Big data terminal data restoration system based on power grid

Abstract

The invention discloses a big data terminal data restoration system based on a power grid. The system comprises a data acquisition module which is used for acquiring operation data of a power transmission line in real time through a sensor and an electric meter; the method relates to the technical field of power grid data processing, and comprises the following steps: when a data missing or error condition occurs, predicting missing or error data through a data restoration numerical value providing unit in a risk prediction module; in the data restoration process, other device data and operation information related to the data can be obtained through the association model, the association data are utilized to construct a data inference model of the association data, and missing data or error data are inferred; and the data restoration module calculates a restoration value of missing data or wrong data through a linear interpolation algorithm, integrates deduction values of the data restoration value providing unit and the data deduction model of the associated data, and takes an average value of the three data as a final restoration value. The data recovery accuracy is improved while the data is effectively recovered.

Description

Technical Field

[0001] The present invention relates to the technical field of power grid data processing, and in particular to a big data terminal data repair system based on a power grid. Background Art

[0002] Grid terminal data refers to the data collected from various grid-related terminal devices in the power system. These data reflect the operating status of the grid, equipment status, and user electricity consumption.

[0003] Publication No. CN118964874B discloses a distribution network measurement data repair method and device. When it is determined that there is missing data in the distribution network measurement data set, the data type of the missing data and the data type of each of the multiple measurement data in the distribution network measurement data set are determined; when it is determined that the data types of the multiple measurement data are different and there is only one missing data at the same monitoring time, based on multiple measurement data at the same monitoring time as the missing data and an electrical correlation function, first initial repair data of the missing data is determined; based on the distribution network measurement data set, a distribution network measurement data tensor is generated; data repair of the distribution network measurement data tensor by tensor decomposition is performed to obtain second initial repair data matching the missing data; and repair data matching the missing data is obtained based on the first initial repair data and the second initial repair data.

[0004] However, the above application still has the following problems: the data repair method of the above application mainly uses electrical correlation function and tensor decomposition to repair data, and the data repair method is relatively simple;

[0005] The electrical correlation function mainly focuses on establishing the relationship between data from the perspective of electrical principles. However, the data relationship in the actual power grid is extremely complex. In addition to electrical correlation, it is also affected by non-electrical factors such as equipment aging.

[0006] Tensor decomposition mainly analyzes and processes data from the perspective of mathematical structure and dimension. Although it can mine some potential patterns in the data, it cannot fully consider the physical meaning and actual operation behind the data. For example, in the distribution network, the impact of data loss on different devices is different. Tensor decomposition cannot accurately identify it, and thus cannot perform effective data repair for specific situations, which affects the accuracy of the repair results.

[0007] In summary, the accuracy of the above application data repair needs to be improved. Summary of the invention

[0008] In order to solve the technical problems existing in the background technology, the present invention proposes a big data terminal data repair system based on power grid.

[0009] A big data terminal data repair system based on the power grid proposed by the present invention includes:

[0010] Data acquisition module: used to collect the operation data of the transmission line in real time through sensors and electric meters;

[0011] Abnormality detection module: used to identify whether there is data loss or error in the operation data of the transmission line obtained in real time in the data acquisition module through data mining technology and statistical analysis methods;

[0012] Risk prediction module, the risk prediction module includes:

[0013] Data prediction unit: used to collect various basic data, equipment parameter data, historical operation data and fault record data of the power grid;

[0014] The basic data includes power grid topology structure data;

[0015] Obtain the operation data of the transmission line obtained in real time by the data acquisition module;

[0016] Build a data prediction model through big data analysis and machine learning algorithms;

[0017] Data repair value providing unit: used to generate the data repair value X when there is data loss or error 修复1 ;

[0018] Data association module: By building a data association model, the data association model is used to obtain the associated data of a certain data; when there is data loss or error in a certain data, the data association model is used to predict the influence range and degree of the data on other related data; and then the potential influence on the operation state of the power grid equipment can be further inferred.

[0019] Data repair module: used to generate the final repair value when there is data loss or error, fill in the missing data or replace the wrong data.

[0020] Preferably, in the data acquisition module, the collected data is preliminarily cleaned to remove obvious errors or duplicate data.

[0021] Preferably, in the data acquisition module, the sensors include temperature sensors, current sensors, voltage sensors, and power sensors.

[0022] Preferably, the risk prediction module further includes:

[0023] Data repair priority determination unit: used to determine the data repair priority.

[0024] Preferably, in the data repair priority determination unit, the data repair priority is determined as follows:

[0025] When there are multiple pieces of missing data or incorrect data simultaneously, start the data repair priority determination unit;

[0026] Set the line overload risk probability, power outage range, and power outage duration as the evaluation indicators for the impact degree of data missing or incorrect;

[0027] Evaluate the line overload risk probability by calculating the ratio of the actual current to the rated current and using statistical analysis methods;

[0028] Collect various basic data of the power grid through the risk prediction module, construct the topological model of the power grid, abstract the equipment in the power grid as nodes and edges, and use graph theory algorithms to analyze the connection relationship between faulty equipment or lines and other equipment; when a certain transmission line fails, determine the substations and users that lose power supply through topological analysis to determine the power outage range;

[0029] Collect historical fault repair data, and based on the average repair time of different types of faults in the historical data, when a new fault occurs, directly use the average repair time of this type of fault as the estimated repair time;

[0030] Set the weights of the line overload risk probability, power outage range, and power outage duration as wx, wy, and wz respectively, and wx + wy + wz = 1;

[0031] Inject the data missing or data error scenario into the power grid model to simulate the operation state of the power grid under abnormal data conditions;

[0032] According to the simulation results, calculate the line overload risk probability, power outage range, and power outage duration caused by data missing or incorrect, and calculate the impact degree according to the weights of the line overload risk probability, power outage range, and power outage duration. The data repair priority with a higher impact degree is greater than that with a lower impact degree. When data repair is performed in the data repair module, repair in sequence according to the priority sorting result.

[0033] Preferably, in the data repair value providing unit, the generation of the data repair value when there is data missing or incorrect is as follows:

[0034] When the anomaly detection module identifies that there is data missing or incorrect, set the data with data missing or incorrect as X, and assume that at time t i There is data missing or incorrect;

[0035] Among the data collected from the data prediction unit, assume that the number of key influencing factors affecting X is m, which are F1, F2,..., F m ;

[0036] Through the analysis of historical data, use statistical methods to determine the weights of each key influencing factor, which are w1, w2,..., w m, and w1 + w2 +... + w m = 1;

[0037] Construct a prediction model f through a neural network algorithm;

[0038] Input the data of each key influencing factor collected in real time by the data acquisition module into the data prediction model to predict the predicted values of each key influencing factor at time t i Let the predicted values of each key influencing factor obtained be F1 t i at time t i , F2 t i ,..., F m t i ;

[0039] Substitute F1 t i , F2 t i ,..., F m t i into the prediction model f to calculate the predicted data value Xf of X at time t i ;

[0040] Extract the value of X in the time period [t i , t i-k , t i+k near time t from historical data, and calculate the average value X 平均 of these values, where k is a set time window size;

[0041] Determine the repaired data value of X through the weighted average method where is the weight coefficient, and its value range is [0, 1].

[0042] Preferably, in the data association module, the data association model is constructed based on graph database technology. First, obtain the power grid operation data, use the data items in the power grid operation data as nodes, and the potential association relationships between the data as edges to construct a graph structure, and analyze in this graph structure through graph algorithms to locate the data associated with a specific data;

[0043] The power grid operation data is a large and complex data set, covering the relevant information of all links of the power grid from power generation, transmission, transformation, distribution to power consumption; the power grid operation data includes the operation data of the transmission lines obtained in real time in the data acquisition module.

[0044] Preferably, in the data repair module, the final repair value when data is missing or incorrect is generated in the following manner:

[0045] Obtain the data repair value X through the data repair value providing unit修复1 ;

[0046] Obtain associated data through a data association model, construct a data inference model for the associated data through a machine learning algorithm. When the anomaly detection module identifies data loss or errors, the data inference model of the associated data is used to calculate the inferred value X of the missing data or incorrect data 修复2 ;

[0047] When the anomaly detection module identifies data loss or errors, calculate the repaired value X of the missing data or incorrect data through a linear interpolation algorithm 修复3 ;

[0048] Then the final repaired value when there is data loss or errors

[0049] In the present invention, the proposed big data terminal data repair system based on the power grid has the following beneficial technical effects:

[0050] 1. The anomaly detection module uses data mining techniques and statistical analysis methods to accurately identify data loss or error situations. Through the data repair value providing unit in the risk prediction module, predict the missing or incorrect data;

[0051] During the data repair process, the data association module can provide associated data support for the repair of missing or incorrect data. When a certain data has problems, other device data and operation information related to this data can be obtained through the association model, and a data inference model of the associated data is constructed using these associated data to infer the missing data or incorrect data;

[0052] The data repair module calculates the repaired value of the missing data or incorrect data through a linear interpolation algorithm, and combines the inferred value calculated by the data inference model of the associated data and the data repair value obtained by the data repair value providing unit. The average value of the three data is determined as the final repaired value. While effectively repairing the data, ensure the integrity of the big data terminal data of the power grid and improve the accuracy of data repair.

[0053] 2. The data repair priority determination unit sets evaluation indicators for line overload risk probability, power outage range, and power outage duration. By quantifying the potential impact of data loss or errors on the operation of the power grid, and determining the data repair priority based on these quantification results, ensure that data problems with a greater impact on the safe and stable operation of the power grid are processed first, and improve the anti-risk ability of the power grid.

[0054] 3. The data prediction model in the risk prediction module predicts the operation parameters of power grid equipment in a future period by long-term monitoring and analysis of the operation data of transmission lines, which is conducive to identifying potential risks of equipment aging and performance degradation in advance; when a certain data is missing or incorrect, the data association model is used to predict the influence range and degree of this data on other relevant data, and further infer the potential impact on the power grid operation status, which is conducive to helping power grid operation and maintenance personnel take measures in time to prevent the further spread of risks.

[0055] Additional aspects and advantages of the present invention will be given in part in the following description, become apparent in part from the following description, or be learned through the practice of the present invention. Brief Description of the Drawings

[0056] Figure 1 It is a schematic block diagram of the system of the present invention. Detailed Embodiments

[0057] Embodiments of the present invention will be described in detail below. Examples of the embodiments are shown in the drawings, where the same or similar symbols represent the same or similar elements or elements with the same or similar functions throughout. The embodiments described below with reference to the drawings are exemplary and are only used to explain the present invention and should not be construed as a limitation to the present invention.

[0058] As Figure 1 shown, a big data terminal data repair system based on a power grid includes:

[0059] Data acquisition module: Real-time collects the operation data of transmission lines through sensors and electric meters;

[0060] In an optional embodiment, in the data acquisition module, the collected data is preliminarily cleaned to remove obviously incorrect or duplicate data; the value range and time sequence rules can be set to determine incorrect data, and duplicate data is removed through hash calculation;

[0061] In an optional embodiment, in the data acquisition module, the sensors include temperature sensors, current sensors, voltage sensors, and power sensors;

[0062] The data acquisition module real-time collects the operation data of transmission lines with the help of a variety of sensors and electric meters, and conducts preliminary cleaning to remove obviously incorrect or duplicate data. This ensures the basic quality of the data entering the system, provides an accurate and reliable data source for subsequent analysis and decision-making, and avoids interfering with subsequent risk assessment and data repair work due to incorrect or redundant initial data.

[0063] Abnormality detection module: For the operation data of the transmission lines obtained in real time in the data acquisition module, it identifies whether there is data missing or incorrect through data mining techniques and statistical analysis methods;

[0064] Risk assessment module, which includes:

[0065] Data prediction unit: By collecting various types of basic data of the power grid, equipment parameter data, historical operation data, and fault record data;

[0066] The basic data includes power grid topology data; The power grid topology data such as the connection relationship of transmission lines, the location and connection of substations;

[0067] The fault record data includes the time, location, and impact range of the fault;

[0068] Obtain the operation data of the transmission line obtained in real time by the data acquisition module;

[0069] Through big data analysis and machine learning algorithms, construct a data prediction model;

[0070] The data prediction model is used to predict the values of power grid equipment operation parameters, line load, and power outage probability in a future period of time;

[0071] The risk assessment module further includes:

[0072] Data repair priority determination unit, used to determine the data repair priority;

[0073] In an optional embodiment, in the data repair priority determination unit, the data repair priority is determined as follows:

[0074] When there are multiple missing data or incorrect data at the same time, start the data repair priority determination unit;

[0075] Set the line overload risk probability, power outage range, and power outage duration as the evaluation indicators of the impact degree of data loss or error;

[0076] By calculating the ratio of the actual current to the rated current, evaluate the line overload risk probability through statistical analysis methods;

[0077] Through various types of basic data of the power grid collected by the risk assessment module, construct a topology model of the power grid, abstract the equipment in the power grid as nodes and edges, and use graph theory algorithms to analyze the connection relationship between faulty equipment or lines and other equipment; When a certain transmission line fails, determine the substations and users that lose power supply through topology analysis, and determine the power outage range;

[0078] The equipment in the power grid includes substations, transmission lines, and transformers;

[0079] Collect historical fault repair data, and according to the average repair time of different types of faults in the historical data, when a new fault occurs, directly use the average repair time of this type of fault as the estimated repair time;

[0080] Let the weights of the line overload risk probability, power outage scope, and power outage duration be \(w_x\), \(w_y\), and \(w_z\) respectively, and \(w_x + w_y + w_z = 1\);

[0081] The weights of the line overload risk probability, power outage scope, and power outage duration can be set manually;

[0082] Inject data missing or data error scenarios into the power grid model to simulate the operation state of the power grid under abnormal data conditions;

[0083] According to the simulation results, calculate the line overload risk probability, power outage scope, and power outage duration caused by data missing or errors, and calculate the impact degree according to the weights of the line overload risk probability, power outage scope, and power outage duration. The data repair priority with a higher impact degree is greater than that with a lower impact degree. When data repair is performed in the data repair module, repair is carried out in sequence according to the priority sorting result;

[0084] The power grid model adopts the power grid model in the existing technology;

[0085] The data repair priority determination unit sets the evaluation indicators for the line overload risk probability, power outage scope, and power outage duration, quantifies the potential impact of data missing or errors on the operation of the power grid, and determines the data repair priority according to these quantification results to ensure that data problems with a greater impact on the safe and stable operation of the power grid are processed first, and improve the anti-risk ability of the power grid.

[0086] The risk prediction module further includes:

[0087] A data repair value providing unit for generating a data repair value \(X\) when data is missing or in error 修复1 ;

[0088] In an optional embodiment, in the data repair value providing unit, the generation of the data repair value when data is missing or in error is as follows:

[0089] When the anomaly detection module identifies that data is missing or in error, let the data with missing or error be \(X\), and assume that at time \(t\) i data is missing or in error;

[0090] Among the data collected from the data prediction unit, let the number of key influencing factors affecting \(X\) be \(m\), which are \(F_1\), \(F_2\),..., \(F\) m ;

[0091] Through the analysis of historical data, use statistical methods to determine the weights of each key influencing factor, which are \(w_1\), \(w_2\),..., \(w\) m , and \(w_1 + w_2 +... + w\) m = 1;

[0092] Construct a prediction model f through a neural network algorithm;

[0093] Input the data of each key influencing factor collected in real time by the data acquisition module into the data prediction model to predict the predicted values of each key influencing factor at time t i Let the predicted values of each key influencing factor obtained be F1 t i at time t i , F2 t i ,..., m F i t

[0094] Substitute F1 t i , F2 t i ,..., m F i t i into the prediction model f to calculate the predicted data value Xf of X at time t

[0095] Extract the value of X in the time period [t i , t i-k , t i+k near time t from historical data, and calculate the average value X 平均 of these values, where k is a set time window size; used to determine the statistical range of historical data;

[0096] Determine the repaired data value of X by the method of weighted average where is the weight coefficient, and its value range is [0, 1];

[0097] Data association module: Construct a data association model, which is used to obtain the associated data of a certain data; when a certain data has data missing or errors, the data association model is used to predict the influence range and degree of this data on other related data; and further infer the potential impact on the operation state of power grid equipment;

[0098] In an optional embodiment, in the data association module, the data association model is constructed based on graph database technology. First, obtain power grid operation data, use each data in the power grid operation data as a node, and the potential association relationship between data as an edge to construct a graph structure, and analyze in this graph structure through graph algorithms to locate the data associated with a certain specific data;

[0099] Power grid operation data is a huge and complex data set, covering relevant information of all links of the power grid from power generation, transmission, transformation, distribution to power consumption; power grid operation data includes the operation data of transmission lines obtained in real time in the data acquisition module;

[0100] The data prediction model in the risk prediction module predicts the operating parameters of grid equipment in a future period through long-term monitoring and analysis of the operating data of transmission lines, which is beneficial to identifying potential risks of equipment aging and performance degradation in advance; when a certain data is missing or incorrect, the data association model is used to predict the influence range and degree of this data on other relevant data, and further infer the potential impact on the grid operating state, which is beneficial to helping grid operation and maintenance personnel take measures in time to prevent the further spread of risks.

[0101] Data repair module: used to generate the final repair value when data is missing or incorrect, fill in the missing data or replace the incorrect data;

[0102] In an optional embodiment, in the data repair module, the final repair value when data is missing or incorrect is generated in the following manner:

[0103] Obtain the data repair value X through the data repair value providing unit 修复1 ;

[0104] Obtain associated data through the data association model, construct a data inference model for the associated data through machine learning algorithms, and when the anomaly detection module identifies that data is missing or incorrect, the data inference model for the associated data is used to calculate the inferred value X of the missing data or incorrect data 修复2 ;

[0105] When the anomaly detection module identifies that data is missing or incorrect, calculate the repair value X of the missing data or incorrect data through the linear interpolation algorithm 修复3 ;

[0106] Then the final repair value when data is missing or incorrect

[0107] The anomaly detection module uses data mining technology and statistical analysis methods to accurately identify data missing or incorrect situations. Through the data repair value providing unit in the risk prediction module, predict the missing or incorrect data;

[0108] During the data repair process, the data association module can provide associated data support for the repair of missing or incorrect data. When a certain data has problems, other device data and operation information related to this data can be obtained through the association model, and a data inference model for the associated data is constructed using these associated data to infer the missing data or incorrect data;

[0109] The data repair module calculates the repair values of missing or incorrect data through the linear interpolation algorithm, and combines the inferred values calculated by the data inference model of the associated data. The data repair value provided by the data repair value providing unit obtains the data repair value, and the average value of the three pieces of data is determined as the final repair value. While effectively repairing the data, the integrity of the power grid big data terminal data is ensured, and the accuracy of data repair is improved.

[0110] In the working process of this embodiment, the following steps are included:

[0111] S1. Real-time collect the operation data of the transmission line through sensors and meters;

[0112] S2. For the operation data of the transmission line obtained in real time in S1, identify whether there is missing or incorrect data through data mining technology and statistical analysis methods;

[0113] S3. Collect various basic data, equipment parameter data, historical operation data, and fault record data of the power grid; the basic data includes power grid topology structure data;

[0114] Obtain the operation data of the transmission line obtained in real time by the data acquisition module;

[0115] Build a data prediction model through big data analysis and machine learning algorithms;

[0116] S4. Generate the data repair value X when there is data missing or incorrect through the data repair value providing unit 修复1 ;

[0117] S5. Build a data association model, which is used to obtain the associated data of a certain piece of data; when there is data missing or incorrect in a certain piece of data, the data association model is used to predict the influence range and degree of this data on other related data; and then the potential impact on the operation state of power grid equipment can be further inferred.

[0118] S6. Generate the final repair value when there is data missing or incorrect through the data repair module to fill in the missing data or replace the incorrect data.

[0119] At the same time, the content not described in detail in this specification belongs to the prior art well known to those skilled in the art.

[0120] In the embodiments provided by the present invention, it should be understood that the disclosed system or method can be implemented in other ways. For example, the above-described invention embodiments are merely illustrative. For example, the division of modules is only a logical function division, and there may be other division methods in actual implementation.

[0121] The modules described as separate components may or may not be physically separated. The components shown as modules may or may not be physical modules, and they may be located in one place or distributed across multiple network modules. Some or all of the modules can be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0122] In addition, the functional modules in various embodiments of the present invention may be integrated into one processing module, or each module may exist physically alone, or two or more modules may be integrated into one module. The above integrated modules may be implemented in the form of hardware, or in the form of a combination of hardware and software functional modules.

[0123] For those skilled in the art of operation and maintenance, it is obvious that the present invention is not limited to the details of the above-mentioned exemplary embodiments, and without departing from the basic features of the present invention, the present invention can be implemented in other specific forms.

[0124] The above is only a preferred specific embodiment of the present invention, but the protection scope of the present invention is not limited thereto. Any person skilled in the art within the technical scope disclosed by the present invention, according to the technical solution and inventive concept of the present invention, makes equivalent substitutions or changes, and should be covered by the protection scope of the present invention.

Claims

1. A big data terminal data repair system based on a power grid, characterized in that, Including: Data acquisition module: used to collect the operation data of the transmission line in real time through sensors and electricity meters; Abnormality detection module: used to identify whether there is data loss or error in the operation data of the transmission line obtained in real time in the data acquisition module through data mining technology and statistical analysis methods; Risk prediction module, the risk prediction module includes: Data prediction unit: used to collect various basic data of the power grid, equipment parameter data, historical operation data and fault record data; The basic data includes power grid topology structure data; Obtain the operation data of the transmission line obtained in real time by the data acquisition module; Construct a data prediction model through big data analysis and machine learning algorithms; Data repair value providing unit: used to generate a data repair value X when data is missing or incorrect 修复1 ; Data association module: By constructing a data association model, the data association model is used to obtain the associated data of a certain data; when a certain data has data loss or error, the data association model is used to predict the influence range and degree of the data on other related data; Data repair module: used to generate the final repair value when there is data loss or error, fill in the missing data or replace the error data.

2. The big data terminal data repair system based on the power grid according to claim 1, characterized in that, In the data acquisition module, the collected data is preliminarily cleaned to remove obvious errors or duplicate data.

3. The big data terminal data repair system based on the power grid according to claim 1, characterized in that, In the data acquisition module, the sensors include temperature sensors, current sensors, voltage sensors, and power sensors.

4. The big data terminal data repair system based on the power grid according to claim 1, characterized in that, The risk prediction module also includes: Data repair priority determination unit, used to determine the data repair priority.

5. The big data terminal data repair system based on the power grid according to claim 4, characterized in that In the data repair priority determination unit, determine the data repair priority as follows: When there are multiple missing data or error data at the same time, start the data repair priority determination unit; Set the line overload risk probability, power outage range, and power outage duration as evaluation indicators for the influence degree of data loss or error; Evaluate the line overload risk probability through statistical analysis methods by calculating the ratio of the actual current to the rated current; Construct a topology model of the power grid through various basic data of the power grid collected by the risk prediction module, abstract the equipment in the power grid as nodes and edges, and use graph theory algorithms to analyze the connection relationship between the faulty equipment or line and other equipment; when a certain transmission line fails, determine the power supply lost substation and users through topology analysis to determine the power outage range; Collect historical fault repair data, and according to the average repair time of different types of faults in the historical data, when a new fault occurs, directly use the average repair time of this type of fault as the estimated repair time; Set the weights of the line overload risk probability, power outage range, and power outage duration as wx, wy, and wz respectively, and wx + wy + wz = 1; Inject the data loss or data error scenario into the power grid model to simulate the operation state of the power grid under abnormal data conditions; According to the simulation results, calculate the line overload risk probability, power outage range, and power outage duration caused by data loss or error, and calculate the influence degree according to the weights of the line overload risk probability, power outage range, and power outage duration. The data repair priority with a higher influence degree is greater than the data repair priority with a lower influence degree. When performing data repair in the data repair module, repair in sequence according to the priority sorting result.

6. The big data terminal data repair system based on the power grid according to claim 5, characterized in that, In the data repair value providing unit, when data is missing or incorrect, the generation of the data repair value is as follows: When the anomaly detection module identifies data loss or errors, let the data with data loss or errors be X, and at time t i there is data loss or errors; Among the data collected by the data prediction unit, let the number of key influencing factors affecting X be m, which are F1, F2, ..., F m ; By analyzing historical data, the weights of each key influencing factor are determined using statistical methods, which are w1, w2, ..., w m , and w1 + w2 +... + w m = 1; Construct a prediction model f through a neural network algorithm; The data of each key influencing factor collected in real time by the data acquisition module is input into the data prediction model to predict the predicted values of each key influencing factor at time t i Let the predicted values of each key influencing factor obtained at time t i be F1(t) i , F2(t) i ,..., F m (t) i ; Substitute F1 t i , F2 t i ,..., F m t i into the prediction model f to calculate the predicted data value Xf of X at time t i ; Extract time t from historical data i For the nearby time period [t i-k , t i+k , extract the values of X and calculate the average value X of these values 平均 , where k is a set time window size; Determine the repair data value of X by the method of weighted average wherein is the weight coefficient, and its value range is [0, 1].

7. The big data terminal data repair system based on the power grid according to claim 1, characterized in that, In the data association module, the data association model is constructed based on graph database technology. First, obtain the power grid operation data, use each item of data in the power grid operation data as nodes, and the potential association relationships between the data as edges to construct a graph structure. Analyze in this graph structure through graph algorithms to locate the data associated with a specific piece of data.

8. The big data terminal data repair system based on the power grid according to claim 7, characterized in that, In the data repair module, the final repair value when data is missing or incorrect is generated in the following manner: Obtain the data repair value X through the data repair value providing unit 修复1 ; Obtain associated data through a data association model, construct a data inference model for the associated data through a machine learning algorithm. When the anomaly detection module identifies data loss or errors, the data inference model of the associated data is used to calculate the inferred value X of the missing data or incorrect data 修复2 ; When the anomaly detection module identifies data loss or errors, the repaired value X of the missing or incorrect data is calculated through the linear interpolation algorithm 修复3 ; The final repair value in case of data loss or error

Images