A large grounding net corrosion state evaluation system based on big data analysis

Abstract

This invention discloses a large-scale grounding grid corrosion status assessment system based on big data analysis, comprising: a data acquisition module for collecting node resistance, potential, conductivity, humidity, and temperature, and recording time and node data; a data processing module for denoising, normalizing, and correcting data to generate a structured data matrix; a feature extraction module for extracting corrosion feature vectors based on historical records; a corrosion assessment module for probabilistic model inference and scoring to generate corrosion levels, risk matrices, and health trends; a corrosion distribution module for interpolation mapping to generate a full-network corrosion distribution map and marking high-risk nodes; a trend prediction module for simulating evolution to generate future corrosion level change curves and high-risk rankings; and a report generation module for generating a status assessment report. This invention utilizes big data analysis technology to achieve accurate corrosion assessment, risk visualization, and health trend prediction for grounding grids, improving maintenance decision-making efficiency.

Description

Technical Field

[0001] This invention relates to the field of corrosion monitoring technology for large grounding grids, and in particular to a corrosion status assessment system for large grounding grids based on big data analysis. Background Technology

[0002] Large-scale grounding grids are crucial infrastructure for ensuring the safe operation of power systems. Their nodes and lines are exposed to complex soil environments for extended periods, making them susceptible to corrosion due to variations in humidity, soil conductivity, chemical composition, and temperature. Existing methods for assessing the corrosion status of grounding grids primarily rely on manual inspections and single-point measurements, periodically collecting data such as node resistance and potential. However, these methods suffer from drawbacks such as long data collection cycles, limited coverage, small data volumes, and difficulty in achieving continuous monitoring, failing to accurately reflect the real-time corrosion status of the entire large-scale grounding grid.

[0003] In recent years, some studies have attempted to introduce information technology to collect node data through sensors and perform local statistical analysis. However, most of these methods only process single types of measurement data and lack the ability to comprehensively analyze multi-dimensional environmental factors and the historical state of nodes. In addition, existing assessment methods usually use static threshold judgments or simple linear models to classify corrosion levels, which cannot fully consider the spatial correlation and temporal evolution trends between nodes. This leads to uncertainty and bias in corrosion status determination, making it difficult to provide scientific risk ranking and prediction guidance for operation and maintenance.

[0004] With the gradual application of big data analytics and PHM (Property Management and Health) technologies to power system asset management, existing technologies still have the following shortcomings: The lack of synchronized processing of node-collected data in time and space leads to missing and inconsistent data matrices; corrosion feature extraction relies on traditional statistical methods, making it difficult to consider the influence of multiple factors and the correlation of high-dimensional features; the lack of continuity and visualization methods for the corrosion distribution across the entire network makes it impossible to clearly mark high-risk areas and risk boundaries; and the lack of dynamic models and spatial correlation analysis in corrosion evolution prediction makes it difficult to provide priority ranking references for maintenance decisions.

[0005] Therefore, how to provide a large-scale grounding grid corrosion status assessment system based on big data analysis is a problem that urgently needs to be solved by those skilled in the art. Summary of the Invention

[0006] One objective of this invention is to propose a large-scale grounding grid corrosion status assessment system based on big data analysis. This invention fully utilizes intelligent sensors, big data processing technology, weighted principal component analysis, and PHM health management algorithm. It details the technical process of multi-dimensional data acquisition, structured processing, corrosion-sensitive feature extraction, probabilistic model inference, visualization of the corrosion distribution across the entire grid, and dynamic evolution prediction. This system enables accurate assessment of the corrosion status of nodes in large-scale grounding grids and priority maintenance of high-risk areas. It has the advantages of high real-time performance, accurate assessment, intuitive visualization, and scientific maintenance decision-making.

[0007] A large-scale grounding grid corrosion status assessment system based on big data analysis according to an embodiment of the present invention includes: The data acquisition module is used to collect data on node resistance, potential, soil conductivity, humidity, and temperature, and to simultaneously record the acquisition time and node identifier. The data processing module is used to perform wavelet denoising and adaptive normalization on the collected data, and to generate a structured data matrix through a time series correction algorithm. The feature extraction module combines the structured data matrix with historical inspection records, uses weighted principal component analysis to extract corrosion-sensitive features, and constructs a set of corrosion feature vectors for the corresponding nodes. The corrosion assessment module is used to perform probabilistic model inference and state score calculation on the corrosion feature vector set, generate node corrosion level and risk index matrix, and calculate node health degradation trend. The corrosion distribution module is used to perform kriging interpolation and thermal mapping calculations in a spatial coordinate system using the node corrosion level and risk index matrix to generate a network-wide corrosion distribution map and mark high-risk nodes and risk boundaries. The trend prediction module is used to perform corrosion evolution simulation and generate corrosion level change curves at future time points and a priority maintenance ranking table for high-risk areas. The report generation module is used to generate a status assessment report, which includes a structured data matrix, corrosion feature tensor, a network-wide corrosion distribution map, a risk index matrix, and maintenance and optimization suggestions.

[0008] Optionally, modules can be integrated using the following methods: S1. Deploy smart sensors at grounding grid nodes to collect node resistance, potential, soil conductivity, humidity and temperature data, and simultaneously record the collection time and node identification. S2. Perform wavelet denoising and adaptive normalization on the collected data, and generate a structured data matrix through a time series correction algorithm to ensure data synchronization and consistency between different sensor nodes; S3. Combine the structured data matrix with historical inspection records, use weighted principal component analysis to extract corrosion-sensitive features, and construct a set of corrosion feature vectors for the corresponding nodes. S4. Based on the PHM health management algorithm, perform probabilistic model inference and state score calculation on the corrosion feature vector set to generate node corrosion level and risk index matrix, and calculate node health decay trend at the same time. S5. Using the node corrosion level and risk index matrix, perform Kriging interpolation and thermal mapping calculations in the spatial coordinate system to generate a network corrosion distribution map and mark high-risk nodes and risk boundaries. S6. Combining the node health decay trend and PHM dynamic prediction model, perform corrosion evolution simulation to generate node corrosion level change curves and priority maintenance ranking table for high-risk areas at future time points. S7. Generate a status assessment report, including a structured data matrix, corrosion feature tensor, a network-wide corrosion distribution map, a risk index matrix, and maintenance and optimization suggestions.

[0009] Optionally, the time series correction algorithm achieves data synchronization and consistency by aligning timestamps and unifying sampling frequencies of node data collected by different smart sensors, specifically including: Based on the data collection timestamps of each node, a global time axis is constructed, and data with inconsistent timestamps are mapped to a unified time point through linear interpolation or spline interpolation. For data with missing or delayed sampling, forward padding and weighted moving average processing are performed to ensure continuity and smoothness; To address the time-series offset between sensor nodes, the time delay is calculated using a cross-correlation function, and the delayed data is corrected accordingly to ensure that the data between nodes are synchronized and aligned on the global time axis. The corrected time series data is normalized to unify the dimensions and ranges of different nodes to a standard interval, forming a structured data matrix.

[0010] Optionally, S3 specifically includes: S31. Perform node-by-node comparison between the structured data matrix and historical inspection records, establish data mapping relationship based on node number and collection time, interpolate and fill missing or abnormal historical data and remove anomalies to form a complete joint data table. S32. Aggregate the sensor feature values ​​of various types in the joint data table by node, calculate the variance and weight coefficient of each feature, calculate the covariance matrix by weighted principal component analysis, and perform eigenvalue decomposition to extract principal component features sensitive to corrosion state, and generate corrosion principal component vector for each node. S33. Normalize the principal component vector of each node to unify the vector dimension and numerical range. Construct a set of corrosion feature vectors for each node through the vector set and index and store them according to the node spatial layout for corrosion assessment and health trend calculation.

[0011] Optionally, the calculation scheme of the weighted principal component analysis method specifically includes: Based on the sensor characteristics of each node in the joint data table, the mean and variance of each characteristic are calculated, and the variance is used as the initial weighting factor. At the same time, outliers are corrected by weighting to form a weighted feature factor. The weighted feature factor is applied to the structured data matrix, and the feature vector of each node is multiplied according to the corresponding weight to obtain the weighted feature matrix; The weighted feature matrix is ​​centered column by column, and the weighted mean is subtracted from each feature to form a zero-mean matrix. The weighting factors are then incorporated into the covariance calculation to obtain the covariance matrix. Eigenvalue decomposition is performed on the covariance matrix to obtain eigenvalues ​​and corresponding eigenvectors. The eigenvalues ​​are sorted according to their contribution rate, and the top few eigenvectors whose cumulative contribution rate reaches a preset threshold are selected as corrosion-sensitive principal components. The selected principal component vectors are orthogonalized to ensure that each principal component is independent, and the principal components are locally weighted and adjusted in combination with the node space layout. The weighted principal component vector of each node is normalized to unify the range of vector values, and a set of erosion feature vectors is constructed according to the node index.

[0012] Optionally, S4 specifically includes: S41. Input the corrosion feature vector set into the PHM health management algorithm, construct a multi-dimensional state vector based on node features, and initialize the prior parameters and state transition matrix of the probability model based on the node's historical corrosion records. S42. Perform Bayesian probability model inference on each node state vector. By calculating the conditional probability of the current state vector and the historical state vector of the node, obtain the probability distribution of the node under different corrosion levels, and normalize the probability values ​​to obtain the corresponding node corrosion level. S43. Based on the node corrosion level and combined with the node's state change rate in the time series, calculate the node risk index, multiply the corrosion level probability by the node state evolution trend to form a weighted risk value, and at the same time perform smoothing filtering on nodes with abnormal fluctuations to generate a risk index matrix. S44. Perform time series fitting on the risk index matrix, use the exponential weighted moving average method and cumulative decay coefficient to calculate the node health decay trend curve, and generate a health decay trend set according to the node spatial layout to record the future corrosion development trend and potential risk changes of each node. S45. Store the node corrosion level, risk index matrix and health decay trend set in a unified database and create an index table, while ensuring the time sequence consistency between the data of each node and the historical inspection records.

[0013] Optionally, the inference process of the Bayesian probability model specifically includes: The node state vector is normalized according to the feature weights, and a dynamic weight adjustment mechanism is introduced so that the weights of sensitive features can change adaptively under different environmental conditions. A Markov model is constructed to map the evolution sequence of node historical corrosion levels into a state transition matrix, and a time decay factor and a spatial correlation factor are introduced into the calculation of the transition probability. In the state inference phase, the Bayesian update algorithm is used to iteratively calculate the posterior probability of the current node state, combining the current observation data with the historical state probabilities. At the same time, during the iteration process, adaptive smoothing filtering is applied to abnormal and sudden data to suppress the interference of noise on the probability calculation. The node corrosion level is determined by the maximum a posteriori probability principle, and the uncertainty of the node state is quantified by the variance index.

[0014] Optionally, S5 specifically includes: S51. Map the corrosion level and risk index matrix of each node to a spatial coordinate system, and calculate the spatial weight coefficient based on the geographical distance and electrical connectivity between nodes to provide a weighted basis for interpolation calculations. S52. Perform Kriging interpolation on node attributes, using corrosion level and risk index as variables, and calculate the predicted value of each non-measurement point by combining spatial weight coefficients. Use the variogram adaptive fitting method to ensure that the interpolation process can maintain accuracy in areas with low node density or uneven distribution. S53. Perform thermal mapping calculation on the interpolation results, map the corrosion level and risk index into color gradient values, form a visualized full network corrosion distribution map, mark the nodes whose risk values ​​exceed the preset threshold in the map, and generate risk boundary lines along the continuous high-risk areas. S54. The heat mapping results are smoothed and the resolution is optimized to ensure that high-risk nodes and boundaries remain clearly distinguishable at different scaling levels, covering both vector and raster data formats.

[0015] Optionally, S6 specifically includes: S61. Input the set of node health decay trends and the matrix of node corrosion level and risk index into the PHM dynamic prediction model respectively. Construct a state space model based on the node state vector and historical evolution data, and initialize the evolution simulation parameters, including time step, prediction window and node weight coefficient. S62. During the prediction process, a multi-step recursive prediction is performed on the corrosion level of each node, and the state transition calculation is performed in combination with the spatial correlation between nodes. The current state probability distribution is dynamically fused with the risk information of neighboring nodes to generate a corrosion level probability vector for future time points. S63. Determine the maximum a posteriori value of the corrosion level probability vector to form the node corrosion level change curve. At the same time, calculate the future risk index of each node and identify high-risk areas according to the spatial clustering method to form a dynamically evolving high-risk area map. S64. Combining the node corrosion level change curve and the high-risk area map, prioritize the nodes for maintenance. Use a weighted scoring method to generate maintenance priority values ​​by comprehensively considering corrosion level, risk index, health decay rate and spatial proximity, and output a priority maintenance ranking table for high-risk areas. S65. Store the corrosion evolution simulation results, node corrosion level change curves, and high-risk area priority maintenance sorting table in the database, and associate them with the network-wide corrosion distribution map and historical inspection records.

[0016] The beneficial effects of this invention are: First, this invention uses intelligent sensors to collect multi-dimensional data such as resistance, potential, soil conductivity, humidity and temperature of large grounding grid nodes in real time, and uses wavelet denoising, adaptive normalization and time series correction algorithms to generate a structured data matrix, thereby achieving high-precision synchronization and consistency of data and improving the reliability of corrosion status monitoring of the entire network. Secondly, this invention uses weighted principal component analysis to extract node corrosion features, and combines the PHM health management algorithm to perform probabilistic model inference and state score calculation to generate node corrosion level, risk index matrix and health decay trend, which effectively improves the accuracy of corrosion state assessment and dynamic prediction capability. At the same time, it generates a network-wide corrosion distribution map through Kriging interpolation and thermal mapping, and marks high-risk nodes and risk boundaries, realizing the visualization and spatial continuity of corrosion information. Finally, this invention combines the node health decay trend and PHM dynamic prediction model to simulate corrosion evolution, generating node corrosion level change curves at future time points and a priority maintenance ranking table for high-risk areas, providing a scientific basis for operation and maintenance decisions, realizing the prediction of corrosion evolution trends and the optimized allocation of maintenance resources, thus having significant beneficial effects of high real-time performance, high accuracy and strong operability in the safe operation management of large grounding grids. Attached Figure Description

[0017] The accompanying drawings are provided to further illustrate the invention and form part of the specification. They are used in conjunction with embodiments of the invention to explain the invention and do not constitute a limitation thereof. In the drawings: Figure 1This is a module structure diagram of a large-scale grounding grid corrosion status assessment system based on big data analysis proposed in this invention. Figure 2 This is a flowchart of a method for assessing the corrosion status of a large grounding grid based on big data analysis, as proposed in this invention. Figure 3 This is a schematic diagram illustrating the probabilistic model inference and corrosion status assessment of a large-scale grounding grid corrosion status assessment system based on big data analysis proposed in this invention. Detailed Implementation

[0018] The present invention will now be described in further detail with reference to the accompanying drawings. These drawings are simplified schematic diagrams, illustrating only the basic structure of the invention, and therefore only show the components relevant to the invention.

[0019] refer to Figure 1 A large-scale grounding grid corrosion status assessment system based on big data analysis includes: The data acquisition module is used to collect data on node resistance, potential, soil conductivity, humidity, and temperature, and to simultaneously record the acquisition time and node identifier. The data processing module is used to perform wavelet denoising and adaptive normalization on the collected data, and to generate a structured data matrix through a time series correction algorithm. The feature extraction module combines the structured data matrix with historical inspection records, uses weighted principal component analysis to extract corrosion-sensitive features, and constructs a set of corrosion feature vectors for the corresponding nodes. The corrosion assessment module is used to perform probabilistic model inference and state score calculation on the corrosion feature vector set, generate node corrosion level and risk index matrix, and calculate node health degradation trend. The corrosion distribution module is used to perform kriging interpolation and thermal mapping calculations in a spatial coordinate system using the node corrosion level and risk index matrix to generate a network-wide corrosion distribution map and mark high-risk nodes and risk boundaries. The trend prediction module is used to perform corrosion evolution simulation and generate corrosion level change curves at future time points and a priority maintenance ranking table for high-risk areas. The report generation module is used to generate a status assessment report, which includes a structured data matrix, corrosion feature tensor, a network-wide corrosion distribution map, a risk index matrix, and maintenance and optimization suggestions.

[0020] refer to Figure 2-3 In this embodiment, the modules are interconnected using the following method: S1. Deploy smart sensors at grounding grid nodes to collect node resistance, potential, soil conductivity, humidity and temperature data, and simultaneously record the collection time and node identification. S2. Perform wavelet denoising and adaptive normalization on the collected data, and generate a structured data matrix through a time series correction algorithm to ensure data synchronization and consistency between different sensor nodes; S3. Combine the structured data matrix with historical inspection records, use weighted principal component analysis to extract corrosion-sensitive features, and construct a set of corrosion feature vectors for the corresponding nodes. S4. Based on the PHM health management algorithm, perform probabilistic model inference and state score calculation on the corrosion feature vector set to generate node corrosion level and risk index matrix, and calculate node health decay trend at the same time. S5. Using the node corrosion level and risk index matrix, perform Kriging interpolation and thermal mapping calculations in the spatial coordinate system to generate a network corrosion distribution map and mark high-risk nodes and risk boundaries. S6. Combining the node health decay trend and PHM dynamic prediction model, perform corrosion evolution simulation to generate node corrosion level change curves and priority maintenance ranking table for high-risk areas at future time points. S7. Generate a status assessment report, including a structured data matrix, corrosion feature tensor, a network-wide corrosion distribution map, a risk index matrix, and maintenance and optimization suggestions.

[0021] In this embodiment, the time series correction algorithm achieves data synchronization and consistency by aligning timestamps and unifying sampling frequencies of node data collected by different smart sensors. Specifically, it includes: Based on the data collection timestamps of each node, a global time axis is constructed, and data with inconsistent timestamps are mapped to a unified time point through linear interpolation or spline interpolation. For data with missing or delayed sampling, forward padding and weighted moving average processing are performed to ensure continuity and smoothness; To address the time-series offset between sensor nodes, the time delay is calculated using a cross-correlation function, and the delayed data is corrected accordingly to ensure that the data between nodes are synchronized and aligned on the global time axis. The corrected time series data is normalized to unify the dimensions and ranges of different nodes to a standard interval, forming a structured data matrix.

[0022] In this embodiment, S3 specifically includes: S31. Perform node-by-node comparison between the structured data matrix and historical inspection records, establish data mapping relationship based on node number and collection time, interpolate and fill missing or abnormal historical data and remove anomalies to form a complete joint data table. S32. Aggregate the sensor feature values ​​of various types in the joint data table by node, calculate the variance and weight coefficient of each feature, calculate the covariance matrix by weighted principal component analysis, and perform eigenvalue decomposition to extract principal component features sensitive to corrosion state, and generate corrosion principal component vector for each node. S33. Normalize the principal component vector of each node to unify the vector dimension and numerical range. Construct a set of corrosion feature vectors for each node through the vector set and index and store them according to the node spatial layout for corrosion assessment and health trend calculation.

[0023] In this embodiment, the calculation scheme of the weighted principal component analysis method specifically includes: Based on the sensor characteristics of each node in the joint data table, the mean and variance of each characteristic are calculated, and the variance is used as the initial weighting factor. At the same time, outliers are corrected by weighting to form a weighted feature factor. The weighted feature factor is applied to the structured data matrix, and the feature vector of each node is multiplied according to the corresponding weight to obtain the weighted feature matrix; The weighted feature matrix is ​​centered column by column, and the weighted mean is subtracted from each feature to form a zero-mean matrix. The weighting factors are then incorporated into the covariance calculation to obtain the covariance matrix. Eigenvalue decomposition is performed on the covariance matrix to obtain eigenvalues ​​and corresponding eigenvectors. The eigenvalues ​​are sorted according to their contribution rate, and the top few eigenvectors whose cumulative contribution rate reaches a preset threshold are selected as corrosion-sensitive principal components. The selected principal component vectors are orthogonalized to ensure that each principal component is independent, and the principal components are locally weighted and adjusted in combination with the node space layout. The weighted principal component vector of each node is normalized to unify the range of vector values, and a set of erosion feature vectors is constructed according to the node index.

[0024] In this embodiment, S4 specifically includes: S41. Input the corrosion feature vector set into the PHM health management algorithm, construct a multi-dimensional state vector based on node features, and initialize the prior parameters and state transition matrix of the probability model based on the node's historical corrosion records. S42. Perform Bayesian probability model inference on each node state vector. By calculating the conditional probability of the current state vector and the historical state vector of the node, obtain the probability distribution of the node under different corrosion levels, and normalize the probability values ​​to obtain the corresponding node corrosion level. S43. Based on the node corrosion level and combined with the node's state change rate in the time series, calculate the node risk index, multiply the corrosion level probability by the node state evolution trend to form a weighted risk value, and at the same time perform smoothing filtering on nodes with abnormal fluctuations to generate a risk index matrix. S44. Perform time series fitting on the risk index matrix, use the exponential weighted moving average method and cumulative decay coefficient to calculate the node health decay trend curve, and generate a health decay trend set according to the node spatial layout to record the future corrosion development trend and potential risk changes of each node. S45. Store the node corrosion level, risk index matrix and health decay trend set in a unified database and create an index table, while ensuring the time sequence consistency between the data of each node and the historical inspection records.

[0025] In this embodiment, the inference process of the Bayesian probability model specifically includes: The node state vector is normalized according to the feature weights, and a dynamic weight adjustment mechanism is introduced so that the weights of sensitive features can change adaptively under different environmental conditions. A Markov model is constructed to map the evolution sequence of node historical corrosion levels into a state transition matrix, and a time decay factor and a spatial correlation factor are introduced into the calculation of the transition probability. In the state inference phase, the Bayesian update algorithm is used to iteratively calculate the posterior probability of the current node state, combining the current observation data with the historical state probabilities. At the same time, during the iteration process, adaptive smoothing filtering is applied to abnormal and sudden data to suppress the interference of noise on the probability calculation. The node corrosion level is determined by the maximum a posteriori probability principle, and the uncertainty of the node state is quantified by the variance index.

[0026] In this embodiment, S5 specifically includes: S51. Map the corrosion level and risk index matrix of each node to a spatial coordinate system, and calculate the spatial weight coefficient based on the geographical distance and electrical connectivity between nodes to provide a weighted basis for interpolation calculations. S52. Perform Kriging interpolation on node attributes, using corrosion level and risk index as variables, and calculate the predicted value for each non-measurement point by combining spatial weighting coefficients. Employ an adaptive fitting method using the variogram function to maintain accuracy in areas with low or uneven node density. Specifically, this includes: An initial semivariance matrix is ​​constructed based on the spatial coordinates, corrosion level, and risk index of all network nodes. A weighting factor is introduced to the spatial correlation between nodes in different regions so that the influence of sparse or edge nodes can be reasonably amplified or attenuated. By using the adaptive variogram fitting method, the semivariance matrix is ​​fitted in a regional segment. The fitting parameters of each segment are adjusted according to the local node density and attribute variance, which significantly improves the interpolation accuracy in sparse node regions. In the weighting coefficient calculation stage, the node geographical distance, soil conductivity difference and electrical connectivity are integrated into a comprehensive weight, and a regularization term is added when solving the Kriging equation to ensure matrix invertibility and computational stability. For each prediction point, the interpolation result is calculated by weighted linear combination of the attribute values ​​of neighboring nodes, while the set of neighboring nodes is dynamically updated to make the interpolation in the high gradient change region smoother and more accurate. The interpolation results are corrected for errors by comparing them with the actual measured values ​​of the nodes, and local iterative optimization is performed on areas with large deviations. S53. Perform thermal mapping calculation on the interpolation results, map the corrosion level and risk index into color gradient values, form a visualized full network corrosion distribution map, mark the nodes whose risk values ​​exceed the preset threshold in the map, and generate risk boundary lines along the continuous high-risk areas. S54. The heat mapping results are smoothed and the resolution is optimized to ensure that high-risk nodes and boundaries remain clearly distinguishable at different scaling levels, covering both vector and raster data formats.

[0027] In this embodiment, S6 specifically includes: S61. Input the set of node health decay trends and the matrix of node corrosion level and risk index into the PHM dynamic prediction model respectively. Construct a state space model based on the node state vector and historical evolution data, and initialize the evolution simulation parameters, including time step, prediction window and node weight coefficient. S62. During the prediction process, a multi-step recursive prediction is performed on the corrosion level of each node, and the state transition calculation is performed in combination with the spatial correlation between nodes. The current state probability distribution is dynamically fused with the risk information of neighboring nodes to generate a corrosion level probability vector for future time points. S63. Determine the maximum a posteriori value of the corrosion level probability vector to form the node corrosion level change curve. At the same time, calculate the future risk index of each node and identify high-risk areas according to the spatial clustering method to form a dynamically evolving high-risk area map. S64. Combining the node corrosion level change curve and the high-risk area map, prioritize the nodes for maintenance. Use a weighted scoring method to generate maintenance priority values ​​by comprehensively considering corrosion level, risk index, health decay rate and spatial proximity, and output a priority maintenance ranking table for high-risk areas. S65. Store the corrosion evolution simulation results, node corrosion level change curves, and high-risk area priority maintenance sorting table in the database, and associate them with the network-wide corrosion distribution map and historical inspection records. Example 1:

[0028] To verify the feasibility of this invention in practice, it was applied to a large-scale grounding grid operating environment to comprehensively assess the corrosion status of its nodes. This grounding grid consists of multiple power facility nodes, which are exposed to complex variations in humidity, soil conductivity, temperature, and chemical composition over long periods. Traditional manual inspections struggle to comprehensively grasp the corrosion status of the entire grid, resulting in insufficient data coverage, inaccurate assessments, and difficulty in predicting high-risk areas. In this scenario, this invention deploys intelligent sensors at each node to collect real-time data on node resistance, potential, soil conductivity, humidity, and temperature, while simultaneously recording the collection time and node identifier to ensure the continuity and integrity of data acquisition. The collected data undergoes wavelet denoising and adaptive normalization, and is further processed using a time-series correction algorithm for timestamp alignment and sampling frequency unification, generating a structured data matrix for all nodes in the grid. This ensures that data from different nodes can be directly compared in terms of time and dimensions.

[0029] After obtaining the structured data matrix, this invention combines it with historical inspection records and uses weighted principal component analysis to extract corrosion-sensitive features. This method weights multi-dimensional data such as node resistance, potential, soil conductivity, humidity, and temperature, prioritizing features highly correlated with corrosion risk while reducing the impact of noise features, thus constructing a set of corrosion feature vectors for each node. Based on this, the PHM health management algorithm is used to perform probabilistic model inference and state scoring calculations, generating node corrosion levels, risk index matrices, and health degradation trends. By performing Kriging interpolation and thermal mapping calculations on the corrosion levels and risk indices of all nodes in the network, this invention generates a network-wide corrosion distribution map and marks high-risk nodes and risk boundaries, ensuring that maintenance personnel can clearly identify potential corrosion hotspots.

[0030] By further combining node health degradation trends and the PHM dynamic prediction model, this invention simulates the evolution of node corrosion levels at future time points, generating corrosion level change curves and a priority maintenance ranking table for high-risk areas. Simulation results show that for nodes with a risk index higher than 0.7, the probability of corrosion level increase within the next six months reaches 85%, while for nodes with a risk index lower than 0.3, the probability of corrosion level change within the next six months is only 12%. Through the priority maintenance ranking table, maintenance personnel can scientifically allocate resources and prioritize the maintenance of high-risk nodes, thereby reducing the risk of network-wide failures and maintenance costs. Throughout the implementation process, this invention achieves full data collection coverage, accurate extraction of corrosion-sensitive features, visualized network-wide distribution, and prediction of future corrosion evolution, solving the problem that traditional methods struggle to achieve continuous network-wide monitoring and high-precision prediction.

[0031] Table 1. Example of Corrosion Status of Nodes Across the Entire Network

[0032] Table 1 shows the multidimensional measurements of different nodes at the same time period, the extracted corrosion-sensitive feature vectors, and the corrosion level and risk index calculated based on the PHM health management algorithm. These data can be used to generate a network-wide heat map and evolution prediction curve, thereby guiding maintenance decisions and prioritizing the treatment of high-risk nodes.

[0033] The above description is only a preferred embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any equivalent substitutions or modifications made by those skilled in the art within the scope of the technology disclosed in the present invention, based on the technical solution and inventive concept of the present invention, should be covered within the scope of protection of the present invention.

Claims

1. A large-scale grounding grid corrosion status assessment system based on big data analysis, characterized in that, include: The data acquisition module is used to collect data on node resistance, potential, soil conductivity, humidity, and temperature, and to simultaneously record the acquisition time and node identifier. The data processing module is used to perform wavelet denoising and adaptive normalization on the collected data, and to generate a structured data matrix through a time series correction algorithm. The feature extraction module combines the structured data matrix with historical inspection records, uses weighted principal component analysis to extract corrosion-sensitive features, and constructs a set of corrosion feature vectors for the corresponding nodes. The corrosion assessment module is used to perform probabilistic model inference and state score calculation on the corrosion feature vector set, generate node corrosion level and risk index matrix, and calculate node health degradation trend. The corrosion distribution module is used to perform kriging interpolation and thermal mapping calculations in a spatial coordinate system using the node corrosion level and risk index matrix to generate a network-wide corrosion distribution map and mark high-risk nodes and risk boundaries. The trend prediction module is used to perform corrosion evolution simulation and generate corrosion level change curves at future time points and a priority maintenance ranking table for high-risk areas. The report generation module is used to generate a status assessment report, which includes a structured data matrix, corrosion feature tensor, a network-wide corrosion distribution map, a risk index matrix, and maintenance and optimization suggestions.

2. The large-scale grounding grid corrosion status assessment system based on big data analysis according to claim 1, characterized in that, The modules are connected in the following way: S1. Deploy smart sensors at grounding grid nodes to collect node resistance, potential, soil conductivity, humidity and temperature data, and simultaneously record the collection time and node identification. S2. Perform wavelet denoising and adaptive normalization on the collected data, and generate a structured data matrix through a time series correction algorithm to ensure data synchronization and consistency between different sensor nodes; S3. Combine the structured data matrix with historical inspection records, use weighted principal component analysis to extract corrosion-sensitive features, and construct a set of corrosion feature vectors for the corresponding nodes. S4. Based on the PHM health management algorithm, perform probabilistic model inference and state score calculation on the corrosion feature vector set to generate node corrosion level and risk index matrix, and calculate node health decay trend at the same time. S5. Using the node corrosion level and risk index matrix, perform Kriging interpolation and thermal mapping calculations in the spatial coordinate system to generate a network corrosion distribution map and mark high-risk nodes and risk boundaries. S6. Combining the node health decay trend and PHM dynamic prediction model, perform corrosion evolution simulation to generate node corrosion level change curves and priority maintenance ranking table for high-risk areas at future time points. S7. Generate a status assessment report, including a structured data matrix, corrosion feature tensor, a network-wide corrosion distribution map, a risk index matrix, and maintenance and optimization suggestions.

3. The large-scale grounding grid corrosion status assessment system based on big data analysis according to claim 2, characterized in that, The time series correction algorithm achieves data synchronization and consistency by aligning timestamps and unifying sampling frequencies of node data collected from different smart sensors. Specifically, it includes: Based on the data collection timestamps of each node, a global time axis is constructed, and data with inconsistent timestamps are mapped to a unified time point through linear interpolation or spline interpolation. For data with missing or delayed sampling, forward padding and weighted moving average processing are performed to ensure continuity and smoothness; To address the time-series offset between sensor nodes, the time delay is calculated using a cross-correlation function, and the delayed data is corrected accordingly to ensure that the data between nodes are synchronized and aligned on the global time axis. The corrected time series data is normalized to unify the dimensions and ranges of different nodes to a standard interval, forming a structured data matrix.

4. The large-scale grounding grid corrosion status assessment system based on big data analysis according to claim 2, characterized in that, S3 specifically includes: S31. Perform node-by-node comparison between the structured data matrix and historical inspection records, establish data mapping relationship based on node number and collection time, interpolate and fill missing or abnormal historical data and remove anomalies to form a complete joint data table. S32. Aggregate the sensor feature values ​​of various types in the joint data table by node, calculate the variance and weight coefficient of each feature, calculate the covariance matrix by weighted principal component analysis, and perform eigenvalue decomposition to extract principal component features sensitive to corrosion state, and generate corrosion principal component vector for each node. S33. Normalize the principal component vector of each node to unify the vector dimension and numerical range. Construct a set of corrosion feature vectors for each node through the vector set and index and store them according to the node spatial layout for corrosion assessment and health trend calculation.

5. The large-scale grounding grid corrosion status assessment system based on big data analysis according to claim 4, characterized in that, The calculation scheme of the weighted principal component analysis method specifically includes: Based on the sensor characteristics of each node in the joint data table, the mean and variance of each characteristic are calculated, and the variance is used as the initial weighting factor. At the same time, outliers are corrected by weighting to form a weighted feature factor. The weighted feature factor is applied to the structured data matrix, and the feature vector of each node is multiplied according to the corresponding weight to obtain the weighted feature matrix; The weighted feature matrix is ​​centered column by column, and the weighted mean is subtracted from each feature to form a zero-mean matrix. The weighting factors are then incorporated into the covariance calculation to obtain the covariance matrix. Eigenvalue decomposition is performed on the covariance matrix to obtain eigenvalues ​​and corresponding eigenvectors. The eigenvalues ​​are sorted according to their contribution rate, and the top few eigenvectors whose cumulative contribution rate reaches a preset threshold are selected as corrosion-sensitive principal components. The selected principal component vectors are orthogonalized to ensure that each principal component is independent, and the principal components are locally weighted and adjusted in combination with the node space layout. The weighted principal component vector of each node is normalized to unify the range of vector values, and a set of erosion feature vectors is constructed according to the node index.

6. The large-scale grounding grid corrosion status assessment system based on big data analysis according to claim 2, characterized in that, S4 specifically includes: S41. Input the corrosion feature vector set into the PHM health management algorithm, construct a multi-dimensional state vector based on node features, and initialize the prior parameters and state transition matrix of the probability model based on the node's historical corrosion records. S42. Perform Bayesian probability model inference on each node state vector. By calculating the conditional probability of the current state vector and the historical state vector of the node, obtain the probability distribution of the node under different corrosion levels, and normalize the probability values ​​to obtain the corresponding node corrosion level. S43. Based on the node corrosion level and combined with the node's state change rate in the time series, calculate the node risk index, multiply the corrosion level probability by the node state evolution trend to form a weighted risk value, and at the same time perform smoothing filtering on nodes with abnormal fluctuations to generate a risk index matrix. S44. Perform time series fitting on the risk index matrix, use the exponential weighted moving average method and cumulative decay coefficient to calculate the node health decay trend curve, and generate a health decay trend set according to the node spatial layout to record the future corrosion development trend and potential risk changes of each node. S45. Store the node corrosion level, risk index matrix and health decay trend set in a unified database and create an index table, while ensuring the time sequence consistency between the data of each node and the historical inspection records.

7. The large-scale grounding grid corrosion status assessment system based on big data analysis according to claim 6, characterized in that, The inference process of the Bayesian probability model specifically includes: The node state vector is normalized according to the feature weights, and a dynamic weight adjustment mechanism is introduced so that the weights of sensitive features can change adaptively under different environmental conditions. A Markov model is constructed to map the evolution sequence of node historical corrosion levels into a state transition matrix, and a time decay factor and a spatial correlation factor are introduced into the calculation of the transition probability. In the state inference phase, the Bayesian update algorithm is used to iteratively calculate the posterior probability of the current node state, combining the current observation data with the historical state probabilities. At the same time, during the iteration process, adaptive smoothing filtering is applied to abnormal and sudden data to suppress the interference of noise on the probability calculation. The node corrosion level is determined by the maximum a posteriori probability principle, and the uncertainty of the node state is quantified by the variance index.

8. The large-scale grounding grid corrosion status assessment system based on big data analysis according to claim 2, characterized in that, S5 specifically includes: S51. Map the corrosion level and risk index matrix of each node to a spatial coordinate system, and calculate the spatial weight coefficient based on the geographical distance and electrical connectivity between nodes to provide a weighted basis for interpolation calculations. S52. Perform Kriging interpolation on node attributes, using corrosion level and risk index as variables, and calculate the predicted value of each non-measurement point by combining spatial weight coefficients. Use the variogram adaptive fitting method to ensure that the interpolation process can maintain accuracy in areas with low node density or uneven distribution. S53. Perform thermal mapping calculation on the interpolation results, map the corrosion level and risk index into color gradient values, form a visualized full network corrosion distribution map, mark the nodes whose risk values ​​exceed the preset threshold in the map, and generate risk boundary lines along the continuous high-risk areas. S54. The heat mapping results are smoothed and the resolution is optimized to ensure that high-risk nodes and boundaries remain clearly distinguishable at different scaling levels, covering both vector and raster data formats.

9. A large-scale grounding grid corrosion status assessment system based on big data analysis according to claim 2, characterized in that, S6 specifically includes: S61. Input the set of node health decay trends and the matrix of node corrosion level and risk index into the PHM dynamic prediction model respectively. Construct a state space model based on the node state vector and historical evolution data, and initialize the evolution simulation parameters, including time step, prediction window and node weight coefficient. S62. During the prediction process, a multi-step recursive prediction is performed on the corrosion level of each node, and the state transition calculation is performed in combination with the spatial correlation between nodes. The current state probability distribution is dynamically fused with the risk information of neighboring nodes to generate a corrosion level probability vector for future time points. S63. Determine the maximum a posteriori value of the corrosion level probability vector to form the node corrosion level change curve. At the same time, calculate the future risk index of each node and identify high-risk areas according to the spatial clustering method to form a dynamically evolving high-risk area map. S64. Combining the node corrosion level change curve and the high-risk area map, prioritize the nodes for maintenance. Use a weighted scoring method to generate maintenance priority values ​​by comprehensively considering corrosion level, risk index, health decay rate and spatial proximity, and output a priority maintenance ranking table for high-risk areas. S65. Store the corrosion evolution simulation results, node corrosion level change curves, and high-risk area priority maintenance sorting table in the database, and associate them with the network-wide corrosion distribution map and historical inspection records.

Images