A power safety intelligent monitoring system and method based on big data

Abstract

The present invention discloses an intelligent monitoring system and method for power safety based on big data, which belongs to the field of power safety technology. The present invention includes: monitoring the power fluctuation in each area of ​​the power grid, collecting the current, voltage and power parameters of the power grid, and generating a power grid fluctuation data set; analyzing the law of power grid power fluctuation, and predicting the power grid power fluctuation trend; according to the power grid power fluctuation law, combined with the current power grid state data, analyzing the arc generation risk in the power grid, obtaining the arc generation risk, and marking the power grid area where the arc generation risk is higher than the preset threshold as a risk area; monitoring the marked risk area, when a fuzzy arc is detected in the risk area, issuing instructions to the power grid protection device, isolating the risk area, and processing the fuzzy arc. The present invention effectively prevents the occurrence of arc events and ensures the stability of power grid operation by intelligently monitoring power safety.

Description

Technical Field

[0001] The present invention relates to the field of power safety technology, and in particular to a power safety intelligent monitoring system and method based on big data. Background Art

[0002] The safe operation of the power system is the foundation of the stability of the power grid, and power safety is directly related to the basic needs of industry, commerce and residents' lives. With the rapid development of my country's economy, the requirements for the power supply capacity and power quality of power companies are getting higher and higher. As a result, the operating environment of the power grid has become increasingly complex, and the safety risk supervision of power grid operation is also facing greater challenges.

[0003] At present, the regulatory authorities of power safety have summarized a set of effective methods and management experience in many years of practice. However, with the large-scale access of renewable energy and changes in electricity demand, power fluctuations in the power grid are inevitable. Although power fluctuations are not the direct cause of arcs, the changes in the power grid conditions caused by them increase the risk of overheating and insulation damage, creating conditions for the formation of arcs. The high temperature generated by the arc is enough to cause burns to the human body, and may also burn the contacts, causing poor circuit contact, further increasing the risk of circuit damage and electric shock, and affecting the reliability and stability of power supply. On the other hand, power fluctuations may cause the grid protection equipment to malfunction or fail, reducing the rapid response and isolation capabilities to arc faults, so that once an arc occurs, it may not be contained in time, posing a great threat to the safe operation of the power system.

[0004] Therefore, in order to prevent the occurrence of arc events, it is necessary to effectively monitor and manage the power grid and ensure the safe operation of electricity. Summary of the invention

[0005] The purpose of the present invention is to provide a big data-based intelligent power safety monitoring system and method to solve the problems raised in the above-mentioned background technology.

[0006] In order to solve the above technical problems, the present invention provides the following technical solutions: a method for intelligent monitoring of power safety based on big data, the method comprising the following steps:

[0007] S1: Monitor the power fluctuations in each area of ​​the power grid through data acquisition equipment installed at each node of the power grid, collect the current, voltage and power parameters of the power grid; align data from different sources in time to generate a power grid fluctuation data set;

[0008] S2: Analyze the power fluctuation patterns of the power grid based on the machine learning model, and predict the power fluctuation trend of the power grid based on the historical data of the power grid;

[0009] S3: Analyze the arc risk in the power grid according to the power grid power fluctuation law and the current power grid status data, obtain the arc risk according to the analysis result, and mark the power grid area where the arc risk is higher than the preset threshold as a risk area;

[0010] S4: monitor the risk area marked in step S3, and when a fuzzy arc is detected in the risk area, issue an instruction to the power grid protection device to isolate the risk area and process the fuzzy arc.

[0011] Further, in step S1, the data acquisition device includes a relay protection device, a sensor, a synchronous phasor measurement device and a data recorder; the power fluctuation conditions in each area of ​​the power grid are integrated through the data acquisition device to generate power grid monitoring data; each power grid area has a corresponding power grid fluctuation data set;

[0012] Among them, the grid fluctuation data set corresponding to the qth area of ​​the grid is recorded as K q , q∈[1,Q], Q represents the total number of regions in the power grid, Each power grid fluctuation data set includes power grid monitoring data at T time nodes, where: represents the current value of the t-th time node in the q-th area of ​​the power grid, represents the voltage value of the t-th time node in the q-th area of ​​the power grid, represents the active power value of the t-th time node in the q-th area of ​​the power grid, It represents the reactive power value of the t-th time node in the q-th area of ​​the power grid.

[0013] Furthermore, the step S2 specifically includes the following steps:

[0014] S2-1: Obtain the power grid fluctuation data set corresponding to each area of ​​the power grid, and perform data preprocessing on the power grid fluctuation data set, wherein the data preprocessing includes processing missing values, filling missing values, and processing abnormal values; sorting the power grid monitoring data in the power grid fluctuation data set in chronological order; defining the window length as L, starting from the power grid monitoring data corresponding to the first time node of the power grid fluctuation data set, and gradually sliding forward according to the window length;

[0015] S2-2: For each window, respectively calculate the change rate of each element in the power grid monitoring data between consecutive windows, calculate the average change rate of each element in the power grid monitoring data corresponding to each window as the eigenvalue, integrate the four eigenvalues ​​corresponding to each window to form a eigenvector, and the four eigenvalues ​​correspond to the current feature, voltage feature, active power feature and reactive power feature in the power grid monitoring data respectively; arrange all windows in chronological order to construct a feature matrix, wherein the feature matrix is ​​a two-dimensional matrix, each row represents the eigenvector of a window, and each column corresponds to a feature dimension;

[0016] S2-3: Calling the LSTM model framework, the LSTM model includes an input layer, an LSTM layer, a fully connected layer, and an output layer; segmenting the input sequence from the feature matrix, defining the output label, and converting the feature matrix into the input and output format required for model training for further predicting the power fluctuation trend of the power grid; the output label represents the target value corresponding to the input sequence, and defines the predicted data output by the model; using the feature matrix to train the LSTM model, and iterating the model through the cross entropy loss function and the gradient descent optimizer;

[0017] It should be noted that the LSTM model framework belongs to the scope of prior art and will not be described in detail here; the focus of the present invention is on how to use the LSTM model to predict the power fluctuation of the power grid, rather than the LSTM model itself.

[0018] S2-4: Deploy the trained LSTM model to the power safety intelligent monitoring system, continuously receive power grid monitoring data, predict power grid power fluctuations in real time, and take corresponding measures based on the prediction results to maintain power grid stability.

[0019] Furthermore, the step S3 specifically includes the following steps:

[0020] S3-1: Predict the grid fluctuation data according to the LSTM model, output the current prediction value, voltage prediction value, active power prediction value and reactive power prediction value, integrate the prediction value output by the LSTM model with the grid monitoring data, and generate the input data set of the arc risk prediction model;

[0021] S3-2: Combined with the current grid status data, the arc risk prediction model is used to calculate the risk of arc generation in any area of ​​the grid. Specifically, the arc risk M in the qth area of ​​the current grid is calculated according to the following formula: q :

[0022] M q =B(q)×Loss1+W(q)×Loss2+F(q)×Loss3;

[0023]

[0024]

[0025]

[0026] Wherein, B(q) represents the harmonic index in the qth region of the power grid, N represents the total number of occurrences of harmonics in the qth region of the power grid, and the harmonics are identified by a harmonic analyzer. represents the peak current of the nth harmonic appearing in the qth area of ​​the power grid, represents the peak value of the fundamental current in the qth area of ​​the power grid; represents the voltage peak value of the nth harmonic appearing in the qth area of ​​the power grid, represents the peak value of the fundamental voltage in the qth area of ​​the power grid;

[0027] W(q) represents the attenuation of the insulation material in the qth area of ​​the power grid, R0 represents the insulation strength of the insulation material at the last power grid detection, represents the predicted active power value of the T+jth time node in the qth area of ​​the power grid, represents the voltage prediction value of the T+jth time node in the qth area of ​​the power grid, λ1 and λ2 are the insulation performance influence coefficients;

[0028] F(q) represents the load impact value in the qth area of ​​the power grid, Indicates the reactive power forecast value at the T+jth time node in the qth area of ​​the power grid; the load impact value is related to the equipment loss and heat generation in the power grid. The higher the load impact value, the more reactive power flows, the greater the loss to the power grid equipment, and the higher the possibility of arc generation in the power grid;

[0029] Loss1, Loss2, and Loss3 are loss factors, which are pre-set by relevant personnel and stored in the database;

[0030] S3-3: The risk of arc generation in each area of ​​the power grid is obtained according to the calculation result of the arc risk prediction model, and the power grid area where the risk of arc generation is higher than a preset threshold is marked as a risk area.

[0031] The above method can determine the location of the area where arcs may occur in the power grid, narrow the scope of data analysis during power grid monitoring, reduce the workload of data processing, and inspect and process the risk area before the arc occurs, reducing the probability of arc occurrence.

[0032] Furthermore, the step S4 specifically includes the following steps:

[0033] S4-1: Monitor the risk area marked in step S3, and detect the monitoring data through the following steps to determine whether there is a fuzzy arc in the risk area:

[0034] S4-1-1: Obtain the historical arc occurrence records of the risk area in the database, and extract the current data of H time nodes before the G arc occurrence events in the historical arc occurrence records;

[0035] S4-1-2: Monitor the current data in the risk area, calculate the current prediction value according to step S2, and integrate to generate a set of data points to be detected, denoted as D = {Y 1′ ,|t′∈[T+jH,T+j]}; where j<H, Y1 t’ Indicates the tth ’ The current value or current prediction value at a time node;

[0036] S4-1-3: Determine whether there is a fuzzy arc in the risk area according to the following formula:

[0037]

[0038] in, is the corresponding symbol, U x Indicates the time point when the arc occurs in the xth arc generation event, U x -H represents the first H time nodes of arc occurrence in the xth arc generation event, θ represents the error term, and there is an error term table in the database, which is selected according to the time characteristics and temperature characteristics of the current risk area; Represents the trend characteristic value of the set of data points to be detected, Indicates the range of previous trend characteristics of historical arc occurrence records;

[0039] When the trend characteristic value of the set of data points to be detected corresponds to the previous trend characteristic range of the historical arc occurrence record, it is determined that there is a fuzzy arc in the current risk area;

[0040] The fuzzy arc is specifically a suspicious arc sign identified based on monitoring data, indicating an early signal of arc occurrence. When a fuzzy arc is detected, an arc is very likely to occur in the risk area;

[0041] S4-2: When a fuzzy arc is detected in a risk area, an instruction is issued to the grid protection device, which may be a circuit breaker, relay or disconnector; the power supply to the affected area is automatically or semi-automatically cut off according to the preset grid protection logic, the risk area is isolated, and an alarm is sent to the operation and maintenance personnel, providing detailed risk area information and preliminary assessment so that they can respond quickly to prevent the arc from further developing into a serious accident.

[0042] An intelligent power safety monitoring system based on big data, the system comprising a power grid data acquisition module, a power grid fluctuation analysis module, a power grid state analysis module and an arc risk detection module;

[0043] The power grid data acquisition module monitors the power fluctuations in each area of ​​the power grid through data acquisition devices installed at each node of the power grid, collects the current, voltage and power parameters of the power grid; aligns data from different sources in time to generate a power grid fluctuation data set;

[0044] The power grid fluctuation analysis module is used to analyze the law of power grid power fluctuation and predict the power grid power fluctuation trend based on the historical data of the power grid;

[0045] The grid state analysis module analyzes the current grid state according to the grid power fluctuation law to determine whether there is an arc risk in the current grid state;

[0046] The arc risk detection module is used to monitor the risk area in the power grid, detect whether a fuzzy arc occurs in the risk area, and process the fuzzy arc that occurs.

[0047] Further, the power grid data acquisition module includes a monitoring data acquisition unit and a monitoring data processing unit;

[0048] The monitoring data acquisition unit collects current, voltage and power parameters in each area of ​​the power grid through a data acquisition device, wherein the data acquisition device includes a relay protection device, a sensor, a synchronous phasor measurement device and a data recorder;

[0049] The monitoring data processing unit integrates the power fluctuation conditions in each area of ​​the power grid through the data acquisition device to generate power grid monitoring data. Each power grid area has a corresponding power grid fluctuation data set.

[0050] Further, the power grid fluctuation analysis module includes a law analysis unit and a trend prediction unit;

[0051] The law analysis unit analyzes the power grid fluctuation data set corresponding to each area of ​​the power grid, and mines the power grid power fluctuation law in each area of ​​the power grid based on the LSTM model;

[0052] The trend prediction unit continuously receives power grid monitoring data and predicts power grid power fluctuations in real time through the LSTM model deployed in the power safety intelligent monitoring system.

[0053] Furthermore, the power grid status analysis module includes a data integration unit and a risk identification unit;

[0054] The data integration unit is used to integrate the predicted value output by the LSTM model and the power grid monitoring data to generate an input data set for the arc risk prediction model;

[0055] The risk identification unit calculates the risk of arc generation in any area of ​​the power grid through the arc risk prediction model, and marks the power grid area where the risk of arc generation is higher than a preset threshold as a risk area according to the calculation result of the arc risk prediction model. By screening the area where arc may occur in the power grid, the data analysis scope during power grid monitoring is narrowed, the workload of data processing is reduced, and the risk area can be checked and processed before the arc occurs, reducing the probability of arc occurrence.

[0056] Further, the arc risk detection module includes a risk monitoring unit, a fuzzy arc detection unit and a risk processing unit;

[0057] The risk monitoring unit is used to monitor the marked risk areas in the power grid according to the output data of the power grid status analysis module, and integrate and generate a set of data points to be detected in combination with the current prediction value output by the trend prediction unit;

[0058] The fuzzy arc detection unit is used to detect whether there is a fuzzy arc in the risk area according to the set of data points to be detected output by the risk monitoring unit;

[0059] The risk processing unit is used to process ambiguous arc events in the risk area; when an ambiguous arc is detected in the risk area, an instruction is issued to the power grid protection device, wherein the power grid protection device can be a circuit breaker, a relay or an isolating switch; according to the preset power grid protection logic, the power supply of the affected area is automatically or semi-automatically cut off, the risk area is isolated, and an alarm is sent to the operation and maintenance personnel.

[0060] Compared with the prior art, the beneficial effects achieved by the present invention are:

[0061] The present invention can realize comprehensive monitoring of the power operation status of the power grid by real-time monitoring of the current, voltage and power parameters in each area of ​​the power grid, timely discover potential problems of power safety, and reduce the probability of power grid instability caused by the superposition of multiple factors. Predicting the trend of power grid power fluctuations based on machine learning models helps to control the development of power data and realize early warning of risks.

[0062] The present invention analyzes the risk of arc generation by analyzing the power fluctuation law of the power grid and the current power grid status data, and marks the power grid areas where the risk of arc generation is higher than a preset threshold as risk areas, which helps to strengthen monitoring and preventive measures in a targeted manner and improves the ability to identify potential hidden dangers to power safety.

[0063] The present invention monitors the marked risk areas in real time, and when a fuzzy arc is detected, an instruction is immediately issued to the power grid protection device, thereby achieving a rapid response to power grid risks and effectively avoiding the danger of arc generation; by isolating the risk area and processing the fuzzy arc, the expansion of the arc event is effectively prevented, the safe operation of the power grid is guaranteed, and the scope of arc damage is reduced in advance.

[0064] The present invention can be flexibly expanded as the scale of the power grid expands, and can adapt to the power safety monitoring needs of different regions and scales. BRIEF DESCRIPTION OF THE DRAWINGS

[0065] The accompanying drawings are used to provide a further understanding of the present invention and constitute a part of the specification. Together with the embodiments of the present invention, they are used to explain the present invention and do not constitute a limitation of the present invention. In the accompanying drawings:

[0066] Figure 1 It is a structural schematic diagram of an intelligent power safety monitoring system based on big data of the present invention;

[0067] Figure 2 It is a flow chart of an intelligent power safety monitoring method based on big data of the present invention. DETAILED DESCRIPTION

[0068] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0069] See also Figure 1 , in the first embodiment of the present invention: a power safety intelligent monitoring system based on big data is provided, the system includes a power grid data acquisition module, a power grid fluctuation analysis module, a power grid state analysis module and an arc risk detection module;

[0070] The power grid data acquisition module monitors the power fluctuations in each area of ​​the power grid through data acquisition devices installed at each node of the power grid, collects the current, voltage and power parameters of the power grid; aligns data from different sources in time to generate a power grid fluctuation data set; includes a monitoring data acquisition unit and a monitoring data processing unit;

[0071] The monitoring data acquisition unit collects the current, voltage and power parameters in each area of ​​the power grid through data acquisition equipment, which includes relay protection devices, sensors, synchronous phasor measurement devices and data recorders;

[0072] The monitoring data processing unit integrates the power fluctuation conditions in each area of ​​the power grid through data acquisition equipment to generate power grid monitoring data. Each power grid area has a corresponding power grid fluctuation data set.

[0073] The power grid fluctuation analysis module is used to analyze the law of power grid power fluctuation and predict the power grid power fluctuation trend based on the historical data of the power grid; it includes a law analysis unit and a trend prediction unit;

[0074] The rule analysis unit analyzes the power grid fluctuation data sets corresponding to each area of ​​the power grid, and mines the power grid power fluctuation rules in each area of ​​the power grid based on the LSTM model; the trend prediction unit continuously receives power grid monitoring data and predicts the power grid power fluctuation in real time through the LSTM model deployed in the power safety intelligent monitoring system.

[0075] The grid status analysis module analyzes the current grid status according to the grid power fluctuation law and determines whether there is an arc risk in the current grid status; it includes a data integration unit and a risk identification unit;

[0076] A data integration unit is used to integrate the prediction value output by the LSTM model with the power grid monitoring data to generate an input data set for the arc risk prediction model;

[0077] The risk identification unit calculates the risk of arc generation in any area of ​​the power grid through an arc risk prediction model, and marks the power grid area where the risk of arc generation is higher than a preset threshold as a risk area based on the calculation result of the arc risk prediction model.

[0078] The arc risk prediction model calculates the arc risk M in the qth area of ​​the current power grid according to the following formula: q :

[0079] M q =B(q)×Loss1+W(q)×Loss2+F(q)×Loss3;

[0080]

[0081]

[0082]

[0083] Where B(q) represents the harmonic index in the qth area of ​​the power grid, N represents the total number of harmonic occurrences in the qth area of ​​the power grid, and the harmonics are identified by a harmonic analyzer. represents the peak current of the nth harmonic appearing in the qth area of ​​the power grid, represents the peak value of the fundamental current in the qth area of ​​the power grid; represents the voltage peak value of the nth harmonic appearing in the qth area of ​​the power grid, represents the peak value of the fundamental voltage in the qth area of ​​the power grid;

[0084] W(q) represents the attenuation of the insulation material in the qth area of ​​the power grid, R0 represents the insulation strength of the insulation material at the last power grid detection, represents the predicted active power value of the T+jth time node in the qth area of ​​the power grid, represents the voltage prediction value of the T+jth time node in the qth area of ​​the power grid, λ1 and λ2 are the insulation performance influence coefficients;

[0085] F(q) represents the load impact value in the qth area of ​​the power grid, Indicates the reactive power forecast value at the T+jth time node in the qth area of ​​the power grid; the load impact value is related to the equipment loss and heat generation in the power grid. The higher the load impact value, the more reactive power flows, the greater the loss to the power grid equipment, and the higher the possibility of arc generation in the power grid;

[0086] Loss1, Loss2, and Loss3 are loss factors, which are pre-set by relevant personnel and stored in the database;

[0087] The arc risk detection module is used to monitor the risk area in the power grid, detect whether fuzzy arcs appear in the risk area, and process the fuzzy arcs that appear; it includes a risk monitoring unit, a fuzzy arc detection unit and a risk processing unit;

[0088] The risk monitoring unit is used to monitor the marked risk areas in the power grid according to the output data of the power grid status analysis module, and integrate the current prediction value output by the trend prediction unit to generate a set of data points to be detected;

[0089] A fuzzy arc detection unit, used to detect whether there is a fuzzy arc in the risk area according to the set of data points to be detected output by the risk monitoring unit;

[0090] The risk processing unit is used to process ambiguous arc events in the risk area; when an ambiguous arc is detected in the risk area, an instruction is issued to the power grid protection device, where the power grid protection device can be a circuit breaker, a relay or an isolating switch; according to the preset power grid protection logic, the power supply of the affected area is automatically or semi-automatically cut off, the risk area is isolated, and an alarm is sent to the operation and maintenance personnel.

[0091] See also Figure 2 In the second embodiment, a method for intelligent monitoring of power safety based on big data is provided, comprising the following steps:

[0092] S1: Monitor the power fluctuations in each area of ​​the power grid through data acquisition equipment installed at each node of the power grid, collect the current, voltage and power parameters of the power grid; align data from different sources in time to generate a power grid fluctuation data set;

[0093] The data acquisition equipment includes relay protection devices, sensors, synchronous phasor measurement devices, digital power quality analyzers and data recorders. The data acquisition equipment is used to integrate the power fluctuations in each area of ​​the power grid to generate power grid monitoring data. Each power grid area has a corresponding power grid fluctuation data set.

[0094] Among them, the grid fluctuation data set corresponding to the qth area of ​​the grid is recorded as K q , q∈[1,Q], Q represents the total number of regions in the power grid, Each power grid fluctuation data set includes power grid monitoring data at T time nodes, where: represents the current value of the t-th time node in the q-th area of ​​the power grid, represents the voltage value of the t-th time node in the q-th area of ​​the power grid, represents the active power value of the t-th time node in the q-th area of ​​the power grid, It represents the reactive power value of the t-th time node in the q-th area of ​​the power grid.

[0095] S2: Analyze the power fluctuation patterns of the power grid based on the machine learning model, and predict the power fluctuation trend of the power grid based on the historical data of the power grid; specifically:

[0096] S2-1: Obtain the power grid fluctuation data set corresponding to each area of ​​the power grid, and perform data preprocessing on the power grid fluctuation data set, which includes processing missing values, filling missing values, and processing abnormal values; sort the power grid monitoring data in the power grid fluctuation data set in chronological order; define the window length as L, starting from the power grid monitoring data corresponding to the first time node of the power grid fluctuation data set, and gradually slide forward according to the window length;

[0097] S2-2: For each window, calculate the change rate of each element in the power grid monitoring data between consecutive windows, calculate the average change rate of each element in the power grid monitoring data corresponding to each window as the eigenvalue, integrate the four eigenvalues ​​corresponding to each window to form a eigenvector, and the four eigenvalues ​​correspond to the current feature, voltage feature, active power feature and reactive power feature in the power grid monitoring data respectively; arrange all windows in chronological order and construct a feature matrix, which is a two-dimensional matrix, each row represents the eigenvector of a window, and each column corresponds to a feature dimension;

[0098] Exemplarily, the active power change rate ΔP in the power grid monitoring data corresponding to any window can be calculated according to the following formula: Y3 :

[0099]

[0100] Among them, t∈δ means that t is the time node in the δth window;

[0101] S2-3: Call the LSTM model framework. The LSTM model includes an input layer, an LSTM layer, a fully connected layer, and an output layer. For example, the LSTM model is defined such that the input layer has 240 neurons, the LSTM layer has 128 LSTM units, and the output layer has 4 neurons.

[0102] The input sequence is segmented from the feature matrix and the output label is defined. The feature matrix is ​​converted into the input and output format required for model training for further prediction of power grid power fluctuation trends. The output label represents the target value corresponding to the input sequence and defines the predicted data output by the model. The LSTM model is trained using the feature matrix and iterated using the cross entropy loss function and gradient descent optimizer.

[0103] It should be noted that the LSTM model framework belongs to the scope of prior art and will not be described in detail here; the focus of the present invention is on how to use the LSTM model to predict the power fluctuation of the power grid, rather than the LSTM model itself.

[0104] S2-4: Deploy the trained LSTM model to the power safety intelligent monitoring system, continuously receive power grid monitoring data, predict power grid power fluctuations in real time, and take corresponding measures based on the prediction results to maintain power grid stability.

[0105] S3: According to the power fluctuation law of the power grid and the current power grid status data, the arc generation risk in the power grid is analyzed, the arc generation risk is obtained according to the analysis results, and the power grid area where the arc generation risk is higher than the preset threshold is marked as a risk area; specifically:

[0106] S3-1: Predict the grid fluctuation data according to the LSTM model, output the current prediction value, voltage prediction value, active power prediction value and reactive power prediction value, integrate the prediction value output by the LSTM model with the grid monitoring data, and generate the input data set of the arc risk prediction model;

[0107] S3-2: Combined with the current grid status data, the arc risk prediction model is used to calculate the risk of arc generation in any area of ​​the grid. Specifically, the arc risk M in the qth area of ​​the current grid is calculated according to the following formula: q :

[0108] M q =B(q)×Loss1+W(q)×Loss2+F(q)×Loss3;

[0109]

[0110]

[0111]

[0112] Where B(q) represents the harmonic index in the qth area of ​​the power grid, N represents the total number of harmonic occurrences in the qth area of ​​the power grid, and the harmonics are identified by a harmonic analyzer. represents the peak current of the nth harmonic appearing in the qth area of ​​the power grid, represents the peak value of the fundamental current in the qth area of ​​the power grid; represents the voltage peak value of the nth harmonic appearing in the qth area of ​​the power grid, represents the peak value of the fundamental voltage in the qth area of ​​the power grid;

[0113] W(q) represents the attenuation of the insulation material in the qth area of ​​the power grid, R0 represents the insulation strength of the insulation material at the last power grid detection, represents the predicted active power value of the T+jth time node in the qth area of ​​the power grid, represents the voltage prediction value of the T+jth time node in the qth area of ​​the power grid, λ1 and λ2 are the insulation performance influence coefficients;

[0114] F(q) represents the load impact value in the qth area of ​​the power grid, Indicates the reactive power forecast value at the T+jth time node in the qth area of ​​the power grid; the load impact value is related to the equipment loss and heat generation in the power grid. The higher the load impact value, the more reactive power flows, the greater the loss to the power grid equipment, and the higher the possibility of arc generation in the power grid;

[0115] Loss1, Loss2, and Loss3 are loss factors, which are pre-set by relevant personnel and stored in the database;

[0116] S3-3: The risk of arc generation in each area of ​​the power grid is obtained according to the calculation result of the arc risk prediction model, and the power grid area where the risk of arc generation is higher than a preset threshold is marked as a risk area.

[0117] The above method can determine the location of the area where arcs may occur in the power grid, narrow the scope of data analysis during power grid monitoring, reduce the workload of data processing, and inspect and process the risk area before the arc occurs, reducing the probability of arc occurrence.

[0118] For example, power fluctuations in the power grid are usually accompanied by voltage fluctuations. When the voltage in the power grid is unstable, especially when the voltage drops, the arc maintenance voltage will also decrease. In this way, when the switch disconnects the circuit, an arc is likely to be generated between the contacts.

[0119] In one embodiment, when the load in the power grid suddenly increases or decreases, it will cause a rapid change in current. In the case of a sudden increase in current, if the switching device fails to cut off the current in time, an arc between the contacts may be more easily formed due to the increase in current.

[0120] In one embodiment, when current flows through a conductor, a magnetic field is generated around it. Changes in the current will cause changes in the magnetic field, thereby generating electromotive force on the conductor. When the power fluctuates, the electromotive force will change, which may cause arcing between contacts with poor contact.

[0121] In one embodiment, when power fluctuates, the instability of current may cause a rapid change in temperature in a local area, and the increase in temperature may reduce the insulation strength between contacts, thereby causing the generation of an arc.

[0122] S4: monitor the risk area marked in step S3, and when a fuzzy arc is detected in the risk area, issue an instruction to the power grid protection device to isolate the risk area and process the fuzzy arc; specifically:

[0123] S4-1: Monitor the risk area marked in step S3, and detect the monitoring data through the following steps to determine whether there is a fuzzy arc in the risk area:

[0124] S4-1-1: Obtain the historical arc occurrence records of the risk area in the database, and extract the current data of H time nodes before the G arc occurrence events in the historical arc occurrence records;

[0125] S4-1-2: Monitor the current data in the risk area, calculate the current prediction value according to step S2, and integrate to generate a set of data points to be detected, denoted as D = {Y 1′ ,|t′∈[T+jH,T+j]}; where j<H, Y1 t’ represents the current value or current prediction value at the t'th time node;

[0126] It should be noted that in the above step S4-1-2, t'∈[T+jH, T+j]. In the present invention, the current time node of real-time collection is recorded as T, then T+j represents the j-th time node after the current time node, and j represents the time node corresponding to the predicted value output by the LSTM model in step S2; at the same time, if j<H, then T+jH<T, then T+jH is the historical time node before the current time node. In the present invention, the historical time node has the corresponding current value collected by monitoring, and the j-th time node after the current time node has the predicted value output by the LSTM model, so Y1 t’ It can be the current value or the current prediction value;

[0127] S4-1-3: Determine whether there is a fuzzy arc in the risk area according to the following formula:

[0128]

[0129] in, is the corresponding symbol, U x Indicates the time point when the arc occurs in the xth arc generation event, U x -H represents the first H time nodes of arc occurrence in the xth arc generation event, θ represents the error term, and there is an error term table in the database, which is selected according to the time characteristics and temperature characteristics of the current risk area; Represents the trend characteristic value of the set of data points to be detected, Indicates the range of previous trend characteristics of historical arc occurrence records;

[0130] When the trend characteristic value of the set of data points to be detected corresponds to the previous trend characteristic range of the historical arc occurrence record, it is determined that there is a fuzzy arc in the current risk area;

[0131] Fuzzy arc refers to suspicious arc signs identified based on monitoring data, indicating the early signal of arc occurrence. When fuzzy arc is detected, it is very likely that arc will occur in the risk area.

[0132] In another embodiment, determining whether there is a fuzzy arc in the risk area can be achieved by analyzing the voltage;

[0133] S4-2: When a fuzzy arc is detected in a risk area, an instruction is issued to the grid protection device, which can be a circuit breaker, relay or disconnector; the power supply to the affected area is automatically or semi-automatically cut off according to the preset grid protection logic, the risk area is isolated, and an alarm is sent to the operation and maintenance personnel, providing detailed risk area information and preliminary assessment so that they can respond quickly to prevent the arc from further developing into a serious accident. After receiving the alarm, the operation and maintenance personnel remotely monitor or conduct on-site surveys of the risk area to confirm the existence and nature of the arc, repair or replace the damaged equipment, eliminate the hidden dangers of the arc, and record the details of the event, including time, location, arc characteristics, etc., for subsequent analysis and reporting. After confirming that the arc discharge event has been effectively handled, the grid operation in the affected area is restored.

[0134] Optionally, quickly disconnect via a circuit breaker or disconnector to isolate the arc risk area;

[0135] Alternatively, reconfigure the grid topology to bypass risk areas and maintain power supply to other areas.

[0136] It should be noted that, in this article, relational terms such as first and second, etc. are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the terms "include", "comprise" or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or device.

[0137] Finally, it should be noted that the above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art can still modify the technical solutions described in the aforementioned embodiments or replace some of the technical features therein by equivalents. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present invention shall be included in the protection scope of the present invention.

Claims

1. A method for intelligent monitoring of power safety based on big data, characterized in that: The method comprises the following steps: S1: Monitor the power fluctuations in each area of ​​the power grid through data acquisition equipment installed at each node of the power grid, collect the current, voltage and power parameters of the power grid; align data from different sources in time to generate a power grid fluctuation data set; S2: Analyze the power fluctuation patterns of the power grid based on the machine learning model, and predict the power fluctuation trend of the power grid based on the historical data of the power grid; S3: Analyze the arc risk in the power grid according to the power grid power fluctuation law and the current power grid status data, obtain the arc risk according to the analysis result, and mark the power grid area where the arc risk is higher than the preset threshold as a risk area; S4: monitoring the risk area marked in step S3, and when a fuzzy arc is detected in the risk area, issuing instructions to the power grid protection device to isolate the risk area and process the fuzzy arc; The step S3 specifically comprises the following steps: S3-1: Predict the grid fluctuation data according to the LSTM model, output the current prediction value, voltage prediction value, active power prediction value and reactive power prediction value, integrate the prediction value output by the LSTM model with the grid monitoring data, and generate the input data set of the arc risk prediction model; S3-2: Combined with the current grid status data, the arc risk prediction model is used to calculate the risk of arc generation in any area of ​​the grid. Specifically, the arc risk M in the qth area of ​​the current grid is calculated according to the following formula: q : M q =B(q)×Loss1+W(q)×Loss2+F(q)×Loss3; Where B(q) represents the harmonic index in the qth area of ​​the power grid, N represents the total number of harmonic occurrences in the qth area of ​​the power grid, represents the peak current of the nth harmonic appearing in the qth area of ​​the power grid, represents the peak value of the fundamental current in the qth area of ​​the power grid; represents the voltage peak value of the nth harmonic appearing in the qth area of ​​the power grid, represents the peak value of the fundamental voltage in the qth area of ​​the power grid; W(q) represents the insulation material attenuation in the qth area of ​​the power grid, R0 represents the insulation strength of the insulation material, represents the predicted active power value of the T+jth time node in the qth area of ​​the power grid, represents the voltage prediction value of the T+jth time node in the qth area of ​​the power grid, λ1 and λ2 are the insulation performance influence coefficients; F(q) represents the load impact value in the qth area of ​​the power grid, It represents the reactive power prediction value of the T+jth time node in the qth area of ​​the power grid; Loss1, Loss2 and Loss3 are loss factors, which are pre-set by relevant personnel and stored in the database; S3-3: The risk of arc generation in each area of ​​the power grid is obtained according to the calculation result of the arc risk prediction model, and the power grid area where the risk of arc generation is higher than a preset threshold is marked as a risk area.

2. According to the big data-based intelligent power safety monitoring method of claim 1, it is characterized by: In step S1, the data acquisition device includes a relay protection device, a sensor, a synchronous phasor measurement device and a data recorder; the power fluctuation conditions in each area of ​​the power grid are integrated through the data acquisition device to generate power grid monitoring data; each power grid area has a corresponding power grid fluctuation data set; Among them, the grid fluctuation data set corresponding to the qth area of ​​the grid is recorded as K q , q∈[1,Q], Q represents the total number of regions in the power grid, Each power grid fluctuation data set includes power grid monitoring data at T time nodes, where: represents the current value of the t-th time node in the q-th area of ​​the power grid, represents the voltage value of the t-th time node in the q-th area of ​​the power grid, represents the active power value of the t-th time node in the q-th area of ​​the power grid, It represents the reactive power value of the t-th time node in the q-th area of ​​the power grid.

3. The method for intelligent monitoring of power safety based on big data according to claim 1 is characterized in that: The step S2 specifically includes the following steps: S2-1: Obtain the power grid fluctuation data set corresponding to each area of ​​the power grid, and perform data preprocessing on the power grid fluctuation data set; sort the power grid monitoring data in the power grid fluctuation data set in chronological order; define the window length as L, starting from the power grid monitoring data corresponding to the first time node of the power grid fluctuation data set, and gradually slide forward according to the window length; S2-2: For each window, calculate the change rate of each element in the power grid monitoring data between consecutive windows, calculate the average change rate of each element in the power grid monitoring data corresponding to each window as the eigenvalue, and integrate the four eigenvalues ​​corresponding to each window to form a eigenvector; arrange all windows in chronological order to construct a feature matrix; S2-3: Call the LSTM model framework, split the input sequence from the feature matrix, and define the output label; use the feature matrix to train the LSTM model, and iterate the model through the cross entropy loss function and gradient descent optimizer; S2-4: Deploy the trained LSTM model to the power safety intelligent monitoring system, continuously receive power grid monitoring data, and predict power grid power fluctuations in real time.

4. The method for intelligent monitoring of power safety based on big data according to claim 1 is characterized in that: The step S4 specifically comprises the following steps: S4-1: Monitor the risk area marked in step S3, and detect the monitoring data through the following steps to determine whether there is a fuzzy arc in the risk area: S4-1-1: Obtain the historical arc occurrence records of the risk area in the database, and extract the current data of H time nodes before the G arc occurrence events in the historical arc occurrence records; S4-1-2: Monitor the current data in the risk area, calculate the current prediction value according to step S2, and integrate to generate a set of data points to be detected, denoted as D = {Y1 t’ ,|t′∈[T+jH,T+j]}; where j<H, Y1 t’ represents the current value or current prediction value at the t'th time node; S4-1-3: Determine whether there is a fuzzy arc in the risk area according to the following formula: in, is the corresponding symbol, U x Indicates the time point when the arc occurs in the xth arc generation event, U x -H represents the first H time nodes of the arc occurrence in the xth arc generation event, θ represents the error term, Represents the trend characteristic value of the set of data points to be detected, Indicates the range of previous trend characteristics of historical arc occurrence records; When the trend characteristic value of the set of data points to be detected corresponds to the previous trend characteristic range of the historical arc occurrence record, it is determined that there is a fuzzy arc in the current risk area; S4-2: When a fuzzy arc is detected in a risk area, a command is issued to the grid protection device; the power supply to the affected area is automatically or semi-automatically cut off according to the preset grid protection logic, the risk area is isolated, and an alarm is sent to the operation and maintenance personnel.

5. An intelligent power safety monitoring system based on big data, executing an intelligent power safety monitoring method based on big data as claimed in any one of claims 1 to 4, characterized in that: The system includes a power grid data acquisition module, a power grid fluctuation analysis module, a power grid status analysis module and an arc risk detection module; The power grid data acquisition module monitors the power fluctuations in each area of ​​the power grid through data acquisition devices installed at each node of the power grid, collects the current, voltage and power parameters of the power grid; aligns data from different sources in time to generate a power grid fluctuation data set; The power grid fluctuation analysis module is used to analyze the law of power grid power fluctuation and predict the power grid power fluctuation trend based on the historical data of the power grid; The grid state analysis module analyzes the current grid state according to the grid power fluctuation law to determine whether there is an arc risk in the current grid state; The arc risk detection module is used to monitor the risk area in the power grid, detect whether a fuzzy arc occurs in the risk area, and process the fuzzy arc that occurs.

6. The power safety intelligent monitoring system based on big data according to claim 5 is characterized by: The power grid data acquisition module includes a monitoring data acquisition unit and a monitoring data processing unit; The monitoring data acquisition unit collects current, voltage and power parameters in each area of ​​the power grid through a data acquisition device, wherein the data acquisition device includes a relay protection device, a sensor, a synchronous phasor measurement device and a data recorder; The monitoring data processing unit integrates the power fluctuation conditions in each area of ​​the power grid through the data acquisition device to generate power grid monitoring data. Each power grid area has a corresponding power grid fluctuation data set.

7. The power safety intelligent monitoring system based on big data according to claim 5 is characterized by: The power grid fluctuation analysis module includes a law analysis unit and a trend prediction unit; The law analysis unit analyzes the power grid fluctuation data set corresponding to each area of ​​the power grid, and mines the power grid power fluctuation law in each area of ​​the power grid based on the LSTM model; The trend prediction unit continuously receives power grid monitoring data and predicts power grid power fluctuations in real time through the LSTM model deployed in the power safety intelligent monitoring system.

8. The power safety intelligent monitoring system based on big data according to claim 5 is characterized by: The power grid status analysis module includes a data integration unit and a risk identification unit; The data integration unit is used to integrate the predicted value output by the LSTM model and the power grid monitoring data to generate an input data set for the arc risk prediction model; The risk identification unit calculates the risk of arc generation in any area of ​​the power grid through the arc risk prediction model, and marks the power grid area where the risk of arc generation is higher than a preset threshold as a risk area according to the calculation result of the arc risk prediction model.

9. The power safety intelligent monitoring system based on big data according to claim 5 is characterized by: The arc risk detection module includes a risk monitoring unit, a fuzzy arc detection unit and a risk processing unit; The risk monitoring unit is used to monitor the marked risk areas in the power grid according to the output data of the power grid status analysis module, and integrate and generate a set of data points to be detected in combination with the current prediction value output by the trend prediction unit; The fuzzy arc detection unit is used to detect whether there is a fuzzy arc in the risk area according to the set of data points to be detected output by the risk monitoring unit; The risk processing unit is used to process fuzzy arc events in the risk area; When a fuzzy arc is detected in a risk area, a command is issued to the grid protection device to automatically or semi-automatically cut off the power supply to the affected area according to the preset grid protection logic, isolate the risk area, and send an alarm to the operation and maintenance personnel.

Images