Real-time safety monitoring system for substation main grounding grid based on multi-sensor fusion

Abstract

The present application belongs to the technical field of substation grounding grid monitoring, and discloses a real-time safety monitoring system for a substation main grounding grid based on multi-sensor fusion. A dynamic association database module adopts a mixed architecture of MySQL and InfluxDB to store static data such as the location of the substation and the material of the grounding grid, and dynamic data such as soil parameters and corrosion rates. A soil parameter and sampling frequency dynamic mapping table is built in to send a sampling frequency adjustment instruction to the sensor array module and start the soil sample automatic collection device when the soil salinity is greater than 15 mS / cm or the pH is less than 5.5. The module also establishes a material corrosion digital twin model and a climate zoning library to provide scenario data for the scheme dynamic adaptation and operation and maintenance module, so that it can output an adaptive scenario corrosion prevention scheme, and calculate the grounding electrode replacement cycle according to the corrosion rate to build a dynamic association between soil working conditions and corrosion protection, effectively improving the durability of the grounding grid.

Description

Technical Field

[0001] This invention belongs to the field of substation grounding grid monitoring technology, specifically a real-time safety monitoring system for substation main grounding grid based on multi-sensor fusion. Background Technology

[0002] The main grounding grid of a substation is the core guarantee for the safe operation of the power system. It undertakes the key functions of discharging fault current, stabilizing ground potential, and cooperating with lightning protection devices. Its performance directly affects the reliability of the power grid. However, existing technologies still have the following technical problems, making it difficult to meet actual operation and maintenance needs:

[0003] However, existing technologies can only statically record basic soil parameters and have not established a dynamic correlation model between real-time soil conditions, grounding device corrosion rate, and grounding performance degradation. This results in most anti-corrosion solutions being fixed and unable to be dynamically adjusted according to on-site soil conditions. Consequently, the risk of grounding devices losing ground due to corrosion remains high, directly affecting the durability of the grounding grid.

[0004] Existing technologies separate grounding grid monitoring (such as grounding resistance measurement) from lightning protection monitoring (such as lightning current recording), failing to reveal the correlation mechanism between the grounding device status (such as local corrosion, changes in impulse grounding resistance) and the lightning protection effect. When the grounding grid loses ground and its current dissipation capacity decreases, the lightning current cannot be effectively discharged, which can easily lead to excessive residual voltage of the lightning protection device and equipment backflash failure, but cannot be predicted in advance.

[0005] Existing technologies largely rely on manual, periodic offline measurements, such as single-point grounding resistance testing, or monitoring only a single electrical parameter. This fails to achieve multi-dimensional, real-time monitoring of the grounding network's electrical performance, corrosion status, and environmental conditions. Consequently, maintenance personnel cannot promptly detect hidden faults, such as localized corrosion, and struggle to respond quickly to emergencies (such as lightning strikes or soil contamination). Summary of the Invention

[0006] The purpose of this invention is to provide a real-time safety monitoring system for the main grounding grid of a substation based on multi-sensor fusion, so as to solve the problems mentioned in the background art.

[0007] To achieve the above objectives, the present invention provides the following technical solution: a real-time safety monitoring system for the main grounding grid of a substation based on multi-sensor fusion, the system comprising:

[0008] Sensor array module: Configures electrical parameter sensor group, environmental-corrosion sensor group, and lightning protection related sensor group, and transmits the collected raw data to the acquisition and preprocessing module;

[0009] Acquisition and preprocessing module: Upon receiving the raw data, dynamically adjusts the sampling frequency according to the operating conditions, performs noise reduction, converts the analog quantity into a digital quantity, encapsulates it into a standard data frame, and outputs standardized data;

[0010] Fusion Analysis Module: Receives the standardized data and historical data, adopts a three-level fusion architecture, the data layer eliminates random errors, the feature layer introduces cross-modal fusion of electrical and image data, the decision layer outputs the status, sets an abnormal data repair mechanism, and outputs the repaired data;

[0011] Dynamically related database module: Receives the standardized data, repaired data, and execution effect data. It adopts a hybrid architecture of MySQL and InfluxDB, has a built-in dynamic mapping table of soil parameters and sampling frequency, establishes a digital twin model of material corrosion and a climate zoning library, and outputs sampling frequency adjustment instructions and scenario-based data.

[0012] Safety assessment and failure prediction module: Establish a dynamic weighted synergy coefficient model, predict two types of risks based on the LSTM model, output assessment results and prediction information, and provide feedback control effect instructions;

[0013] Real-time early warning and visualization interaction platform module: Establish a three-dimensional model of the grounding grid, set three-level early warning response, and output early warning information and fault location data;

[0014] Dynamic adaptation and operation and maintenance module: Receives the scenario-based data and the fault location data, outputs the scenario-based anti-corrosion solution, calculates the grounding electrode replacement cycle, and outputs the solution execution effect.

[0015] Preferably, the sensor array module is as follows:

[0016] Configure an electrical parameter sensor group, an environmental-corrosion sensor group, and a lightning protection-related sensor group. The electrical parameter sensor group includes a grounding resistance sensor, a ground potential distribution sensor, and a leakage current sensor, which are used to collect electrical performance data, environmental corrosion data, and lightning protection status data.

[0017] The environmental-corrosion sensor group includes a soil moisture sensor, a soil salinity sensor, a soil pH sensor, an electrochemical corrosion sensor, and a heavy metal sensor. The lightning protection-related sensor group includes an impulse grounding resistance sensor, a lightning current sensor, and a surge arrester action counter. Miniature industrial cameras are deployed at key nodes of the grounding grid to collect soil profile images. All sensors adopt a double-layer electromagnetic shielding design, and the collected raw data is transmitted to the acquisition and preprocessing module.

[0018] Formula for dynamic sampling frequency of electrochemical corrosion sensor:

[0019]

[0020] In the formula: f represents the dynamic sampling frequency of the electrochemical corrosion sensor, in Hz;

[0021] f0 represents the fundamental sampling frequency, in Hz;

[0022] k f This represents the sampling frequency adjustment coefficient, with a value of 5.

[0023] S represents soil salinity, with units of mS / cm;

[0024] pH indicates the acidity or alkalinity of the soil.

[0025] Segmentation conditions: S>15mS / cm or pH<5.5 are the criteria for judging abnormal soil conditions, triggering high-frequency sampling to capture dynamic changes in corrosion.

[0026] Preferably, the acquisition and preprocessing module is as follows:

[0027] Upon receiving the raw data from the sensor array module, the acquisition and preprocessing module uses an FPGA as the main control chip and combines GPS timing to dynamically adjust the sampling frequency according to the working conditions. The electrical parameters are 1kHz, the conventional soil parameters are 1Hz, the abnormal soil conditions are 5Hz, and the lightning current is 1MHz.

[0028] Adaptive notch filtering is used to remove 50Hz power frequency interference, and wavelet threshold denoising is used to process transient pulses. Analog quantities are converted into digital quantities and encapsulated into standard data frames of timestamp-sensor ID-parameter value. The processed standardized data is then synchronously pushed to the fusion analysis module and the dynamic association database module.

[0029] Preferably, the fusion analysis module is specifically as follows:

[0030] The system receives standardized data from the acquisition and preprocessing module and historical data from the dynamic correlation database module, and adopts a three-level fusion architecture: the data layer uses a weighted average method to fuse data from similar sensors to eliminate random errors, and calls historical data from the same period in the dynamic correlation database module to assist in the identification of abnormal data;

[0031] The feature layer employs a hybrid model of graph neural network and LSTM to extract the topological and temporal features of the grounding grid. It also introduces cross-modal fusion of electrical and image data to normalize the grayscale features of the soil profile image with the electrical performance data, thereby strengthening the correlation between abnormal areas and parameter mutations. The decision layer combines an expert knowledge base to output three levels of status: safety, early warning, and alarm.

[0032] This module is also configured with an abnormal data repair mechanism. When a sensor fails, it is completed by neighborhood interpolation or historical model. The repaired data is synchronously fed back to the dynamic association database module for supplementary storage. The extracted soil-corrosion and grounding-lightning protection association feature data are transmitted to the safety assessment and failure prediction module in real time. Abnormal status information is pushed to the real-time early warning and visualization interaction platform module.

[0033] Preferably, the dynamic association database module is specifically as follows:

[0034] The system receives standardized data from the acquisition and preprocessing module, repaired data from the fusion analysis module, and execution effect data from the scheme dynamic adaptation and operation and maintenance module. It adopts a hybrid architecture of MySQL and InfluxDB. The MySQL storage includes static data such as substation location and grounding grid material, while the InfluxDB storage includes dynamic data such as soil parameters and corrosion rate.

[0035] The system includes a built-in dynamic mapping table between soil parameters and sampling frequency. When soil salinity is identified as >15 mS / cm or pH <5.5, a sampling frequency adjustment command is sent to the sensor array module, and the automatic soil sample collection device is activated. A digital twin model of material corrosion and a climate zone feature library are established to provide scenario-based data for the dynamic adaptation and operation and maintenance module of the solution, and to provide historical soil data for the safety assessment and failure prediction module.

[0036] Soil salinity-corrosion rate correlation model formula:

[0037]

[0038] In the formula: v corr The electrochemical corrosion rate of the grounding device is expressed in mm / a.

[0039] k1 represents the soil salinity correlation coefficient, with a value ranging from 0.08 to 0.15 mm / (a·mS / cm);

[0040] S represents the real-time monitored soil salinity, in mS / cm;

[0041] k2 represents the soil pH correlation coefficient, with a value ranging from 0.05 to 0.09 mm / (a·pH);

[0042] pH indicates the real-time monitoring of soil acidity and alkalinity;

[0043] k3 represents the soil temperature correction factor;

[0044] T represents the real-time monitored soil temperature, in °C;

[0045] T0 represents the soil reference temperature, measured in °C.

[0046] Preferably, the safety assessment and failure prediction module is as follows:

[0047] Upon receiving the associated feature data from the fusion analysis module and the historical soil data from the dynamic association database module, and combining the impulse grounding resistance, lightning current threshold, and equipment health, a dynamic weighted synergy coefficient model is established. Based on the LSTM model, two types of risks are predicted: local corrosion-excessive lightning protection residual voltage and soil pollution-grounding current dissipation failure.

[0048] The evaluation results and prediction information are pushed to the real-time early warning and visualization interaction platform module to trigger the corresponding early warning level. When the weight coordination coefficient is less than 0.5 or the failure probability reaches the alarm threshold, a control command is sent to the local cathodic protection device, and the control effect data is fed back to the dynamic association database module and the fusion analysis module.

[0049] Formula for dynamic weight coordination coefficient of grounding-lightning protection:

[0050]

[0051] Where: K d This represents the dynamic weighting coordination coefficient between grounding and lightning protection;

[0052] ω1 represents the impulse grounding resistance weighting factor;

[0053] ω2 represents the weighting factor for the lightning current withstand threshold;

[0054] R imp This represents the real-time monitored impulse grounding resistance, expressed in Ω.

[0055] R imp0 This indicates the standard value of impulse grounding resistance, in Ω.

[0056] I light This represents the peak lightning current monitored in real time, in kA.

[0057] I light0 This indicates the lightning current withstand threshold, in kA.

[0058] α represents the equipment health correction factor.

[0059] Formula for the failure probability of localized corrosion-lightning protection residual voltage exceeding the limit:

[0060]

[0061] In the formula: P represents the failure probability of lightning protection residual voltage exceeding the standard due to local corrosion, and the value ranges from 0 to 1;

[0062] 'a' represents the corrosion rate influence coefficient, with a value of 1.2 mm. -1 ·a;

[0063] v corrThis represents the local grounding electrode corrosion rate, expressed in mm / a.

[0064] b represents the influence coefficient of impulse grounding resistance deviation, with a value of 0.1Ω. -1 ;

[0065] R imp -R imp0 This indicates the deviation value of the impulse grounding resistance, in Ω;

[0066] c represents the influence coefficient of lightning current exceeding the limit, with a value of 3.5;

[0067] This indicates the percentage of lightning current exceeding the limit.

[0068] Preferably, the real-time early warning and visualization interaction platform module is specifically as follows:

[0069] The system receives abnormal status information from the fusion analysis module and evaluation results from the safety assessment and failure prediction module, and establishes a three-dimensional model of the grounding grid based on WebGIS to display sensor locations, real-time data, and safety status.

[0070] A three-level early warning response is set up, and the early warning information and fault location data are synchronously pushed to the dynamic adaptation and operation and maintenance module of the solution; it has the functions of data query, curve comparison and report export, and the exported historical data is supplemented to the dynamic association database module;

[0071] Formula for determining the warning level index:

[0072]

[0073] In the formula: S warn This represents the index for determining the warning level, with a value ranging from 0 to 1.

[0074] F represents the output value of the multi-sensor fusion feature layer;

[0075] Th safe Indicates the safety threshold;

[0076] Th alarm Indicates the alarm threshold;

[0077] Th alarm -Th safe This indicates the threshold range.

[0078] Preferably, the dynamic adaptation and maintenance module of the solution is as follows:

[0079] Upon receiving the scenario-based data from the dynamic association database module and the fault warning information and fault location data from the real-time early warning and visualization interaction platform module, the system outputs a scenario-based anti-corrosion scheme based on soil parameters, calculates the grounding electrode replacement cycle based on the corrosion rate, and generates a spare parts list and construction plan in conjunction with the operation and maintenance system.

[0080] The corrosion rate change and risk reduction effect data after the implementation of the solution are fed back to the dynamic correlation database module and synchronized to the fusion analysis module. At the same time, it provides the safety assessment and failure prediction module with post-operation and maintenance equipment status data and corrects the dynamic weighted coordination coefficient model.

[0081] The beneficial effects of this invention are as follows:

[0082] 1. This invention employs a hybrid architecture of MySQL and InfluxDB in its dynamic relational database module to store static data such as substation location and grounding grid material, as well as dynamic data such as soil parameters and corrosion rate. It also includes a built-in dynamic mapping table between soil parameters and sampling frequency. When soil salinity is >15 mS / cm or pH <5.5, the module sends a sampling frequency adjustment command to the sensor array module and activates an automatic soil sample collection device. Furthermore, this module establishes a digital twin model of material corrosion and a climate zoning library, providing scenario-based data for dynamic adaptation and maintenance modules. This enables the module to output suitable scenario-based corrosion protection solutions, such as recommending titanium alloy coating + zinc-magnesium sacrificial anodes for high-salt soils. The module also calculates the grounding electrode replacement cycle based on the corrosion rate, constructing a dynamic correlation between soil conditions and corrosion protection, effectively improving the durability of the grounding grid.

[0083] 2. The safety assessment and failure prediction module of this invention receives the associated feature data from the fusion analysis module and the historical soil data from the dynamic association database module. It then establishes a dynamic weighted coordination coefficient model by combining the impulse grounding resistance, lightning current threshold, and equipment health. Based on the LSTM model, it accurately predicts two types of risks: local corrosion-excessive lightning protection residual voltage and soil pollution-grounding current dissipation failure. When the weighted coordination coefficient is less than 0.5 or the failure probability reaches the alarm threshold, the module pushes information to the real-time early warning and visualization interaction platform module to trigger the corresponding early warning level, and also sends control commands to the local cathodic protection device to achieve coordinated management and control of grounding and lightning protection, and avoids equipment backflash failure in advance.

[0084] 3. This invention utilizes a sensor array module to configure electrical parameter sensor groups, environmental-corrosion sensor groups, lightning protection-related sensor groups, and a miniature industrial camera to achieve multi-dimensional raw data acquisition. The acquisition and preprocessing module uses an FPGA as the main control chip combined with GPS timing to dynamically adjust the sampling frequency and process the data through adaptive notch filtering and wavelet threshold denoising. The fusion analysis module adopts a three-level fusion architecture to eliminate random errors, strengthen parameter correlation, and also has an abnormal data repair mechanism. The entire solution achieves multi-dimensional real-time monitoring and reliable data analysis, enabling maintenance personnel to promptly detect hidden faults and improve the response speed to emergencies. Attached Figure Description

[0085] Figure 1 This is a flowchart of the real-time safety monitoring system for the main grounding grid of a substation based on multi-sensor fusion, as described in this invention.

[0086] Figure 2 This is a flowchart of the data acquisition and preprocessing process of the present invention;

[0087] Figure 3 This is a flowchart of the multi-sensor hierarchical fusion analysis process of the present invention. Detailed Implementation

[0088] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0089] like Figures 1 to 3 As shown in the figure, this embodiment of the invention provides a real-time safety monitoring system for the main grounding grid of a substation based on multi-sensor fusion. The system includes:

[0090] Sensor array module: Configures electrical parameter sensor group, environmental-corrosion sensor group, and lightning protection related sensor group, and transmits the collected raw data to the acquisition and preprocessing module;

[0091] Acquisition and preprocessing module: Upon receiving the raw data, it dynamically adjusts the sampling frequency according to the operating conditions, performs noise reduction, converts the analog quantity into a digital quantity, encapsulates it into a standard data frame, and outputs standardized data.

[0092] Fusion Analysis Module: Receives standardized data and historical data, adopts a three-level fusion architecture, the data layer eliminates random errors, the feature layer introduces cross-modal fusion of electrical and image data, the decision layer outputs the status, sets an abnormal data repair mechanism, and outputs the repaired data;

[0093] Dynamically related database module: Receives standardized data, repaired data, and execution effect data. It adopts a hybrid architecture of MySQL and InfluxDB, has a built-in dynamic mapping table of soil parameters and sampling frequency, establishes a digital twin model of material corrosion and a climate zoning library, and outputs sampling frequency adjustment instructions and scenario-based data.

[0094] Safety assessment and failure prediction module: Establish a dynamic weighted synergy coefficient model, predict two types of risks based on the LSTM model, output assessment results and prediction information, and provide feedback control effect instructions;

[0095] Real-time early warning and visualization interaction platform module: Establish a three-dimensional model of the grounding grid, set three-level early warning response, and output early warning information and fault location data;

[0096] Dynamic adaptation and operation and maintenance module: Receives scenario-based data and fault location data, outputs scenario-based anti-corrosion solutions, calculates the grounding electrode replacement cycle, and outputs the solution execution effect.

[0097] The sensor array module is as follows:

[0098] Configure an electrical parameter sensor group, an environmental-corrosion sensor group, and a lightning protection-related sensor group. The electrical parameter sensor group includes a grounding resistance sensor, a ground potential distribution sensor, and a leakage current sensor, which are used to collect electrical performance data, environmental corrosion data, and lightning protection status data.

[0099] The environmental-corrosion sensor group includes soil moisture sensor, soil salinity sensor, soil pH sensor, electrochemical corrosion sensor, and heavy metal sensor. The lightning protection-related sensor group includes impulse grounding resistance sensor, lightning current sensor, and surge arrester action counter. Miniature industrial cameras are deployed at key nodes of the grounding grid to collect soil profile images. All sensors adopt a double-layer electromagnetic shielding design, which can resist power frequency interference of more than 40dB. The raw data collected is transmitted to the acquisition and preprocessing module.

[0100] The miniature industrial camera has a shooting resolution of 1080P, a soil environmental protection level of IP68, a single shooting coverage area of ​​0.5-1m in diameter, and a shooting frequency that is coordinated with the soil parameter sampling frequency, 1Hz under normal working conditions and 5Hz under abnormal working conditions.

[0101] Formula for dynamic sampling frequency of electrochemical corrosion sensor:

[0102]

[0103] In the formula: f represents the dynamic sampling frequency of the electrochemical corrosion sensor, in Hz, with a base frequency of 1Hz and an increase to 5Hz under abnormal operating conditions;

[0104] f0 represents the basic sampling frequency, in Hz, and is set to 1Hz, which is the sampling frequency under normal soil conditions.

[0105] k f This represents the sampling frequency adjustment factor, with a value of 5, which refers to the frequency amplification factor under abnormal operating conditions.

[0106] S represents soil salinity, measured in mS / cm, and is the basis for determining whether soil is highly saline.

[0107] pH indicates the acidity or alkalinity of soil and is the basis for determining whether soil is acidic.

[0108] Segmentation conditions: S>15mS / cm or pH<5.5 are the criteria for judging abnormal soil conditions, triggering high-frequency sampling to capture dynamic changes in corrosion.

[0109] The acquisition and preprocessing module is as follows:

[0110] Upon receiving the raw data from the sensor array module, the acquisition and preprocessing module uses an FPGA as the main control chip and combines GPS timing to dynamically adjust the sampling frequency according to the working conditions: 1kHz for electrical parameters, 1Hz for conventional soil parameters, 5Hz for abnormal soil conditions, and 1MHz for lightning current.

[0111] The number of FPGA logic units is no less than 500,000, the single-channel data processing rate is ≥1Mbps, and the GPS timing accuracy is at the microsecond level.

[0112] Adaptive notch filtering is used to remove 50Hz power frequency interference, and wavelet threshold denoising is used to process transient pulses to ensure a signal-to-noise ratio of ≥60dB. Analog quantities are converted into digital quantities through a 16-bit ADC and encapsulated into standard data frames of timestamp-sensor ID-parameter value. The processed standardized data is synchronously pushed to the fusion analysis module and the dynamic association database module via IEC61850 Ethernet.

[0113] The fusion analysis module is as follows:

[0114] The system receives standardized data from the acquisition and preprocessing module and historical data from the dynamic correlation database module, and adopts a three-level fusion architecture: the data layer uses a weighted average method to fuse data from similar sensors to eliminate random errors, and calls historical data from the same period in the dynamic correlation database module to assist in the identification of abnormal data;

[0115] The weighting rules of the weighted average method are set according to the accuracy level of the sensor. The higher the accuracy, the greater the weight. For example, the weight of a sensor with an accuracy of ±1% is set to 0.6, and the weight of a sensor with an accuracy of ±3% is set to 0.4.

[0116] The feature layer employs a hybrid model of graph neural network and LSTM to extract the topological and temporal features of the grounding grid (node ​​connection relationships, burial depth differences). It also introduces cross-modal fusion of electrical and image data to normalize the grayscale features of soil profile images and electrical performance data, thereby strengthening the correlation between abnormal areas and parameter mutations. The decision layer combines an expert knowledge base to output three levels of status: safety, early warning, and alarm.

[0117] The graph neural network adopts an undirected graph structure (nodes are key monitoring points of the grounding grid) and the LSTM has 2 hidden layers.

[0118] The normalization algorithm for cross-modal fusion of electrical and image data is min-max normalization, which maps the data to the [0, 1] interval;

[0119] This module is also configured with an abnormal data repair mechanism. When a sensor fails, it is completed by neighborhood interpolation or historical model. The repaired data is synchronously fed back to the dynamic correlation database module for supplementary storage. The extracted soil-corrosion and grounding-lightning protection correlation feature data are transmitted to the safety assessment and failure prediction module in real time. Abnormal status information is pushed to the real-time early warning and visualization interaction platform module.

[0120] The dynamic database module is as follows:

[0121] The system receives standardized data from the acquisition and preprocessing module, data repaired by the fusion analysis module, and execution effect data from the scheme dynamic adaptation and operation and maintenance module. It adopts a hybrid architecture of MySQL and InfluxDB, with MySQL storing static data such as substation location and grounding grid material, and InfluxDB storing dynamic data such as soil parameters and corrosion rate.

[0122] The system incorporates a dynamic mapping table between soil parameters and sampling frequency. When soil salinity is identified as >15 mS / cm or pH <5.5, a sampling frequency adjustment command is sent to the sensor array module, triggering the electrochemical corrosion sensor to switch from 1Hz to 5Hz and simultaneously activating the automatic soil sample collection device. Based on 5 years of data from 100+ substations, a digital twin model of material corrosion and a climate zoning feature library are established to provide scenario-based data for dynamic solution adaptation and operation and maintenance modules, and historical soil data for safety assessment and failure prediction modules.

[0123] The corrosion digital twin model uses finite element analysis combined with random forest algorithm to disclose data dimensions including grounding grid material (galvanized steel, stainless steel, etc.), climate zone (temperate, subtropical, etc.), soil parameters (salt, pH, temperature) and corresponding corrosion rate data.

[0124] Soil salinity-corrosion rate correlation model formula:

[0125]

[0126] In the formula: v corr The electrochemical corrosion rate of the grounding device is expressed in mm / a and is collected by an electrochemical corrosion sensor with a measurement range of 0 to 1 mm / a.

[0127] k1 represents the soil salinity correlation coefficient, with a value ranging from 0.08 to 0.15 mm / (a·mS / cm), and the upper limit is taken for high-salinity soil areas;

[0128] S represents the real-time monitored soil salinity, in mS / cm, collected by a multi-parameter soil sensor. When S > 15 mS / cm, the sampling frequency is adjusted.

[0129] k2 represents the soil pH correlation coefficient, with a value ranging from 0.05 to 0.09 mm / (a·pH), and the upper limit is used in acidic soil areas;

[0130] pH represents the real-time soil acidity and alkalinity, collected by a multi-parameter soil sensor. When pH < 5.5, the sampling frequency is adjusted.

[0131] k3 represents the soil temperature correction factor, which is fixed at 0.03 mm / a;

[0132] T represents the real-time monitored soil temperature, in °C, which is collected by a multi-parameter soil sensor.

[0133] T0 represents the soil reference temperature, in °C, and is set to 25 °C, which is the reference temperature for calibrating the corrosion rate in the laboratory.

[0134] The safety assessment and failure prediction module is as follows:

[0135] The system receives the associated feature data from the fusion analysis module and the historical soil data from the dynamic association database module. It combines the impulse grounding resistance, lightning current threshold, and equipment health to establish a dynamic weighted synergy coefficient model. Based on the LSTM model, it predicts two types of risks: local corrosion-excessive lightning protection residual voltage and soil pollution-grounding current dissipation failure.

[0136] The evaluation results and prediction information are pushed to the real-time early warning and visualization interaction platform module to trigger the corresponding early warning level. When the weight coordination coefficient is less than 0.5 or the failure probability reaches the alarm threshold, a control command is sent to the local cathodic protection device, and the control effect data is fed back to the dynamic association database module and the fusion analysis module.

[0137] Formula for dynamic weight coordination coefficient of grounding-lightning protection:

[0138]

[0139] Where: K d K represents the dynamic weighting coordination coefficient of grounding and lightning protection.d A value ≥0.8 is considered normal. K d A value less than 0.5 indicates an alarm status.

[0140] ω1 represents the impulse grounding resistance weighting factor, which is 0.3 during the rainy season and 0.6 during the period of high soil corrosion.

[0141] ω2 represents the weighting factor for lightning current withstand threshold, which is 0.5 during the rainy season and 0.2 during the period of high soil corrosion incidence.

[0142] R imp This represents the real-time monitored impulse grounding resistance, measured in Ω, and is collected by the lightning protection associated sensor group.

[0143] R imp0 This indicates the standard value of impulse grounding resistance, in Ω, which is determined according to the substation design specifications. For example, it is usually ≤10Ω for 220kV substations.

[0144] I light This represents the peak value of the lightning current monitored in real time, in kA, and is collected by the lightning current sensor.

[0145] I light0 This indicates the lightning current withstand threshold, expressed in kA, and is determined based on the rated parameters of the surge arrester. For example, it is typically set to 50 kA for zinc oxide surge arresters.

[0146] α represents the equipment health correction coefficient, which is calculated based on the service life of the grounding electrode. α = 1 - 0.01·T, where T is the service life of the grounding electrode in years. When T > 10 years, α is fixed at 0.9.

[0147] Formula for the failure probability of localized corrosion-lightning protection residual voltage exceeding the limit:

[0148]

[0149] In the formula: P represents the failure probability of lightning protection residual voltage exceeding the standard due to local corrosion, and the value ranges from 0 to 1. When P>0.7, an alarm is triggered;

[0150] 'a' represents the corrosion rate influence coefficient, with a value of 1.2 mm. -1 ·a;

[0151] v corr This represents the local grounding electrode corrosion rate, expressed in mm / a, and is collected from local node data by an electrochemical corrosion sensor.

[0152] b represents the influence coefficient of impulse grounding resistance deviation, with a value of 0.1Ω. -1 ;

[0153] R imp -R imp0This indicates the deviation value of the impulse grounding resistance, in Ω. A positive value indicates that the resistance is higher than the standard value, and the risk is increased.

[0154] c represents the influence coefficient of lightning current exceeding the limit, with a value of 3.5;

[0155] This indicates the percentage of lightning current exceeding the limit. A positive value indicates that the lightning current exceeds the withstand threshold, increasing the risk.

[0156] The real-time early warning and visualization interaction platform module is as follows:

[0157] It receives abnormal status information from the fusion analysis module and evaluation results from the safety assessment and failure prediction module, and is responsible for realizing the visualization and hierarchical early warning of the grounding grid status, providing an intuitive operation interface for operation and maintenance personnel; it establishes a 3D model of the grounding grid based on WebGIS to display sensor locations, real-time data, and safety status, and supports data drilling from substation to area to grounding electrode.

[0158] The system features a three-level early warning response, including an audible and visual warning pop-up, an alarm SMS notification, and an emergency alarm prompt to prohibit operation. The response time is less than 10 seconds. The system also pushes the early warning information and fault location data to the solution dynamic adaptation and operation and maintenance module simultaneously. It has the functions of data query, curve comparison, and report export. The exported historical data is supplemented to the dynamic association database module.

[0159] Formula for determining the warning level index:

[0160]

[0161] In the formula: S warn This represents the index for determining the warning level, with a value range of 0 to 1, where 0 ≤ S warn <0.3 is considered safe, 0.3≤S warn <0.7 is a warning sign, S warn A value ≥0.7 is considered an alarm.

[0162] F represents the output value of the multi-sensor fusion feature layer, which is calculated by the fusion analysis module and comprehensively reflects the degree of abnormality of electrical, corrosion and lightning protection parameters.

[0163] Th safe This represents the safety threshold, calibrated based on historical normal data, such as 0.2;

[0164] Th alarm This represents the alarm threshold, calibrated based on fault critical data, such as 0.8;

[0165] Th alarm -Th safe This represents the threshold range, used to normalize feature values ​​to an exponent of 0 to 1.

[0166] The specific details of the dynamic adaptation and operation and maintenance module are as follows:

[0167] It receives scenario-based data (partition feature model, digital twin simulation results) from the dynamic relational database module and fault warning information and fault location data from the real-time early warning and visualization interaction platform module. It is responsible for realizing the adaptation of anti-corrosion schemes and closed-loop management of operation and maintenance. Based on soil parameters, it outputs scenario-based anti-corrosion schemes. For high-salt soil, it recommends titanium alloy coating + zinc-magnesium sacrificial anode; for acidic soil, it recommends stainless steel + epoxy resin coating; and for ordinary soil, it recommends hot-dip galvanized steel. It calculates the replacement cycle of grounding electrode according to corrosion rate and links the operation and maintenance system to generate spare parts list and construction plan.

[0168] For high-salt soils, the titanium alloy coating thickness is 0.1-0.2mm, the zinc-magnesium sacrificial anode specifications are 50mm in diameter × 300mm in length, and the installation spacing is 5-8m;

[0169] For acidic soils, the stainless steel material is 304 stainless steel, with two layers of epoxy resin coating, and a curing temperature of 25-30℃.

[0170] The data on corrosion rate changes and risk reduction after the implementation of the solution are fed back to the dynamic correlation database module for iterating the digital twin model and the partition feature library. This data is then synchronized to the fusion analysis module to optimize feature extraction accuracy. Simultaneously, it provides post-maintenance equipment status data to the safety assessment and failure prediction modules, and corrects the dynamic weighted coordination coefficient model.

[0171] It should be noted that, in this document, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such process, method, article, or apparatus.

[0172] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.

Claims

1. A real-time safety monitoring system for substation main grounding grid based on multi-sensor fusion, characterized in that: The system includes: Sensor array module: Configured with electrical parameter sensor group, environmental-corrosion sensor group and lightning protection related sensor group. The electrical parameter sensor group includes grounding resistance sensor, ground potential distribution sensor and leakage current sensor, used to collect electrical performance data, environmental corrosion data and lightning protection status data. The environmental-corrosion sensor group includes a soil moisture sensor, a soil salinity sensor, a soil pH sensor, an electrochemical corrosion sensor, and a heavy metal sensor. The lightning protection-related sensor group includes an impulse grounding resistance sensor, a lightning current sensor, and a surge arrester action counter. Miniature industrial cameras are deployed at key nodes of the grounding grid to collect soil profile images. All sensors adopt a double-layer electromagnetic shielding design, and the collected raw data is transmitted to the acquisition and preprocessing module. Acquisition and preprocessing module: Upon receiving the raw data, dynamically adjusts the sampling frequency according to the operating conditions, performs noise reduction, converts the analog quantity into a digital quantity, encapsulates it into a standard data frame, and outputs standardized data; Fusion Analysis Module: Receives standardized data from the acquisition and preprocessing module and historical data from the dynamic correlation database module. It adopts a three-level fusion architecture: the data layer uses a weighted average method to fuse data from similar sensors to eliminate random errors, and calls historical data from the same period in the dynamic correlation database module to assist in the identification of abnormal data. The feature layer employs a hybrid model of graph neural network and LSTM to extract the topological and temporal features of the grounding grid. It also introduces cross-modal fusion of electrical and image data to normalize the grayscale features of the soil profile image with the electrical performance data, thereby strengthening the correlation between abnormal areas and parameter mutations. The decision layer combines an expert knowledge base to output three levels of status: safety, early warning, and alarm. The fusion analysis module is also configured with an abnormal data repair mechanism. When a sensor fails, it is completed by neighborhood interpolation or historical model. The repaired data is synchronously fed back to the dynamic association database module for supplementary storage. The extracted soil-corrosion and grounding-lightning protection association feature data are transmitted to the safety assessment and failure prediction module in real time. Abnormal status information is pushed to the real-time early warning and visualization interaction platform module. Dynamically related database module: Receives the standardized data, repaired data, and execution effect data. It adopts a hybrid architecture of MySQL and InfluxDB, has a built-in dynamic mapping table of soil parameters and sampling frequency, establishes a digital twin model of material corrosion and a climate zoning library, and outputs sampling frequency adjustment instructions and scenario-based data. Safety assessment and failure prediction module: Establish a dynamic weighted synergy coefficient model, predict two types of risks based on the LSTM model, output assessment results and prediction information, and provide feedback control effect instructions; Real-time early warning and visualization interaction platform module: Establish a three-dimensional model of the grounding grid, set three-level early warning response, and output early warning information and fault location data; Dynamic adaptation and operation and maintenance module: Receives the scenario-based data and the fault location data, outputs the scenario-based anti-corrosion solution, calculates the grounding electrode replacement cycle, and outputs the solution execution effect.

2. The real-time safety monitoring system for substation main grounding grid based on multi-sensor fusion as described in claim 1, characterized in that: The acquisition and preprocessing module is as follows: Upon receiving the raw data from the sensor array module, the acquisition and preprocessing module uses an FPGA as the main control chip and combines GPS timing to dynamically adjust the sampling frequency according to the working conditions. The electrical parameters are 1kHz, the conventional soil parameters are 1Hz, the abnormal soil conditions are 5Hz, and the lightning current is 1MHz. Adaptive notch filtering is used to remove 50Hz power frequency interference, and wavelet threshold denoising is used to process transient pulses. Analog quantities are converted into digital quantities and encapsulated into standard data frames of timestamp-sensor ID-parameter value. The processed standardized data is then synchronously pushed to the fusion analysis module and the dynamic association database module.

3. The real-time safety monitoring system for substation main grounding grid based on multi-sensor fusion as described in claim 2, characterized in that: The dynamic relational database module is specifically as follows: The system receives standardized data from the acquisition and preprocessing module, repaired data from the fusion analysis module, and execution effect data from the scheme dynamic adaptation and operation and maintenance module. It adopts a hybrid architecture of MySQL and InfluxDB. MySQL stores static data including substation location and grounding grid material, while InfluxDB stores dynamic data including soil parameters and corrosion rate. The system includes a built-in dynamic mapping table between soil parameters and sampling frequency. When soil salinity is identified as >15 mS / cm or pH <5.5, a sampling frequency adjustment command is sent to the sensor array module, and the automatic soil sample collection device is activated. A digital twin model of material corrosion and a climate zone feature library are established to provide scenario-based data for the dynamic adaptation and operation and maintenance module of the solution, and to provide historical soil data for the safety assessment and failure prediction module.

4. The real-time safety monitoring system for substation main grounding grid based on multi-sensor fusion as described in claim 3, characterized in that: The specific details of the safety assessment and failure prediction module are as follows: Upon receiving the associated feature data from the fusion analysis module and the historical soil data from the dynamic association database module, and combining the impulse grounding resistance, lightning current threshold, and equipment health, a dynamic weighted synergy coefficient model is established. Based on the LSTM model, two types of risks are predicted: local corrosion-excessive lightning protection residual voltage and soil pollution-grounding current dissipation failure. The evaluation results and prediction information are pushed to the real-time early warning and visualization interaction platform module to trigger the corresponding early warning level. When the weighted coordination coefficient is less than 0.5 or the failure probability reaches the alarm threshold, a control command is sent to the local cathodic protection device, and the control effect data is fed back to the dynamic association database module and the fusion analysis module.

5. The real-time safety monitoring system for substation main grounding grid based on multi-sensor fusion according to claim 4, characterized in that: The specific modules of the real-time early warning and visualization interaction platform are as follows: The system receives abnormal status information from the fusion analysis module and evaluation results from the safety assessment and failure prediction module, and establishes a three-dimensional model of the grounding grid based on WebGIS to display sensor locations, real-time data, and safety status. A three-level early warning response is set up, and the early warning information and fault location data are synchronously pushed to the dynamic adaptation and operation and maintenance module of the solution; it has the functions of data query, curve comparison and report export, and the exported historical data is supplemented to the dynamic association database module.

6. The real-time safety monitoring system for substation main grounding grid based on multi-sensor fusion according to claim 5, characterized in that: The specific details of the dynamic adaptation and operation and maintenance module of the solution are as follows: Upon receiving the scenario-based data from the dynamic association database module and the fault warning information and fault location data from the real-time early warning and visualization interaction platform module, the system outputs a scenario-based anti-corrosion scheme based on soil parameters, calculates the grounding electrode replacement cycle based on the corrosion rate, and generates a spare parts list and construction plan in conjunction with the operation and maintenance system. The corrosion rate change and risk reduction effect data after the implementation of the solution are fed back to the dynamic correlation database module and synchronized to the fusion analysis module. At the same time, it provides the safety assessment and failure prediction module with post-operation and maintenance equipment status data and corrects the dynamic weighted coordination coefficient model.

Images