is p1.

[0172] In this embodiment, the closest distance refers to the two data with the shortest time distance, which may be on the same side of the current position or on both sides of the current position.

[0173] The beneficial effects of the above technical solution are as follows: By classifying the original power data, regenerated data is obtained, the data source of the regenerated data is obtained, a load data set is obtained, and further through clustering processing and curve comparison, bad data can be obtained, data replacement can be performed, the reliability of the data can be improved, and a basis for security protection is provided.

[0174] Obviously, those skilled in the art can make various changes and modifications to the present invention without departing from the spirit and scope of the present invention. Thus, if these modifications and variations of the present invention fall within the scope of the claims of the present invention and its equivalent technologies, the present invention is also intended to include these changes and modifications.

Claims

1. A power remote safety protection system, characterized in that: The system includes: Monitoring module: used to monitor the data information of power terminal equipment in the power grid and obtain terminal data sets; Extraction module: used for extracting features from the terminal data set; Parsing module: obtains the test instructions from the test management platform, parses the test instructions to generate test objects and compliance indicators associated with the test objects; Building module: used to build feature evaluation models based on compliance indicators; Analysis module: Perform situation analysis on the extracted features according to the feature evaluation model, and determine whether to issue an abnormal behavior warning based on the situation analysis results.

2. The electric power remote safety protection system according to claim 1, characterized in that: Monitoring modules, including: A first acquisition unit: acquiring device attributes and device types of power terminal devices in the power grid; Setting unit: setting the data collection period, data collection rate and data items based on the device attributes and device type; Collection unit: collects data information of each data item in the power terminal equipment according to the data collection cycle and data collection rate; Integration unit: Integrate and process the data information of each data item of the power terminal equipment to obtain the terminal data set.

3. The electric power remote safety protection system according to claim 1, characterized in that: Extraction modules, including: Screening unit: Screening the data in the acquired terminal data set by feature values ​​to obtain a basic feature value set; A first classification unit: classifies the feature values ​​under the same data item in the basic feature value set to obtain the classification result under the same data item; A first determining unit: determining a feature distribution of the same data item in the terminal data set based on the classification result; Extraction unit: selects the corresponding feature extraction method from the distribution-method mapping table according to the feature distribution of different data items, and performs feature extraction on all data under the corresponding data item in the terminal data set.

4. The electric power remote safety protection system according to claim 3 is characterized in that: The first classification unit includes: Acquisition subunit: acquiring multiple random eigenvalues ​​in each data item from the basic eigenvalue set by a random acquisition method; Determine the subunit: determine the edit distance between multiple random eigenvalues ​​under the same data item, and obtain a distance similarity array for each random eigenvalue; Classification subunit: Based on the distance similarity array, the feature values ​​under the same data item are classified into similar distances to obtain the classification result.

5. The electric power remote safety protection system according to claim 1, characterized in that: Parsing modules, including: Decryption unit: obtains the test instruction from the test management platform through data acquisition authority, and decrypts the test instruction; A second acquisition unit is used to acquire multiple detection items of power terminals in the power grid; A second determination unit: determines the detection object according to the decrypted detection guide and multiple detection items of the power terminal; The third determination unit: parses the detection instruction to obtain the standard working parameters of each detection object, and determines the compliance indicators associated with each detection object according to the standard working parameters.

6. The electric power remote safety protection system according to claim 1, characterized in that: Build modules, including: Construction unit: obtain the characteristic data of each compliance indicator, obtain the feedback evaluation of each characteristic data in the same compliance indicator, and construct the corresponding characteristic vector; Establishing unit: inputting different feature vectors into the neural network model in sequence to establish a feature evaluation model.

7. The electric power remote safety protection system according to claim 1, characterized in that: Analysis modules, including: Conversion unit: converts the extracted features into standardized data according to the feature-conversion mapping table; Input unit: inputs the extracted features into the feature evaluation model to obtain the initial evaluation results; Analysis unit: performing feature analysis on the standardized data to obtain optimization factors; A judgment unit is configured to optimize the initial evaluation result based on the optimization factor and judge whether to issue an abnormal behavior warning.

8. The electric power remote safety protection system according to claim 1, characterized in that: Also includes: Data determination module: reads the original power data collected by the power terminal in the power grid, determines the data type of the original power data, and obtains the regenerated data; Source matching module: tracing the data source of the regenerated data, matching the regenerated data with the data source one by one according to the source database, and obtaining classified data; Data set acquisition module: sorts the classified data from the same data source in chronological order, and matches the data set at each time point with the load in the system one by one, so as to obtain the target load data set from the same data source; Data clustering module: obtaining the operation characteristics of the same data source, and clustering the corresponding target load data set based on the operation characteristics to obtain a number of operation data groups; Based on matching the initial value of each parameter in the same operation data group with the value mapping table, a target value of the corresponding operation data group is obtained; Locking the operation time points involved in the same operation data group, matching the load value to the data set at each operation time point, and drawing the array load curve of the corresponding operation data group; Curve comparison module: based on the typical daily load curve, array load curve and standard load curve under the same operating data group; Calculate the first similarity S1 between the typical daily load curve y01 and the array load curve y0s under the same operating array and the second similarity S2 between the standard load curve y0b and the array load curve y0s under the same operating array; S1 = sim(y01,y0s); S2 = sim(y0b,y0s); Among them, sim() represents the similarity function; Perform a first comparison between the target value r01 of the typical load curve y01 and the standard value r02 of the standard load curve y0b, and perform a second comparison between the first similarity and the second similarity; Determine an upper boundary and a lower boundary according to the first comparison result and the second comparison result; in, represents the adjustment coefficient; F1 represents the value of the upper boundary; F2 represents the value of the lower boundary; F Z Indicates setting the left range value; F Y Indicates setting the right range value; Based on the upper boundary and the lower boundary of the division, the data outside the boundary of the corresponding array load curve is locked and regarded as bad data; Determine the current position of the bad data, and obtain the value of the first data and the value of the second data that are closest to the bad data, determine to obtain replacement data corresponding to the bad data, and replace it; Among them, J represents the value of the boundary for determining that the data at the current position is bad data; T1 represents the value of the first data; T2 represents the value of the second data; and T0 represents the value of the replacement data.

Images