Power grid load monitoring system and method based on intelligent sensor

Abstract

The invention relates to the field of power grid load monitoring, in particular to a power grid load monitoring system and method based on an intelligent sensor, and the system comprises a data collection module, a data transmission module, a storage backtracking module, a real-time monitoring module, an alarm module and an intelligent decision module. The data transmission module carries out data transmission on power grid load monitoring data, the storage backtracking module carries out storage, backtracking analysis and updating on the data, and the real-time monitoring module carries out spatial-temporal correlation analysis on the data, constructs a spatial-temporal regression model and a dynamic threshold model, and carries out real-time monitoring on the data. The alarm module performs anomaly analysis on the data and performs anomaly alarm, and the intelligent decision-making module performs intelligent decision-making on response measures of the anomaly alarm, so that rapid response to different risk conditions of the deep sea power grid is realized, and safe and stable operation of the power grid is ensured.

Description

Technical Field

[0001] The present invention relates to the technical field of power grid load monitoring, and in particular to a power grid load monitoring system and method based on intelligent sensors. Background Art

[0002] With the rapid development of social economy, the application of electric power energy in various fields is becoming increasingly widespread. As a key link in power supply, the safe and stable operation of the distribution network is crucial for ensuring the quality of power supply. The traditional monitoring methods of the distribution network mainly rely on manual inspections and regular maintenance. This method is not only inefficient but also difficult to grasp the operation status of the distribution network in real time and cannot detect potential safety hazards in a timely manner. To solve the above problems, some distribution network overload early warning and monitoring systems have emerged in the prior art. Most of the existing systems lack the functions of storing and retrieving historical data and cannot use historical data for trend analysis and fault prediction, which is not conducive to the long-term planning and optimal operation of the distribution network. Therefore, it is of great practical significance to develop a more efficient, comprehensive and intelligent distribution network overload early warning and monitoring system.

[0003] Chinese Patent Publication No.: CN109450092A discloses a distribution network overload early warning and monitoring system, but this solution still has the problem that the power grid load monitoring process is too single, only making simple threshold judgments and alarms for the power grid load, unable to comprehensively reflect the complex operation state of the power grid, difficult to detect complex hidden dangers in advance, lacking decision-making support and the function of storing and retrieving data, and unable to mine valuable information from historical data to optimize subsequent monitoring and early warning, which is not conducive to the long-term stable operation and fault prevention of the distribution network. Therefore, the prior art has obvious deficiencies in the real-time monitoring and decision-making output of the operation state of the power grid load and the storage and retrieval of data, and a more comprehensive and adaptable system is needed to solve these problems. Summary of the Invention

[0004] For this reason, the present invention provides a power grid load monitoring system and method based on intelligent sensors to overcome the problems of low efficiency of power grid load monitoring and incomplete prediction and analysis of power grid load caused by the lack of data storage, retrospective analysis and real-time monitoring of data in the prior art.

[0005] To achieve the above object, the present invention provides a power grid load monitoring system and method based on intelligent sensors, including: A data acquisition module for collecting power grid load monitoring data; A data transmission module for transmitting the power grid load monitoring data to obtain target transmission data; A storage and traceback module is used to store the target transmission data, and is also used to perform traceback analysis on the target transmission data and update the target transmission data to obtain the second target transmission data; A real-time monitoring module is used to obtain target-related data based on the second target transmission data, and is also used to construct a dynamic threshold model based on the target-related data; An alarm module is used to perform anomaly analysis on the second target transmission data, the target-related data, and the dynamic threshold output by the dynamic threshold model, and issue an anomaly warning; An intelligent decision-making module is used to make intelligent decisions on countermeasures for anomaly warnings.

[0006] Further, when the data acquisition module acquires power grid load monitoring data, the acquisition interval is once every 3 minutes to obtain the first target transmission data.

[0007] Further, when the data transmission module transmits power grid load monitoring data, the process includes: Step S1, dynamically select the transmission path of the first target transmission data using intelligent routing, and deploy redundant communication devices at key nodes during the transmission of the first target transmission data; Step S2, after dynamically selecting the transmission path of the first target transmission data, compress and encrypt the first target transmission data to obtain the second target transmission data; Step S3, converge the second target transmission data to the local data node to obtain the first node data, transmit the first node data to the offshore platform to obtain the second node data, and transmit the second node data to the onshore control center to obtain the target transmission data.

[0008] Further, when the storage and traceback module stores the target transmission data, it selects the initial storage location of the target transmission data according to the source device of the target transmission data, where: When the source device of the target transmission data is a key power generation device, select the initial storage location of the target transmission data as SSD; When the source device of the target transmission data is an auxiliary device, select the initial storage location of the target transmission data as a tape library; Obtain the usage times A of the target transmission data within a preset period according to historical operations, compare the usage times A of the target transmission data within the preset period with the preset usage frequency A0, judge the usage frequency of the target transmission data according to the comparison result, and adjust the storage location of the target transmission data according to the judgment result, where: When A ≤ A0, the storage backtracking module determines that the usage frequency of the target transmission data within the preset period is low, and adjusts the storage location of the target transmission data to the tape library; When A > A0, the storage backtracking module determines that the usage frequency of the target transmission data within the preset period is high, and adjusts the storage location of the target transmission data to the SSD; When the storage backtracking module stores the target transmission data, it obtains the importance coefficient B of the target transmission data according to expert evaluation, compares the importance coefficient B of the target transmission data with the preset importance coefficient B0, judges the importance of the target transmission data according to the comparison result, and optimizes the usage frequency A of the target transmission data according to the judgment result, where: When B ≤ B0, the storage backtracking module determines that the importance of the target transmission data is low, and does not optimize the usage frequency A of the target transmission data; When B > B0, the storage backtracking module determines that the importance of the target transmission data is high, optimizes the usage frequency A of the target transmission data, and sets the optimized usage frequency of the target transmission data as A 、 , and according to the optimized usage frequency A of the target transmission data 、 Adjust the storage location of the target transmission data with the preset usage frequency A; When the storage backtracking module stores the target transmission data, it obtains the storage time C of the target transmission data according to the time from when the system collects the target transmission data to when the target transmission data is extracted and used, compares the storage time C of the target transmission data with the preset storage time C0, judges the validity of the target transmission data according to the comparison result, and corrects the importance coefficient B of the target transmission data according to the judgment result, where When C ≤ C0, the storage backtracking module determines that the validity of the target transmission data is low, and does not correct the importance coefficient B of the target transmission data, When C > C0, the storage backtracking module determines that the validity of the target transmission data is high, corrects the importance coefficient B of the target transmission data, and sets the corrected importance coefficient of the target transmission data as B 、 , and according to the corrected importance coefficient B of the target transmission data 、 Adjust the storage location of the target transmission data with the preset importance coefficient B0.

[0009] Further, when the storage backtracking module performs backtracking analysis on the target transmission data, the process of performing backtracking analysis on the target transmission data includes: Step G01: Obtain the abnormal deep - sea power grid load reflected by the target transmission data as the target abnormal situation, and take the factors causing the target abnormal situation as the target influencing factors; Step G02: Obtain the number of occurrences D of the target influencing factor within the preset influence period. Compare the number of occurrences D of the target influencing factor within the preset influence period with the preset number of occurrences D0. Judge the occurrence frequency of the target influencing factor according to the comparison result, and make a first adjustment decision based on the judgment result, where: When D ≤ D0, in the storage and back - tracking module, it is determined that the occurrence frequency of the target influencing factor is low - frequency, and no first adjustment decision is made; When D > D0, in the storage and back - tracking module, it is determined that the occurrence frequency of the target influencing factor is high - frequency, and a first adjustment decision is made; Obtain the occupation duration W of the target influencing factor causing the target abnormal situation. Compare the occupation duration W of the target influencing factor causing the target abnormal situation with the preset occupation duration W0, where 10 minutes ≤ W0 ≤ 30 minutes. Judge the damage degree of the target influencing factor according to the comparison result, and make a second adjustment decision based on the judgment result, where: When E ≤ E0, in the storage and back - tracking module, it is determined that the damage degree of the target influencing factor is low, and no second adjustment decision is made; When E > E0, in the storage and back - tracking module, it is determined that the damage degree of the target influencing factor is high, and a second adjustment decision is made; Step G03: Make an adjustment decision according to the judgment of the occurrence frequency of the target influencing factor. Obtain the actual type of the target influencing factor, and output the first adjustment decision and the second adjustment decision as the adjustment decision according to the actual type of the target influencing factor, where: When the actual type of the target influencing factor is a human - induced influencing factor, the adjustment decision is to improve the operation mode of the operation and maintenance personnel to obtain the target normal situation; When the actual type of the target influencing factor is an environmental influencing factor, the adjustment decision is to upgrade and replace the equipment to obtain the target normal situation.

[0010] Further, when the storage and back - tracking module updates the target transmission data in the data acquisition module, it feeds back the target normal situation to the data acquisition module for secondary acquisition, transmission, storage, and back - tracking analysis to obtain the second target transmission data.

[0011] Further, when the real - time monitoring module obtains the target - related data according to the second target transmission data, it includes analyzing the time - series autocorrelation and the spatial - position autocorrelation, where: When the real-time monitoring module analyzes the time series autocorrelation of the second target transmission data, according to the time series \(x_t\), the time series length \(n\), the time interval \(k\), and the mean value of the time series set the time autocorrelation function as ACF(k), and set , calculate the time correlation coefficient \(F\) of the second target transmission data according to the time autocorrelation function ACF(k), compare the time correlation coefficient \(F\) of the second target transmission data with the preset time correlation coefficient \(F_0\), set \(0.556\leq F_0\lt1\), judge the compliance of the time correlation coefficient \(F\) of the second target transmission data according to the comparison result, and output whether to perform spatial position autocorrelation analysis on the second target transmission data according to the judgment result, where: When \(F\lt F_0\), the real-time monitoring module determines that the time correlation coefficient \(F\) of the second target transmission data meets the standard, and does not output the spatial position autocorrelation analysis of the second target transmission data; When \(F\geq F_0\), the real-time monitoring module determines that the time correlation coefficient \(F\) of the second target transmission data does not meet the standard, outputs the spatial position autocorrelation analysis of the second target transmission data, and marks the second target transmission data as time-related data; When the real-time monitoring module performs spatial position autocorrelation analysis on the time-related data, according to the first spatial unit \(i\), the second spatial unit \(j\), each spatial data sample \(x_i\), the number \(n\) of each spatial data sample, and the mean value of the time series , and the correlation variable \(x_i\) of the first spatial unit \(i\), the variable \(x_j\) related to the second spatial unit \(j\), and the spatial weight matrix \(w_{ij}\), substitute the first spatial unit \(i\) and the second spatial unit \(j\) into the spatial autocorrelation function Moran's I , to obtain the function value Moran's I, where when the spatial units \(i\) and \(j\) are adjacent, \(w_{ij}=1\); when the spatial units \(i\) and \(j\) are not adjacent, \(w_{ij}=0\). The real-time monitoring module compares Moran's I with the first preset value \(M_1\) and the second preset value \(M_2\), sets \(M_1\leq Moran's I\leq M_2\), and \(M_1\leq0\leq M_2\). Judge the difference degree of adjacent spatial position variable values of the time-related data according to the comparison result, and label the time-related data according to the judgment result, where: When \(0\lt Moran's I\leq M_2\), the real-time monitoring module determines that the difference degree of adjacent spatial position variable values of the time-related data is small, and labels the time-related data as positively correlated adjacent spatial positions to obtain the first target-related data; When \(M1\leqslant Moran's I < 0\), the real-time monitoring module determines that the difference degree of adjacent spatial position variables of the time-related data is large, marks the time-related data with negative spatial correlation of adjacent spatial positions to obtain second target-related data, and takes the first target-related data and the second target-related data as target-related data; When \(Moran's I = 0\), the real-time monitoring module determines that there is no spatial autocorrelation in the adjacent spatial positions of the time-related data, does not mark the time-related data, and removes the time-related data; When the real-time monitoring module constructs a spatio-temporal regression model according to the target-related data, set \(y_{it}\) as the target-related data of the \(i\)-th position at time \(t\), and \(x_{it}\) as the variable related to time \(t\), which is a variable related to spatial position, substitute the second target-related data into the calculation to obtain the spatio-temporal regression model ; when the real-time monitoring module obtains the target-related data according to the second target transmission data, perform three-dimensional visualization on the target-related data, use time as the \(x\)-axis, longitude as the \(y\)-axis, latitude as the \(z\)-axis, and then use the target-related data corresponding to each time point and spatial position as a data point, plot it in three-dimensional space, and represent the data points with different shades of color according to the size of the target-related data to obtain the three-dimensional visualization model of the target-related data.

[0012] Furthermore, when the real-time monitoring module constructs a dynamic threshold model according to the target-related data, 70% of the target-related data samples calculated by the spatio-temporal regression model are used as the training set, 15% as the validation set, and 15% as the test set to obtain the dynamic threshold model. Then, classify the target-related data samples according to the \(3\sigma\) criterion. The target-related data within the \(3\sigma\) criterion is used as the normal load sample, and the target-related data outside the \(3\sigma\) criterion is used as the abnormal load sample. Then, place the normal load sample into the dynamic threshold model, calculate the distance distribution of the normal load sample to the decision boundary to obtain the target output data, and then process the target output data according to the quantile method to obtain the dynamic threshold. Finally, according to the real-time operation data and risk status of the power grid, dynamically adjust the dynamic threshold in real time by the gradient descent method.

[0013] Furthermore, when the alarm module performs abnormal analysis on the second target transmission data, the target-related data and the dynamic threshold output by the dynamic threshold model and issues an abnormal alarm, compare the target-related data \(G\) obtained in the real-time monitoring module with the preset data threshold \(G_0\), judge the security degree of the target-related data according to the comparison result, and output the conclusion of whether the target-related data triggers the first alarm signal according to the judgment result, where: When G ≤ G0, the alarm module determines that the security level of the target-related data is high and does not trigger the first alarm signal; When G > G0, the alarm module determines that the security level of the target-related data is low and triggers the first alarm signal; The target-related data is input into the spatio-temporal regression model to calculate the predicted value H of the target-related data. The predicted value H of the target-related data is compared with the dynamic threshold H0 output by the dynamic threshold model. According to the comparison result, the abnormal degree of the change trend of the predicted value of the target-related data is judged, and according to the judgment result, a conclusion on whether the predicted value of the target-related data triggers the second alarm signal is output, where: When H ≤ H0, the alarm module determines that the abnormal degree of the change trend of the predicted value of the target-related data is low and does not trigger the second alarm signal; When H > H0, the alarm module determines that the abnormal degree of the change trend of the predicted value of the target-related data is high and triggers the second alarm signal.

[0014] Further, when the intelligent decision-making module makes an intelligent decision on the countermeasures for abnormal alarms, it includes: primary alarm decision-making, intermediate alarm decision-making, and advanced alarm decision-making, where: When the intelligent decision-making module makes a primary alarm decision, it compares the range G1 of the target-related data exceeding the preset data threshold with each preset range of the exceeded threshold. The preset ranges of the exceeded threshold include the first preset range of the exceeded threshold G01 and the first preset range of the exceeded threshold G02. Set G01 = 5% and G02 = 10%. According to the comparison result, the abnormal degree of the range G1 of the target-related data exceeding the preset threshold is judged, and according to the judgment result, an intelligent decision on the countermeasures for the target-related data is made, where: When G1 < G01, the intelligent decision-making module determines that the abnormal degree of the range G1 of the target-related data exceeding the preset threshold is slight. The system automatically records the alarm information and highlights it on the monitoring interface. At the same time, a reminder message is sent to the operation and maintenance personnel, suggesting that the operation and maintenance personnel closely monitor the changes in relevant indicators and do not take emergency measures for the time being; When G01 ≤ G1 ≤ G02, the intelligent decision-making module determines that the abnormal degree of the range G1 of the target-related data exceeding the preset threshold is medium. The system automatically records the alarm information and highlights it on the monitoring interface. At the same time, a preliminary inspection is automatically started.

[0015] When G1 > G02, the intelligent decision-making module determines that the degree of abnormality of the target-related data exceeding the preset threshold range G1 is serious. The system automatically cuts off the relevant circuit, starts the backup power supply, quickly notifies the relevant senior management and professional maintenance team, and at the same time provides detailed fault information and emergency handling suggestions. In combination with the storage and backtracking module, the device with the abnormality is set as a medium-level alarm situation.

[0016] Compared with the prior art, the beneficial effects of the present invention are as follows: by performing data storage and backtracking analysis on the data and performing real-time monitoring on the data, various situations in the power grid operation are dealt with, thereby improving the efficiency of power grid load monitoring and comprehensively analyzing the power grid load prediction. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] Figure 1 It is a schematic structural diagram of the power grid load monitoring system based on intelligent sensors in this embodiment. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0018] In order to make the objectives and advantages of the present invention clearer, the present invention will be further described below in conjunction with embodiments; it should be understood that the specific embodiments described herein are only used to explain the present invention and are not used to limit the present invention.

[0019] The preferred embodiments of the present invention will be described below with reference to the accompanying drawings. Those skilled in the art should understand that these embodiments are only used to explain the technical principles of the present invention and do not limit the protection scope of the present invention.

[0020] In addition, it should be noted that in the description of the present invention, unless otherwise clearly specified and limited, the terms "installation", "connection", and "connection" should be understood in a broad sense. For example, it can be a fixed connection, a detachable connection, or an integral connection; it can be a mechanical connection or an electrical connection; it can be directly connected or indirectly connected through an intermediate medium, and it can be the communication inside two components. For those skilled in the art, the specific meanings of the above terms in the present invention can be understood according to specific situations.

[0021] Please refer to Figure 1 as shown, which is a schematic structural diagram of the power grid load monitoring system based on intelligent sensors in this embodiment. The system includes: A data acquisition module for acquiring power grid load monitoring data; A data transmission module for transmitting the power grid load monitoring data to obtain target transmission data. The data transmission module is connected to the data acquisition module; A storage and traceback module is used to store the target transmission data, and is also used to perform traceback analysis on the target transmission data and update the target transmission data to obtain second target transmission data. The storage and traceback module is connected to the data transmission module; A real-time monitoring module is used to obtain target-related data based on the second target transmission data, and is also used to construct a dynamic threshold model based on the target-related data. The real-time monitoring module is connected to the storage and traceback module; An alarm module is used to perform anomaly analysis on the second target transmission data, the target-related data, and the dynamic threshold output by the dynamic threshold model, and perform anomaly warning. The alarm module is connected to the real-time monitoring module; An intelligent decision-making module is used to make intelligent decisions on the countermeasures for anomaly warnings. The intelligent decision-making module is connected to the alarm module.

[0022] Specifically, the power grid load monitoring system based on intelligent sensors is applied to data monitoring of deep-sea power grids under extreme weather conditions. It is set in the terminal of the deep-sea power grid data monitoring device, such as the offshore power generation management terminal. The operation data of each power generation device and the environmental data are collected through sensors, and then the collected data is stored and retrospectively analyzed, so as to make timely countermeasures for the risks occurring in the deep-sea power grid under extreme weather conditions. Among them, the power grid load monitoring system based on intelligent sensors uses the data acquisition module to collect power generation device data, transmission device data, power grid operation data, and environmental data through sensors, improving the reliability of the power grid load monitoring system based on intelligent sensors, ensuring the safety of personnel and equipment, and making the power grid load monitoring system operate more efficiently. The power grid load monitoring system based on intelligent sensors realizes full-coverage data transmission from the deep sea to the land through the data transmission module. The power grid load monitoring system based on intelligent sensors stores and analyzes the monitoring data through the storage and traceback module, providing a reference for subsequent power grid planning and operation and maintenance management. The power grid load monitoring system based on intelligent sensors monitors the power grid load situation of the deep-sea power grid under extreme weather conditions through the real-time monitoring module, facilitating subsequent analysis. The power grid load monitoring system based on intelligent sensors makes alarms for situations with different risk levels through the alarm module, and finally transmits them to the intelligent decision-making module to make corresponding alarm decisions, timely coping with the major risks of the deep-sea power grid under extreme weather conditions.

[0023] Specifically, when the data acquisition module collects power grid load monitoring data, the collection interval is once every 3 minutes to obtain first target transmission data.

[0024] Specifically, the power grid load monitoring data refers to a series of data related to the power grid load collected through various monitoring devices and technical means during the operation of the power system, including the operation data of each power generation device and environmental data. In this embodiment, the operation data of each power generation device is not limited, and relevant technical personnel in the field can freely set it according to actual needs, as long as it meets the power grid data generated during the operation of the power generation device. For example, a power sensor is used to collect the power data of the device, and a speed sensor is used to collect the speed data of the device. In this embodiment, the environmental data is not limited, and relevant technical personnel in the field can freely set it according to actual needs, as long as it meets the data of the environment around the power generation device during the operation of the power generation device. For example, a current meter is used to collect the current speed and direction data, and an anemometer is used to collect the wind speed and direction data.

[0025] Specifically, the data acquisition module collects the operation data and environmental data of the deep-sea power grid through sensors, and can monitor the key parameters in the deep-sea power grid in real time, so as to facilitate the timely discovery of possible problems such as overload or insulation failure. For the obtained data, the usage frequency, importance, and data storage time are determined, which is convenient for subsequent classification storage and retrospective analysis of the data, thereby improving the data monitoring efficiency of the deep-sea power grid in extreme weather conditions.

[0026] Specifically, when the data transmission module transmits the power grid load monitoring data, the process includes: Step S1, dynamically select the transmission path of the first target transmission data using an intelligent router, and deploy redundant communication devices at the key nodes of the first target transmission data transmission; Step S2, after dynamically selecting the transmission path of the first target transmission data, compress and encrypt the first target transmission data to obtain the second target transmission data; Step S3, converge the second target transmission data to the local data node to obtain the first node data, transmit the first node data to the offshore platform to obtain the second node data, and transmit the second node data to the onshore control center to obtain the target transmission data.

[0027] Specifically, the intelligent routing refers to an algorithm used in network communication to dynamically select the optimal transmission path based on the real-time state of the network and various factors. The redundant communication device refers to an additional backup communication device in the communication system. The compression and encryption refer to the method of performing Huffman coding compression on data, using an asymmetric encryption algorithm to exchange the symmetric encryption key for data, and then using a symmetric encryption algorithm to encrypt the target transmission data. The power generation device refers to a device used in the deep-sea environment to convert other forms of energy into electrical energy. The local data acquisition node refers to a device responsible for collecting various types of data within a specific area, performing preliminary processing and aggregation. The offshore platform refers to a large engineering structure built and used in the ocean for data reception. The onshore control center refers to a facility for centrally monitoring, managing, and operating various remote devices. In this embodiment, the dynamic selection is not limited, and those skilled in the relevant art can freely set it according to actual needs, as long as it meets the requirement of flexibly selecting the transmission path according to actual needs. For example, when the submarine optical cable fails, it can be automatically adjusted to satellite communication. The redundant communication device refers to an additional backup communication device in the communication system. In this embodiment, the transmission path is not limited, and those skilled in the relevant art can freely set it according to actual needs, as long as it meets the requirement of transmitting data, such as satellite communication and submarine optical cable.

[0028] Specifically, in the data transmission module, intelligent routing and redundancy design, data compression and encryption, and a hierarchical communication architecture are used to transmit and protect the collected target transmission data in sequence, and reduce the latency and ensure the efficiency of important information in the target transmission data. The intelligent routing and redundancy design improve the network performance and efficiency by dynamically selecting the optimal path and encrypting the data, and at the same time improve the reliability, stability, and fault tolerance of the system.

[0029] Specifically, when the storage and traceback module stores the target transmission data, the initial storage location of the target transmission data is selected according to the source device of the target transmission data, where: When the source device of the target transmission data is a key power generation device, the initial storage location of the target transmission data is selected as an SSD; When the source device of the target transmission data is an auxiliary device, the initial storage location of the target transmission data is selected as a tape library; Obtain the usage times A of the target transmission data within a preset period according to historical operations, compare the usage times A of the target transmission data within the preset period with the preset usage frequency A0, judge the usage frequency of the target transmission data according to the comparison result, and adjust the storage location of the target transmission data according to the judgment result, where: When A ≤ A0, the storage backtracking module determines that the usage frequency of the target transmission data within the preset period is low, and adjusts the storage location of the target transmission data to the tape library; When A > A0, the storage backtracking module determines that the usage frequency of the target transmission data within the preset period is high, and adjusts the storage location of the target transmission data to the SSD; When the storage backtracking module stores the target transmission data, it obtains the importance coefficient B of the target transmission data according to expert evaluation, compares the importance coefficient B of the target transmission data with the preset importance coefficient B0, judges the importance of the target transmission data according to the comparison result, and optimizes the usage frequency A of the target transmission data according to the judgment result, where: When B ≤ B0, the storage backtracking module determines that the importance of the target transmission data is low, and does not optimize the usage frequency A of the target transmission data; When B > B0, the storage backtracking module determines that the importance of the target transmission data is high, optimizes the usage frequency A of the target transmission data, and sets the optimized usage frequency of the target transmission data to A 、 , and according to the optimized usage frequency A of the target transmission data 、 adjusts the storage location of the target transmission data with the preset usage frequency A; When the storage backtracking module stores the target transmission data, it obtains the storage time C of the target transmission data according to the time from when the system collects the target transmission data to when the target transmission data is extracted and used, compares the storage time C of the target transmission data with the preset storage time C0, judges the validity of the target transmission data according to the comparison result, and corrects the importance coefficient B of the target transmission data according to the judgment result, where When C ≤ C0, the storage backtracking module determines that the validity of the target transmission data is low, and does not correct the importance coefficient B of the target transmission data. When C > C0, the storage backtracking module determines that the validity of the target transmission data is high, corrects the importance coefficient B of the target transmission data, and sets the corrected importance coefficient of the target transmission data to B 、 , and according to the corrected importance coefficient B of the target transmission data 、 adjusts the storage location of the target transmission data with the preset importance coefficient B0; Specifically, according to the evaluation by power grid experts, the correlation degree U between the target device and the core operating functions of the power grid is obtained. The correlation degree U between the target device and the core operating functions of the power grid is compared with the preset correlation degree U0. According to the comparison result, the criticality of the target device is judged, and the function definition of the target device is output according to the judgment result, where: When U ≤ U0, it is determined that the criticality of the device is low, and the function of the target device is defined as an auxiliary device; When U > U0, it is determined that the criticality of the device is high, and the function of the target device is defined as a key power generation device; The power grid expert evaluation refers to a system in which power grid information is pushed to expert users with the ability to set power grid operation plans for their terminals, so as to obtain power grid operation knowledge, experience and research results. The tape library refers to a system for storing and managing tapes, and the SSD refers to a non-volatile storage device based on flash chips.

[0030] Specifically, when the storage and backtracking module stores the target transmission data, it reduces the storage cost and ensures the high efficiency of data retrieval, so as to facilitate the subsequent backtracking analysis of the target transmission data and improve the deep-sea power grid data monitoring efficiency.

[0031] Specifically, when the storage and backtracking module performs backtracking analysis on the target transmission data, the process of performing backtracking analysis on the target transmission data includes: Step G01, obtaining the deep-sea power grid load anomaly reflected by the target transmission data as the target abnormal situation, and taking the factors causing the target abnormal situation as the target influencing factors; Step G02, obtaining the number of occurrences D of the target influencing factors within the preset influence period, comparing the number of occurrences D of the target influencing factors within the preset influence period with the preset number of occurrences D0, judging the occurrence frequency of the target influencing factors according to the comparison result, and making a first adjustment decision according to the judgment result, where: When D ≤ D0, the storage and backtracking module determines that the occurrence frequency of the target influencing factor is low frequency and does not make a first adjustment decision; When D > D0, the storage and backtracking module determines that the occurrence frequency of the target influencing factor is high frequency and makes a first adjustment decision; Obtain the occupation duration W of the target influencing factor causing the target abnormal situation, compare the occupation duration W of the target influencing factor causing the target abnormal situation with the preset occupation duration W0, set 10 minutes ≤ W0 ≤ 30 minutes, judge the damage degree of the target influencing factor according to the comparison result, and make a second adjustment decision according to the judgment result, where: When E ≤ E0, the storage and backtracking module determines that the damage degree of the target influencing factor is low and does not make a second adjustment decision; When E > E0, the storage and backtracking module determines that the damage degree of the target influencing factor is high and makes a second adjustment decision. Step G03: Make an adjustment decision based on the judgment of the occurrence frequency of the target influencing factor, obtain the actual type of the target influencing factor, and output the first adjustment decision and the second adjustment decision as the adjustment decision according to the actual type of the target influencing factor, where: When the actual type of the target influencing factor is a human influencing factor, the adjustment decision is to improve the operation mode of the operation and maintenance personnel to obtain the target normal situation. When the actual type of the target influencing factor is an environmental influencing factor, the adjustment decision is to upgrade and replace the equipment to obtain the target normal situation. When the storage and backtracking module updates the target transmission data in the data acquisition module, it feeds back the target normal situation to the data acquisition module for secondary acquisition, transmission, storage, and backtracking analysis to obtain the second target transmission data.

[0032] Specifically, this embodiment does not limit the personnel factor. Those skilled in the relevant art can freely set it according to actual needs, as long as it satisfies the continuous deterioration of the target load abnormal situation caused by human reasons, such as improper operations of power grid operation and maintenance personnel and dispatchers during extreme weather, and untimely deployment of prevention and repair responses. This embodiment includes but is not limited to the limitation of the equipment factor. Those skilled in the relevant art can freely set it according to actual needs, as long as it satisfies the continuous deterioration of the target load abnormal situation caused by equipment reasons, such as the moisture in the generator winding affecting the insulation performance and the swing of the line caused by strong winds. This embodiment does not limit the abnormal load of the deep-sea power grid. Those skilled in the relevant art can freely set it according to actual needs, as long as it satisfies the situation that the power consumption load of the power grid in the deep-sea area deviates from the normal state, such as the abnormal power consumption load caused by natural disasters damaging the power grid facilities. This embodiment does not limit the preset influence period. Those skilled in the relevant art can freely set it according to actual needs, as long as it satisfies the need for time limitation, such as setting two days as a preset influence period. The judgment of the occurrence frequency of the target influencing factor refers to the result of the backtracking analysis of the target transmission data.

[0033] Specifically, the storage and backtracking module can quickly determine the type of the target influencing factor affecting the target abnormal situation and make a decision for adjustment in time by classifying and storing the target transmission data and performing backtracking analysis on the stored target transmission data, so as to avoid the continuous deterioration of the abnormal power grid load.

[0034] Specifically, when the real-time monitoring module obtains the target-related data according to the second target transmission data, it includes the analysis of time series autocorrelation and the analysis of spatial position autocorrelation, where: When the real-time monitoring module analyzes the time series autocorrelation of the second target transmission data, according to the time series xt, the time series length n, the time interval k, and the mean value of the time series Set the time autocorrelation function as ACF(k), and set , calculate the time correlation coefficient F of the second target transmission data according to the time autocorrelation function ACF(k), compare the time correlation coefficient F of the second target transmission data with the preset time correlation coefficient F0, set 0.556 ≤ F0 < 1, judge the compliance of the time correlation coefficient F of the second target transmission data according to the comparison result, and output whether to perform spatial autocorrelation analysis on the second target transmission data according to the judgment result, where: When F < F0, the real-time monitoring module determines that the time correlation coefficient F of the second target transmission data meets the standard, and does not output the spatial autocorrelation analysis of the second target transmission data; When F ≥ F0, the real-time monitoring module determines that the time correlation coefficient F of the second target transmission data does not meet the standard, outputs the spatial autocorrelation analysis of the second target transmission data, and marks the second target transmission data as time-related data; When the real-time monitoring module performs spatial autocorrelation analysis on time-related data, according to the first spatial unit i, the second spatial unit j, each spatial data sample xi, the number n of each spatial data sample, and the mean value of the time series , the correlation variable xi of the first spatial unit i, the variable xj related to the second spatial unit j, and the spatial weight matrix wij, substitute the first spatial unit i and the second spatial unit j into the spatial autocorrelation function Moran's I , and obtain the function value Moran's I. When the spatial units i and j are adjacent, wij = 1; when the spatial units i and j are not adjacent, wij = 0. The real-time monitoring module compares Moran's I with the first preset value M1 and the second preset value M2, sets M1 ≤ Moran's I ≤ M2, and M1 ≤ 0 ≤ M2, judges the difference degree of adjacent spatial position variable values of time-related data according to the comparison result, and marks the time-related data according to the judgment result, where: When 0 < Moran's I ≤ M2, the real-time monitoring module determines that the difference degree of adjacent spatial position variable values of the time-related data is small, and marks the time-related data as positively correlated with adjacent spatial positions to obtain the first target-related data; When M1 ≤ Moran's I < 0, the real-time monitoring module determines that the difference degree of adjacent spatial position variables of the time-related data is large, marks the time-related data with negative spatial correlation of adjacent spatial positions to obtain second target-related data, and uses the first target-related data and the second target-related data as target-related data; When Moran's I = 0, the real-time monitoring module determines that there is no spatial autocorrelation in the adjacent spatial positions of the time-related data, does not mark the time-related data, and removes the time-related data; When the real-time monitoring module constructs a spatio-temporal regression model based on the target-related data, set yit as the target-related data of the i-th position at time t, and xit as the variable related to time t, as the variable related to the spatial position, substitute the second target-related data into the calculation to obtain the spatio-temporal regression model ; When the real-time monitoring module obtains the target-related data according to the second target transmission data, perform three-dimensional visualization on the target-related data. Use time as the x-axis, longitude as the y-axis, and latitude as the z-axis. Then, use the target-related data corresponding to each time point and spatial position as a data point, draw it in three-dimensional space, and represent the data points with different shades of color according to the size of the target-related data to obtain the three-dimensional visualization model of the target-related data.

[0035] Specifically, when the real-time monitoring module constructs a dynamic threshold model based on the target-related data, 70% of the target-related data samples after being calculated by the spatio-temporal regression model are used as the training set, 15% as the validation set, and 15% as the test set to obtain the dynamic threshold model. Then, classify the target-related data samples according to the 3σ criterion. The target-related data within the 3σ criterion is used as the normal load sample, and the target-related data outside the 3σ criterion is used as the abnormal load sample. Then, place the normal load sample into the dynamic threshold model, calculate the distance distribution of the normal load sample to the decision boundary to obtain the target output data, and then process the target output data according to the quantile method to obtain the dynamic threshold. Finally, according to the real-time operation data and risk status of the power grid, dynamically adjust the dynamic threshold in real time through the gradient descent method.

[0036] Specifically, the time series autocorrelation refers to the correlation existing between the observed values of time series data at different time points. The spatial location autocorrelation refers to the correlation existing between the observed value at a certain position in space and the observed values at other positions. The spatial unit refers to the basic unit used to divide and define space in spatial analysis and research. The spatial data sample refers to a representative data subset extracted from the overall spatial data. The spatial data sample refers to a data sample with spatial location information sampled from the spatial dataset. The spatial weight matrix refers to a rectangular array used to describe the mutual relationship between spatial objects. The Moran's I is a statistical index used to measure the autocorrelation of spatial data. The spatio-temporal regression model refers to a statistical model used to analyze data with spatio-temporal characteristics. The training set refers to a data subset selected from the original data and specifically used for the training process of the model. The validation set refers to the data set used to assist in adjusting the hyperparameters of the model and evaluating the model performance during the model training process. The test set refers to the data set used to finally evaluate the model performance and generalization ability after the model training and tuning are completed. The 3σ criterion refers to the outlier determination criterion based on the normal distribution. The decision boundary refers to the boundary or surface that separates data of different categories. The quantile method refers to the value at a specific position after the data is arranged in ascending order. The gradient descent method refers to an optimization algorithm used to solve the minimum value of the objective function.

[0037] Specifically, the real-time monitoring module accurately screens the second target transmission data through the method of correlation function analysis, ensures the effectiveness of the second target transmission data, and establishes a spatio-temporal regression model that can be used to predict the load value. Finally, a dynamic threshold model is established to obtain a dynamic threshold, effectively reducing the false alarms of the power grid load situation during extreme weather.

[0038] Specifically, when the alarm module performs anomaly analysis on the second target transmission data, the target-related data, and the dynamic threshold output by the dynamic threshold model and issues an anomaly alarm, the target-related data G obtained from the real-time monitoring module is compared with the preset data threshold G0. According to the comparison result, the security level of the target-related data is judged, and according to the judgment result, a conclusion on whether the target-related data triggers the first alarm signal is output, where: When G ≤ G0, the alarm module determines that the security level of the target-related data is high and does not trigger the first alarm signal; When G > G0, the alarm module determines that the security level of the target-related data is low and triggers the first alarm signal; Bring the target-related data into the spatio-temporal regression model to calculate the predicted value H of the target-related data. Compare the predicted value H of the target-related data with the dynamic threshold H0 output by the dynamic threshold model. Judge the abnormal degree of the change trend of the predicted value of the target-related data according to the comparison result, and output the conclusion of whether the predicted value of the target-related data triggers the second alarm signal according to the judgment result, where: When H ≤ H0, the alarm module determines that the abnormal degree of the change trend of the predicted value of the target-related data is low and does not trigger the second alarm signal; When H > H0, the alarm module determines that the abnormal degree of the change trend of the predicted value of the target-related data is high and triggers the second alarm signal; Obtain the second target transmission data L according to the storage and backtracking module. Compare the second target transmission data L with the preset second target transmission data threshold L0. Judge the potential risk of the second target transmission data according to the comparison result, and output the conclusion of whether the second target transmission data triggers the third alarm signal according to the judgment result, where: When L ≤ L0, the alarm module determines that the potential risk of the second target transmission data is too low and does not trigger the third alarm signal; When L ≤ L0, the alarm module determines that the potential risk of the second target transmission data is too high and triggers the third alarm signal.

[0039] Specifically, this embodiment does not limit the primary alarm signal. Those skilled in the relevant art can freely set it according to actual needs, as long as it satisfies that when the emergency coefficient of the target-related data exceeds the preset threshold, it generates an alarm signal different from the intermediate alarm signal and the high-level alarm signal. For example, a loud sound and a flashing light signal are emitted by the sound and light alarm. This embodiment does not limit the intermediate alarm signal. Those skilled in the relevant art can freely set it according to actual needs, as long as it satisfies that when the emergency coefficient of the predicted value of the target-related data exceeds the preset threshold, it generates an alarm signal different from the primary alarm signal and the high-level alarm signal, such as popping up a warning prompt box. This embodiment does not limit the primary alarm signal. Those skilled in the relevant art can freely set it according to actual needs, as long as it satisfies that when the emergency coefficient of the second target transmission data exceeds the preset threshold, it generates an alarm signal different from the primary alarm signal and the intermediate alarm signal, such as automatically triggering an emergency meeting notification mechanism.

[0040] Specifically, the alarm module classifies and alarms the detection results of the storage and backtracking module and the real-time monitoring module, which is convenient for quickly judging the risk level according to different detection results, so as to take correct countermeasures to ensure the safety of deep-sea power grid data monitoring.

[0041] Specifically, when the intelligent decision-making module makes intelligent decisions on the countermeasures for abnormal alarms, it includes: primary alarm decision-making, intermediate alarm decision-making, and advanced alarm decision-making, where: When the intelligent decision-making module makes a primary alarm decision, it compares the range G1 where the target-related data exceeds the preset data threshold with each preset range of the exceeded threshold. The preset ranges of the exceeded threshold include the first preset range of the exceeded threshold G01 and the first preset range of the exceeded threshold G02. Set G01 = 5% and G02 = 10%. It judges the abnormality degree of the range G1 where the target-related data exceeds the preset threshold according to the comparison result, and makes an intelligent decision on the countermeasures for the target-related data according to the judgment result, where: When G1 < G01, the intelligent decision-making module determines that the abnormality degree of the range G1 where the target-related data exceeds the preset threshold is slight. The system automatically records the alarm information, highlights it on the monitoring interface, and sends a reminder message to the operation and maintenance personnel, suggesting that the operation and maintenance personnel closely monitor the changes of relevant indicators and do not take emergency measures for the time being; When G01 ≤ G1 ≤ G02, the intelligent decision-making module determines that the abnormality degree of the range G1 where the target-related data exceeds the preset threshold is medium. The system automatically records the alarm information, highlights it on the monitoring interface, and automatically starts a preliminary inspection.

[0042] When G1 > G02, the intelligent decision-making module determines that the abnormality degree of the range G1 where the target-related data exceeds the preset threshold is serious. The system automatically cuts off the relevant circuit, starts the standby power supply, quickly notifies the relevant senior management and professional maintenance team, provides detailed fault information and emergency handling suggestions at the same time, and sets the device with abnormalities as an intermediate alarm situation in combination with the storage and backtracking module; When the intelligent decision-making module makes an intermediate alarm decision, it compares the time T when the predicted value of the target-related data exceeds the dynamic threshold range output by the dynamic threshold model with each preset dynamic threshold range time. The preset dynamic threshold range times include the first preset dynamic threshold range time T1 and the second preset dynamic threshold range time T2. Set T1 = 1h and T2 = 3h. It judges the risk degree of the time T when the predicted value of the target-related data exceeds the dynamic threshold range output by the dynamic threshold model according to the comparison result, and makes an intelligent decision on the countermeasures for the predicted value of the target-related data according to the judgment result, where: When T < T1, the time T when the intelligent decision-making module determines that the predicted value of the target-related data exceeds the dynamic threshold range output by the dynamic threshold model has a low risk level. The system generates a trend analysis report, which details the change trend of the indicators, the prediction results, and the possible influencing factors, and sends the trend analysis report to the operation and maintenance personnel and relevant technical personnel, and recommends regular monitoring and maintenance of the relevant equipment to make preparations in advance; When T1 ≤ T ≤ T2, the time T when the intelligent decision-making module determines that the predicted value of the target-related data exceeds the dynamic threshold range output by the dynamic threshold model has a medium risk level. The system generates a trend analysis report and formulates a preventive adjustment plan; When T > T2, the time T when the intelligent decision-making module determines that the predicted value of the target-related data exceeds the dynamic threshold range output by the dynamic threshold model has a high risk level. The system immediately arranges for the shutdown and maintenance of the equipment in advance, allocates standby equipment to run, notifies relevant personnel to take emergency measures, and combines with the storage and backtracking module to set the equipment with anomalies as a high-level alarm situation; When the intelligent decision-making module makes a high-level alarm decision, it obtains the correlation degree L1 between the second target transmission data and other factors and compares it with each preset correlation degree. The preset correlation degrees include the first preset correlation degree L01 and the second preset correlation degree L02 for comparison. Set L01 = 0.68 and L02 = 1. According to the comparison results, judge the comprehensive risk of the correlation degree L1 between the second target transmission data and other factors, and make an intelligent decision on the response measures for the second target transmission data according to the judgment results, where: When L1 < L01, the intelligent decision-making module determines that the comprehensive risk of the correlation degree L1 between the second target transmission data and other factors is low, and the system proposes optimization suggestions; When L01 ≤ L1 ≤ L02, the intelligent decision-making module determines that the comprehensive risk of the correlation degree L1 between the second target transmission data and other factors is medium, and the system proposes a risk control plan; When L1 > L02, the intelligent decision-making module determines that the comprehensive risk of the correlation degree L1 between the second target transmission data and other factors is serious, and the system immediately activates the highest-level emergency response mechanism to fully mobilize resources for processing.

[0043] Specifically, this embodiment does not limit the optimization suggestions, and those skilled in the relevant art can freely set them according to actual needs, as long as the need to reduce potential risks is met. For example, optimizing the operation strategy of the power grid and strengthening the maintenance management of equipment. This embodiment does not limit the risk control scheme, and those skilled in the relevant art can freely set it according to actual needs, as long as the need to improve the stability and risk resistance of the power grid is met. For example, adjusting the topological structure of the power grid and adding reactive power compensation devices. This embodiment does not limit the highest-level emergency response mechanism, and those skilled in the relevant art can freely set it according to actual needs, as long as the need to cope with major power grid risks is met. For example, large-scale load shedding, emergency allocation of power generation resources, and coordination with surrounding power grids for support, etc.

[0044] Specifically, the intelligent decision-making module classifies different security risks and performs corresponding decision-making processing for different security risks, which can quickly respond to different risk situations of the deep-sea power grid, prevent and repair in a timely manner, avoid the occurrence of potential failures, and ensure the safe and stable operation of the power grid.

[0045] So far, the technical solution of the present invention has been described in conjunction with the preferred embodiments shown in the drawings. However, it is easy for those skilled in the art to understand that the protection scope of the present invention is obviously not limited to these specific embodiments. Without departing from the principle of the present invention, those skilled in the art can make equivalent changes or substitutions to the relevant technical features, and the technical solutions after these changes or substitutions will all fall within the protection scope of the present invention.

Claims

1. A power grid load monitoring system based on intelligent sensors, characterized in that, Including: A data acquisition module for acquiring power grid load monitoring data; A data transmission module for transmitting the power grid load monitoring data to obtain target transmission data; A storage and traceback module for storing the target transmission data, and also for performing traceback analysis on the target transmission data and updating the target transmission data to obtain second target transmission data; A real-time monitoring module for obtaining target-related data based on the second target transmission data, and also for constructing a dynamic threshold model based on the target-related data; An alarm module for performing anomaly analysis on the second target transmission data, the target-related data, and the dynamic threshold output by the dynamic threshold model, and for giving anomaly alarms; An intelligent decision-making module for making intelligent decisions on countermeasures for anomaly alarms.

2. The power grid load monitoring system based on intelligent sensors according to claim 1, wherein, When the data acquisition module acquires power grid load monitoring data, the acquisition interval is once every 3 minutes to obtain first target transmission data.

3. The power grid load monitoring system based on intelligent sensors according to claim 1, characterized in that, When the data transmission module transmits the power grid load monitoring data, its process includes: Step S1, dynamically selecting the transmission path of the first target transmission data using an intelligent router, and deploying redundant communication devices at key nodes during the transmission of the first target transmission data; Step S2, after dynamically selecting the transmission path of the first target transmission data, compressing and encrypting the first target transmission data to obtain second target transmission data; Step S3, aggregating the second target transmission data to a local data node to obtain first node data, transmitting the first node data to an offshore platform to obtain second node data, and transmitting the second node data to a land control center to obtain target transmission data.

4. The power grid load monitoring system based on intelligent sensors according to claim 1, wherein, When the storage and traceback module stores the target transmission data, it selects the initial storage location of the target transmission data according to the source device of the target transmission data, where: When the source device of the target transmission data is a key power generation device, the initial storage location of the target transmission data is selected as an SSD; When the source device of the target transmission data is an auxiliary device, the initial storage location of the target transmission data is selected as a tape library; Obtaining the usage times A of the target transmission data within a preset period according to historical operations, comparing the usage times A of the target transmission data within the preset period with a preset usage frequency A0, judging the usage frequency of the target transmission data according to the comparison result, and adjusting the storage location of the target transmission data according to the judgment result, where: When A ≤ A0, the storage and traceback module determines that the usage frequency of the target transmission data within the preset period is low frequency, and adjusts the storage location of the target transmission data to a tape library; When A > A0, the storage and traceback module determines that the usage frequency of the target transmission data within the preset period is high frequency, and adjusts the storage location of the target transmission data to an SSD; When storing the target transmission data, the storage backtracking module obtains the importance coefficient B of the target transmission data according to expert evaluation, compares the importance coefficient B of the target transmission data with the preset importance coefficient B0, judges the importance of the target transmission data according to the comparison result, and optimizes the usage frequency A of the target transmission data according to the judgment result, where: When B ≤ B0, the storage backtracking module determines that the importance of the target transmission data is low and does not optimize the usage frequency A of the target transmission data; When B > B0, the storage backtracking module determines that the importance of the target transmission data is high, optimizes the usage frequency A of the target transmission data, and sets the optimized usage frequency of the target transmission data to A 、 , and according to the optimized usage frequency A of the target transmission data 、 adjusts the storage location of the target transmission data with the preset usage frequency A; When storing the target transmission data, the storage backtracking module obtains the storage time C of the target transmission data according to the time from when the system collects the target transmission data to when the target transmission data is extracted and used, compares the storage time C of the target transmission data with the preset storage time C0, judges the validity of the target transmission data according to the comparison result, and corrects the importance coefficient B of the target transmission data according to the judgment result, where When C ≤ C0, the storage backtracking module determines that the validity of the target transmission data is low and does not correct the importance coefficient B of the target transmission data. When C > C0, the storage backtracking module determines that the validity of the target transmission data is high, modifies the importance coefficient B of the target transmission data, and sets the importance coefficient of the target transmission data after modification to B 、 , and adjusts the storage location of the target transmission data according to the importance coefficient B 、 of the target transmission data after modification and the preset importance coefficient B0.

5. The power grid load monitoring system based on intelligent sensors according to claim 4, wherein When the storage backtracking module performs backtracking analysis on the target transmission data, the process of performing backtracking analysis on the target transmission data includes: Step G01: Obtain the abnormal deep-sea power grid load reflected by the target transmission data as the target abnormal situation, and take the factors causing the target abnormal situation as the target influencing factors; Step G02: Obtain the number of occurrences D of the target influencing factors within the preset influence period, compare the number of occurrences D of the target influencing factors within the preset influence period with the preset number of occurrences D0, judge the occurrence frequency of the target influencing factors according to the comparison result, and make a first adjustment decision according to the judgment result, where: When D ≤ D0, the storage backtracking module determines that the occurrence frequency of the target influencing factors is low-frequency and does not make a first adjustment decision; When D > D0, the storage backtracking module determines that the occurrence frequency of the target influencing factors is high-frequency and makes a first adjustment decision; Obtain the occupation duration W of the target influencing factors causing the target abnormal situation, compare the occupation duration W of the target influencing factors causing the target abnormal situation with the preset occupation duration W0, set 10 minutes ≤ W0 ≤ 30 minutes, judge the damage degree of the target influencing factors according to the comparison result, and make a second adjustment decision according to the judgment result, where: When E ≤ E0, the storage backtracking module determines that the damage degree of the target influencing factors is low and does not make a second adjustment decision; When E > E0, the storage backtracking module determines that the damage degree of the target influencing factors is high and makes a second adjustment decision; Step G03: Make an adjustment decision according to the judgment of the occurrence frequency of the target influencing factors, obtain the actual type of the target influencing factors, and output the first adjustment decision and the second adjustment decision as the adjustment decision according to the actual type of the target influencing factors, where: When the actual type of the target influencing factor is a human influencing factor, the adjustment decision is to improve the operation mode of the operation and maintenance personnel to obtain the target normal situation; When the actual type of the target influencing factor is an environmental influencing factor, the adjustment decision is to upgrade and replace the equipment to obtain the target normal situation.

6. The power grid load monitoring system based on intelligent sensors according to claim 5, wherein When the storage and backtracking module updates the target transmission data in the data acquisition module, it feeds back the target normal situation to the data acquisition module for secondary acquisition, transmission, storage, and backtracking analysis to obtain the second target transmission data.

7. The power grid load monitoring system based on intelligent sensors according to claim 1, characterized in that When the real-time monitoring module obtains the target-related data based on the second target transmission data, it includes analyzing the time series autocorrelation and the spatial position autocorrelation. Among them: When the real-time monitoring module analyzes the time series autocorrelation of the second target transmission data, according to the time series xt, the time series length n, the time interval k, and the mean value of the time series Set the time autocorrelation function to ACF(k), and set , calculate the time correlation coefficient F of the second target transmission data according to the time autocorrelation function ACF(k), compare the time correlation coefficient F of the second target transmission data with the preset time correlation coefficient F0, set 0.556 ≤ F0 < 1, judge the compliance of the time correlation coefficient F of the second target transmission data according to the comparison result, and output whether to perform spatial position autocorrelation analysis on the second target transmission data according to the judgment result, where: When F < F0, the real-time monitoring module determines that the time correlation coefficient F of the second target transmission data meets the standard, and does not perform spatial position autocorrelation analysis on the second target transmission data for output; When F ≥ F0, the real-time monitoring module determines that the time correlation coefficient F of the second target transmission data does not meet the standard, performs spatial position autocorrelation analysis on the second target transmission data for output, and marks the second target transmission data as time-related data; When the real-time monitoring module performs spatial autocorrelation analysis on time-related data, according to the first spatial unit i, the second spatial unit j, each spatial data sample xi, the number n of each spatial data sample, and the mean value of the time series , the correlation variable xi of the first spatial unit i, the variable xj related to the second spatial unit j, and the spatial weight matrix wij, substitute the first spatial unit i and the second spatial unit j into the spatial autocorrelation function Moran's I , to obtain the function value Moran's I, where when the spatial units i and j are adjacent, wij = 1; when the spatial units i and j are not adjacent, wij = 0. The real-time monitoring module compares Moran's I with the first preset value M1 and the second preset value M2, sets M1 ≤ Moran's I ≤ M2, and M1 ≤ 0 ≤ M2. According to the comparison result, judge the difference degree of the adjacent spatial position variable values of the time-related data, and label the time-related data according to the judgment result, where: When 0 < Moran's I ≤ M2, the real-time monitoring module determines that the difference degree of adjacent spatial position variable values of the time-related data is small, and marks the time-related data as positively correlated with adjacent spatial positions to obtain the first target-related data; When M1 ≤ Moran's I < 0, the real-time monitoring module determines that the difference degree of adjacent spatial position variables of the time-related data is large, marks the time-related data as negatively correlated with adjacent spatial positions to obtain the second target-related data, and takes the first target-related data and the second target-related data as the target-related data; When Moran's I = 0, the real-time monitoring module determines that there is no spatial autocorrelation in the adjacent spatial positions of the time-related data, does not mark the time-related data, and removes the time-related data; When the real-time monitoring module constructs a spatio-temporal regression model based on the target-related data, set yit as the target-related data at the i-th location at time t, and xit as the variable related to time t, which is a variable related to the spatial position, and substitute the second target-related data for calculation to obtain the spatio-temporal regression model ; when the real-time monitoring module obtains the target-related data according to the second target transmission data, perform three-dimensional visualization on the target-related data, use time as the x-axis, longitude as the y-axis, latitude as the z-axis, and then use the target-related data corresponding to each time point and spatial position as a data point, plot it in three-dimensional space, and represent the data point with different shades of color according to the size of the target-related data to obtain the three-dimensional visualization model of the target-related data.

8. The power grid load monitoring system based on intelligent sensors according to claim 1, characterized in that When the real-time monitoring module constructs the dynamic threshold model based on the target-related data, 70% of the target-related data samples after being calculated by the spatio-temporal regression model are used as the training set, 15% as the validation set, and 15% as the test set to obtain the dynamic threshold model. Then, the target-related data samples are classified according to the 3σ criterion. The target-related data within the 3σ criterion are used as normal load samples, and the target-related data outside the 3σ criterion are used as abnormal load samples. Then, the normal load samples are placed into the dynamic threshold model to calculate the distance distribution of the normal load samples to the decision boundary to obtain the target output data. Then, the target output data is processed according to the quantile method to obtain the dynamic threshold. Finally, according to the real-time operation data and risk status of the power grid, the dynamic threshold is dynamically adjusted in real time by the gradient descent method.

9. The power grid load monitoring system based on intelligent sensors according to claim 1, characterized in that When the alarm module performs anomaly analysis on the second target transmission data, target-related data, and the dynamic threshold output by the dynamic threshold model, and issues an anomaly alert, it compares the target-related data G obtained from the real-time monitoring module with the preset data threshold G0, determines the security level of the target-related data based on the comparison result, and outputs a conclusion on whether the target-related data triggers the first alarm signal according to the determination result, where: When G ≤ G0, the alarm module determines that the security level of the target-related data is high and does not trigger the first alarm signal; When G > G0, the alarm module determines that the security level of the target-related data is low and triggers the first alarm signal; The target-related data is brought into the spatio-temporal regression model to calculate the predicted value H of the target-related data. The predicted value H of the target-related data is compared with the dynamic threshold H0 output by the dynamic threshold model. The anomaly degree of the change trend of the predicted value of the target-related data is judged based on the comparison result, and a conclusion on whether the predicted value of the target-related data triggers the second alarm signal is output according to the judgment result, where: When H ≤ H0, the alarm module determines that the anomaly degree of the change trend of the predicted value of the target-related data is low and does not trigger the second alarm signal; When H > H0, the alarm module determines that the anomaly degree of the change trend of the predicted value of the target-related data is high and triggers the second alarm signal.

10. The power grid load monitoring system based on intelligent sensors according to claim 9, characterized in that, When the intelligent decision-making module makes an intelligent decision on the countermeasures for anomaly alerts, it includes: primary alarm decision-making, intermediate alarm decision-making, and advanced alarm decision-making, where: When the intelligent decision-making module makes a primary alarm decision, it compares the range G1 by which the target-related data exceeds the preset data threshold with each preset range of exceeded thresholds. The preset ranges of exceeded thresholds include the first preset range of exceeded threshold G01 and the first preset range of exceeded threshold G02. It is set that G01 = 5% and G02 = 10%. The anomaly degree of the range G1 by which the target-related data exceeds the preset threshold is judged based on the comparison result, and an intelligent decision on the countermeasures for the target-related data is made according to the judgment result, where: When G1 < G01, the intelligent decision-making module determines that the anomaly degree of the range G1 by which the target-related data exceeds the preset threshold is slight. The system automatically records the alarm information, highlights it on the monitoring interface, and at the same time sends a reminder message to the operation and maintenance personnel, suggesting that the operation and maintenance personnel closely monitor the changes in relevant indicators and do not take emergency measures for the time being; When G01 ≤ G1 ≤ G02, the intelligent decision-making module determines that the anomaly degree of the range G1 by which the target-related data exceeds the preset threshold is medium. The system automatically records the alarm information, highlights it on the monitoring interface, and automatically starts a preliminary inspection; When G1 > G02, the intelligent decision-making module determines that the abnormality degree of the target-related data exceeding the preset threshold range G1 is serious. The system automatically cuts off the relevant circuit, starts the backup power supply, quickly notifies the relevant senior management personnel and professional maintenance teams, and at the same time provides detailed fault information and emergency handling suggestions, and combines with the storage and backtracking module to set the device with abnormalities as a medium-level alarm situation.

Images