Large power grid grounding body corrosion condition estimation system

Abstract

The invention provides a large power grid grounding body corrosion condition estimation system, and belongs to the technical field of power grid grounding. Comprising a carrier signal generation module, a signal injection module, a multi-channel signal acquisition module, a signal conditioning module, a voltage / current measurement module, a soil parameter detection module, a temperature detection module, a GPS positioning module, a wireless communication module and a 32-bit control chip. Wherein the carrier signal generation module generates a carrier signal with a specific frequency, and the carrier signal is injected into a grounding network after power amplification and impedance matching of the signal injection module; the multi-channel signal acquisition module is provided with more than eight sampling channels and a 16-bit AD converter, the sampling rate exceeds 100 kHz, and the signals are processed by the signal conditioning module and then input to the control chip. A corrosion estimation software module is arranged in the control chip and is used for corrosion detection and estimation; the problems that in the prior art, it is difficult to comprehensively and quantitatively evaluate the corrosion condition of the large-scale earth screen grounding body and predict the future corrosion trend are solved.

Description

Technical Field

[0001] The present invention belongs to the technical field of power grid grounding, and specifically relates to a system for predicting the corrosion situation of large power grid grounding electrodes. Background Art

[0002] The existing corrosion monitoring technologies for large grounding electrodes of grounding grids mainly include two methods: regular inspection and electrochemical corrosion monitoring. Regular inspection is to manually check the appearance and grounding resistance of the grounding grid, and perform repairs when obvious corrosion problems are found. This method cannot monitor the corrosion status of the grounding grid in real time and is prone to missing hidden corrosion defects. Electrochemical corrosion monitoring uses a reference electrode to measure the potential distribution on the surface of the grounding grid to infer the corrosion degree, but this method can only reflect the corrosion situation in a local area and cannot comprehensively evaluate the corrosion status of the entire grounding grid. In addition, both of these methods require regular manual inspections and data collection, with a large workload, low efficiency, and inability to perform long-term continuous monitoring. To solve the above problems, in recent years, some researchers have proposed a corrosion monitoring technology for grounding grids based on electromagnetic signals. This method utilizes the transmission characteristics of electromagnetic signals in conductors and infers the corrosion status of the conductors by detecting the attenuation degree of the signals. For example, some scholars have proposed a corrosion monitoring system for grounding grids based on multi-channel electromagnetic signal acquisition, which can monitor the overall state of the grounding grid in real time and predict future corrosion trends through electromagnetic signal model analysis. However, this method still has some defects, such as being unable to accurately locate the corrosion points, difficult to distinguish the contribution degrees of different types of corrosion, and unable to quantitatively evaluate the corrosion degree. In addition, most of the existing electromagnetic signal analysis models are based on the idealized transmission line theory and are difficult to fully reflect the complex structure of the actual grounding grid and the non-uniform soil environment, thus affecting the accuracy of prediction. That is to say, the existing technologies have problems in comprehensively and quantitatively evaluating the corrosion status of large grounding electrodes of grounding grids and predicting future corrosion trends. Summary of the Invention

[0003] In view of this, the present invention provides a system for predicting the corrosion situation of large power grid grounding electrodes, which can solve the problems in the existing technologies that it is difficult to comprehensively and quantitatively evaluate the corrosion status of large grounding electrodes of grounding grids and predict future corrosion trends.

[0004] The present invention is implemented as follows:

[0005] The present invention provides a system for predicting the corrosion situation of large power grid grounding electrodes, including:

[0006] A carrier signal generating module, which is used to generate a carrier signal with a specific frequency;

[0007] A signal injection module, which includes a power amplifier and an impedance matching circuit, and is used to inject the carrier signal into the grounding network;

[0008] Multi-channel signal acquisition module, including at least 8 sampling channels, each channel is equipped with a 16-bit AD converter, and the sampling rate is not less than 100 kHz;

[0009] Signal conditioning module, including a preamplifier, an anti-aliasing filter and a band-pass filter;

[0010] Voltage / current measurement module, used to measure the voltage and current between each node of the grounding grid;

[0011] Soil parameter detection module, including a resistivity sensor, a pH value sensor and a water content sensor;

[0012] Temperature detection module, used to detect the soil temperature around the grounding grid;

[0013] GPS positioning module, used to determine the precise geographical location of each detection point;

[0014] Wireless communication module, supporting 4G / 5G networks, used for data transmission;

[0015] And a 32-bit control chip, the control chip is electrically connected to all the above modules, and is used for data acquisition, processing and corrosion prediction calculation. The corrosion prediction calculation is carried out by adopting a deep learning network to establish a corrosion prediction model according to the changes of spatio-temporal distribution; among them, the carrier signal generation module is connected in series with the signal injection module; the signal injection module is connected in parallel with the multi-channel signal acquisition module; the multi-channel signal acquisition module is connected to the control chip through the signal conditioning module; the voltage / current measurement module, the soil parameter detection module, the temperature detection module and the GPS positioning module are all connected in parallel with the control chip; the wireless communication module is connected in series with the control chip.

[0016] On the basis of the above technical solutions, a large-scale power grid grounding body corrosion situation prediction system of the present invention can also be improved as follows:

[0017] Among them, the number of sampling channels of the multi-channel signal acquisition module is 8-16, and each channel is equipped with an independent signal conditioning circuit, including:

[0018] Preamplifier, the adjustable range of gain is 20 dB - 60 dB;

[0019] Low-pass filter, the cut-off frequency can be adjusted in the range of 1 kHz - 50 kHz;

[0020] Band-pass filter, the center frequency is the same as the carrier frequency, and the bandwidth is ±5% of the carrier frequency;

[0021] 16-bit AD converter, the sampling rate can be adjusted in the range of 50 kHz - 200 kHz;

[0022] The multi-channel signal acquisition module further includes a multiplexer for channel switching and a clock circuit for signal synchronization.

[0023] Furthermore, the soil parameter detection module includes:

[0024] A sensor array for measuring soil resistivity by the four-electrode method, with a measurement range of 0.1 - 10000 Ω·m;

[0025] A glass electrode pH sensor, with a measurement range of 0 - 14 pH and an accuracy of ±0.1 pH;

[0026] A capacitive soil moisture sensor, with a measurement range of 0 - 100% and an accuracy of ±2%;

[0027] The soil parameter detection module further includes a signal acquisition circuit and a data processing circuit for real-time acquisition and processing of sensor data, and the sampling period is set between 1 minute and 24 hours.

[0028] Among them, a corrosion prediction software module is set in the control chip for performing the following steps:

[0029] S10. Obtain the grounding grid node position data collected by the GPS positioning module, establish a node connection relationship matrix according to the grounding grid node position data, generate a grounding grid topology structure diagram, and divide the grounding grid topology structure diagram into multiple detection units;

[0030] S20. Based on the carrier signal generated by the carrier signal generation module, within the sampling period of the multi-channel signal acquisition module, calculate the signal transmission integrity between each node in the detection unit to form a node transmission matrix, and calculate a real-time packet loss rate matrix according to the node transmission matrix; establish a network transmission loss evaluation model according to the real-time packet loss rate matrix to determine the signal attenuation degree of each detection unit of the grounding grid;

[0031] S30. Collect the node voltage data and current data measured by the voltage and current measurement module, obtain the soil data measured by the soil parameter detection module, and the soil temperature data measured by the temperature detection module, and combine the signal attenuation degree, input the node voltage data, the soil data, and the soil temperature data into a corrosion state evaluation function to generate an actual corrosion matrix; arrange the actual corrosion matrix in the order of sampling time to form an actual corrosion matrix sequence;

[0032] S40. Establish a theoretical corrosion calculation model according to the geometric dimension parameters and material characteristic parameters of the detection unit, input the soil data and the soil temperature data into the theoretical corrosion calculation model to generate a theoretical corrosion matrix; arrange the theoretical corrosion matrix in the order of sampling time to form a theoretical corrosion matrix sequence;

[0033] S50. Calculate the difference degree between the actual corrosion matrix sequence and the theoretical corrosion matrix sequence, extract the time variation characteristics, spatial distribution characteristics and amplitude characteristics of the difference degree, and combine the time variation characteristics, the spatial distribution characteristics and the amplitude characteristics to form a corrosion feature vector;

[0034] S60. Build a corrosion prediction model using a deep learning network, construct a training data set using the corrosion feature vector, input the training data set into the deep learning network, and train the deep learning network through the backpropagation algorithm to obtain an initial corrosion prediction model;

[0035] S70. Use the sliding time window method to obtain multiple groups of verification data, input the verification data into the initial corrosion prediction model for verification, calculate the prediction error, optimize the network parameters of the initial corrosion prediction model according to the prediction error, and obtain an optimized corrosion prediction model; Repeat this step multiple times, select the network parameters with the smallest prediction error as the final model parameters, and generate a final corrosion prediction model;

[0036] S80. Input the newly collected corrosion feature vector into the final corrosion prediction model to obtain the predicted results of the corrosion states of the grounding grid nodes, and classify the corrosion state prediction results according to a preset corrosion degree threshold to generate corrosion warning level information.

[0037] Further, the nodes of the grounding grid topology structure diagram represent the connection points of conductors and the positions of grounding electrodes in the grounding grid. The nodes include node numbers, geographic coordinate information, and node type identifiers, where the node type identifiers are used to distinguish ordinary connection points, test points and grounding electrodes; The edges of the grounding grid topology structure diagram represent the conductor connection paths in the grounding grid. The edges include the starting node number, the ending node number, the conductor type, the conductor length, and the buried depth information, where the conductor type is used to distinguish horizontally laid conductors and spiral vertical grounding bodies.

[0038] Further, the soil data includes soil resistivity data, pH data, and water content data.

[0039] Compared with the prior art, the beneficial effects of a large-scale power grid grounding body corrosion situation prediction system provided by the present invention are:

[0040] 1. It realizes the comprehensive monitoring of the corrosion status of the grounding grid. This system not only uses the method of electromagnetic signal detection to obtain the overall transmission characteristics of the grounding grid, but also combines means such as soil parameter detection, voltage and current measurement, etc. to comprehensively obtain various physical and chemical parameters affecting corrosion, laying a foundation for accurately evaluating the corrosion status.

[0041] 2. It can quantitatively analyze the corrosion degree of the grounding grid. The system has established a comparative analysis method of corrosion status based on measured data and theoretical models, which can not only obtain the actual corrosion depth, but also evaluate the deviation from the theoretical expectation, providing a basis for the quantification of the corrosion degree.

[0042] 3. It has accurate corrosion prediction ability. The system uses deep learning technology to construct a corrosion prediction model, which can predict the future corrosion trend of the grounding grid based on the corrosion characteristics monitored in real time, providing a basis for taking timely maintenance measures.

[0043] In summary, the present invention solves the problems in the prior art that it is difficult to comprehensively and quantitatively evaluate the corrosion status of large grounding grid conductors and predict the future corrosion trend. BRIEF DESCRIPTION OF THE DRAWINGS

[0044] Figure 1 It is a schematic diagram of the composition of the system provided by the present invention;

[0045] Figure 2 It is a flowchart of the steps executed by the corrosion prediction software module;

[0046] Figure 3 It is the corrosion prediction curve in the embodiment. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0047] To make the objectives, technical solutions and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the accompanying drawings in the embodiments of the present invention.

[0048] As Figure 1 shown, it is a schematic diagram of the composition of a system for predicting the corrosion situation of large power grid grounding conductors provided by the present invention, including:

[0049] A carrier signal generating module, used to generate a carrier signal with a specific frequency;

[0050] A signal injection module, including a power amplifier and an impedance matching circuit, used to inject the carrier signal into the grounding network;

[0051] A multi-channel signal acquisition module, including at least 8 sampling channels, each channel equipped with a 16-bit AD converter, and the sampling rate is not less than 100 kHz;

[0052] A signal conditioning module, including a preamplifier, an anti-aliasing filter and a band-pass filter;

[0053] A voltage / current measurement module, used to measure the voltage and current between each node of the grounding grid;

[0054] A soil parameter detection module, including a resistivity sensor, a pH value sensor and a water content sensor;

[0055] A temperature detection module, which is used to detect the soil temperature around the grounding grid;

[0056] A GPS positioning module, which is used to determine the precise geographical location of each detection point;

[0057] A wireless communication module, which supports 4G / 5G networks and is used for data transmission;

[0058] And a 32-bit control chip, which is electrically connected to all the above modules and is used for data acquisition, processing and corrosion prediction calculation. The corrosion prediction calculation is carried out by adopting a deep learning network to establish a corrosion prediction model according to the changes in spatio-temporal distribution; wherein, the carrier signal generation module is connected in series with the signal injection module; the signal injection module is connected in parallel with the multi-channel signal acquisition module; the multi-channel signal acquisition module is connected to the control chip through a signal conditioning module; the voltage / current measurement module, the soil parameter detection module, the temperature detection module and the GPS positioning module are all connected in parallel with the control chip; the wireless communication module is connected in series with the control chip.

[0059] As Figure 2 shown, a corrosion prediction software module is arranged in the control chip and is used to execute the following steps:

[0060] S10. Obtain the grounding grid node position data collected by the GPS positioning module, establish a node connection relationship matrix according to the grounding grid node position data, generate a grounding grid topology structure diagram, and divide the grounding grid topology structure diagram into multiple detection units;

[0061] S20. Based on the carrier signal generated by the carrier signal generation module, calculate the signal transmission integrity between each node in the detection unit within the sampling period of the multi-channel signal acquisition module, form a node transmission matrix, and calculate a real-time packet loss rate matrix according to the node transmission matrix; establish a network transmission loss evaluation model according to the real-time packet loss rate matrix, and determine the signal attenuation degree of each detection unit of the grounding grid;

[0062] S30. Collect the node voltage data and current data measured by the voltage and current measurement module, obtain the soil data measured by the soil parameter detection module, and the soil temperature data measured by the temperature detection module. Combine the signal attenuation degree, input the node voltage data, the soil data and the soil temperature data into a corrosion state evaluation function to generate an actual corrosion matrix; arrange the actual corrosion matrix in the order of sampling time to form an actual corrosion matrix sequence;

[0063] S40. Establish a theoretical corrosion calculation model based on the geometric dimension parameters and material property parameters of the detection unit, input the soil data and the soil temperature data into the theoretical corrosion calculation model to generate a theoretical corrosion matrix; arrange the theoretical corrosion matrix in the order of sampling time to form a theoretical corrosion matrix sequence;

[0064] S50. Calculate the difference degree between the actual corrosion matrix sequence and the theoretical corrosion matrix sequence, extract the time variation characteristics, spatial distribution characteristics and amplitude characteristics of the difference degree, and combine the time variation characteristics, the spatial distribution characteristics and the amplitude characteristics to form a corrosion feature vector;

[0065] S60. Construct a corrosion prediction model using a deep learning network, use the corrosion feature vector to construct a training data set, input the training data set into the deep learning network, and train the deep learning network through the backpropagation algorithm to obtain an initial corrosion prediction model;

[0066] S70. Adopt the sliding time window method to obtain multiple groups of verification data, input the verification data into the initial corrosion prediction model for verification, calculate the prediction error, optimize the network parameters of the initial corrosion prediction model according to the prediction error to obtain an optimized corrosion prediction model; repeat this step multiple times, select the network parameters with the minimum prediction error as the final model parameters, and generate a final corrosion prediction model;

[0067] S80. Input the newly collected corrosion feature vector into the final corrosion prediction model to obtain the predicted results of the corrosion states of the earthing grid nodes, and classify the corrosion state prediction results according to a preset corrosion degree threshold to generate corrosion warning level information.

[0068] The following describes the specific implementation manners of the above steps in detail:

[0069] The specific implementation manner of step S10 is as follows: First, obtain the position data of the earthing grid nodes from the GPS positioning module. Next, establish a node connection relationship matrix A based on the node position data, where Based on this matrix, generate the topological structure diagram of the earthing grid where represents the node set, and ε represents the edge set. On this basis, divide the topological structure diagram into multiple detection units Ω k , k = 1, 2,..., K. The nodes of the topological structure diagram represent the conductor connection points and the earthing electrode positions in the earthing grid. Each node contains a node number i and geographical coordinates (x i , y i) and the node type identifier t i ∈ {0, 1, 2}, where t i = 0 represents a normal connection point, t i = 1 represents a test point, t i = 2 represents a grounding electrode. And the edge (v i , v j ) ∈ ε represents the connection path of the conductors in the grounding grid. Each edge includes the starting node number i, the ending node number j, the conductor type m ij ∈ {0, 1} (0 represents a horizontally laid conductor, 1 represents a spiral vertical grounding electrode), the conductor length l ij and the burial depth h ij . By establishing such a detailed grounding grid topology model, it lays a foundation for the subsequent corrosion state assessment.

[0070] The specific implementation manner of step S20 is: First, based on the carrier signal U0e generated by the carrier signal generation module jωt , within the sampling period of the multi-channel signal acquisition module, calculate the signal transmission integrity T k between each node in the detection unit Ω ij , and form a node transmission matrix T. Each element T ij represents the voltage transmission coefficient between node i and node j, and satisfies the formula where Z L is the load impedance, and the transmission impedance matrix is expressed as:

[0071]

[0072] where each element satisfies the system of equations:

[0073]

[0074] In the formula, Z ij is the transmission impedance between node i and j; Z0 is the characteristic impedance; Z L is the load impedance; γ is the transmission constant; l ij is the conductor length; R is the resistance per unit length; L is the inductance per unit length; G is the conductance per unit length; C is the capacitance per unit length; ω is the angular frequency.

[0075] According to the node transmission matrix T:

[0076]

[0077] where each element satisfies:

[0078]

[0079] In the formula, T ijis the voltage transfer coefficient; V i , V j are the voltages at nodes i and j respectively; α is the attenuation constant; μ is the magnetic permeability of the medium; ε is the dielectric constant of the medium; σ is the conductivity of the medium; P ij is the power transfer ratio. The power transfer ratio P ij = |T ij | 2 , thus obtaining the real-time packet loss rate matrix. In this way, the overall signal transmission quality of the grounding grid can be evaluated, and a basis for subsequent corrosion state assessment can be provided.

[0080] The specific implementation of step S30 is as follows: First, collect the node voltage data V ij and current data I ij measured by the voltage-current measurement module, and obtain the soil resistivity data ρ ij , pH value data pH ij , water content data θ ij measured by the soil parameter detection module, and the soil temperature data T ij measured by the temperature detection module. Then, combining the signal attenuation degree obtained in step S20, input these data into the corrosion state evaluation function to generate the actual corrosion matrix C actual :

[0081]

[0082] where each element satisfies the partial differential equation system:

[0083]

[0084] In the formula, C ij is the corrosion depth at position (i, j); D is the diffusion coefficient; R ij is the corrosion reaction rate, calculated by the formula ; E ij is the electric field strength, calculated by the formula ; V ij is the potential; I ij is the current density; r is the distance; E a is the activation energy; R is the gas constant; T ij is the temperature; is the dissolved oxygen concentration; k1, k2, k3 are rate constants; n, m are reaction orders. The purpose of this step is to obtain the current actual corrosion condition of the grounding grid.

[0085] The specific implementation of step S40 is as follows: First, according to the geometric dimension parameter A i , V iA theoretical corrosion calculation model is established with the material property parameters ρ, n, and F. The mathematical description of this model is as follows: Where M ij represents the metal mass loss at position (i, j), and I ij is the corrosion current density, which is calculated by the formula , where I0 is the exchange current density, α is the transfer coefficient, and η ij = E ij - E eq,ij is the overpotential, is the equilibrium potential, E0 is the standard electrode potential, a ox,ij , a red,ij are the activities of the oxidized and reduced substances. Next, the soil resistivity data ρ ij , the pH data of the soil acidity and alkalinity, ij , the water content data θ ij , and the temperature data T ij obtained in step S30 are input into the theoretical corrosion calculation model to generate a theoretical corrosion matrix C theory :

[0086]

[0087] Each element of which satisfies:

[0088]

[0089] In the formula, M ij is the metal mass loss at position (i, j); I ij is the corrosion current density; n is the number of electron transfers; F is the Faraday constant; ρ is the metal density; I0 is the exchange current density; α is the transfer coefficient; η ij is the overpotential; E ij is the electrode potential; E eq,ij is the equilibrium potential; E0 is the standard electrode potential; a ox,ij , a red,ij are the activities of the oxidized and reduced substances.

[0090] Finally, the theoretical corrosion matrix C theory is arranged in the order of sampling time to form a theoretical corrosion matrix sequence. The purpose of this step is to establish a theoretical corrosion model of the grounding grid, providing a benchmark for subsequent comparison with the actual corrosion situation.

[0091] The specific implementation of step S50 is as follows: First, calculate the difference matrix ΔC between the actual corrosion matrix sequence C actual and the theoretical corrosion matrix sequence C theory :

[0092]

[0093] Each element satisfies:

[0094]

[0095] In the formula, σ ij is the conductivity at position (i, j); J corr,ij is the corrosion current density; J0 is the exchange current density; β is the Tafel slope; ΔC ij is the corrosion difference degree; x k is an influencing factor such as voltage V ij , current I ij , resistivity ρ ij , pH value ij , water content θ ij , and temperature T ij etc.; Γ ij is the boundary flux; n is the boundary normal vector.

[0096] Next, extract the time-varying characteristics spatial distribution characteristics and amplitude characteristics ||ΔC ij || of the difference degree matrix ΔC, and combine these three characteristics to form a corrosion feature vector The purpose of this step is to quantify the difference between the actual corrosion situation and the theoretical expectation, and provide a basis for the subsequent training of the corrosion prediction model.

[0097] The specific implementation of step S60 is as follows: First, construct a corrosion prediction model using a deep learning network. Specifically, use the corrosion feature vector Γ obtained in step S50 to construct a training dataset where C (n) represents the actual corrosion depth of the nth sample. Input the training dataset into the deep learning network, and optimize the network parameters θ through the backpropagation algorithm to minimize the loss function to obtain an initial corrosion prediction model It should be noted that the network structure and hyperparameters of the deep learning network need to be adjusted and optimized through a large amount of experimental data to improve the prediction accuracy of the model. The purpose of this step is to establish a grounding grid corrosion prediction model using machine learning methods and provide a basis for subsequent predictions.

[0098] The specific implementation of step S70 is as follows: First, use the sliding time window method to obtain multiple sets of validation data Then, input these validation data into the initial corrosion prediction model obtained in step S60 for verification and calculate the prediction error Optimize the network parameters θ of the initial model according to the estimated error ε (0) to obtain the optimized corrosion prediction model This process needs to be repeated many times until the network parameters θ with the minimum estimated error ε are selected * as the final model parameters to generate the final corrosion prediction model (*) The purpose of this step is to improve the accuracy and reliability of corrosion prediction by continuously optimizing the model parameters

[0099] The specific implementation of step S80 is as follows: First, input the newly collected corrosion feature vector Γ into the final corrosion prediction model generated in step S70 to obtain the predicted result of the corrosion state of the grounding grid node Then, according to the preset corrosion degree thresholds of 10%, 30%, and 50%, classify the predicted result C of the corrosion state pred to generate corrosion warning level information. Specifically, when C pred <10%, the warning level is low; when 10% ≤ C pred <30%, the warning level is medium; when C pred ≥30%, the warning level is high. The purpose of this step is to timely discover the potential risks of grounding grid corrosion and provide a decision-making basis for operation and maintenance personnel

[0100] It should be noted that:

[0101] 1. Transmission parameter acquisition:

[0102] Use an impedance analyzer to scan and measure the conductor impedance in the frequency range of 1 Hz - 1 MHz

[0103] Adopt the least squares method to fit and obtain the R, L, G, C parameters

[0104] Measure the environmental temperature, humidity, etc. to determine the medium parameters μ, ε, σ

[0105] 2. Corrosion parameter acquisition:

[0106] Use an electrochemical workstation to measure the polarization curve to obtain I0, α, β

[0107] Use a reference electrode to measure the potential distribution

[0108] Adopt the gravimetric method to regularly measure the mass loss of the metal specimen

[0109] Use a pH meter and a conductivity meter to measure the soil parameters

[0110] Specifically, the principle of the present invention is:

[0111] First, the system uses the GPS positioning module to obtain the spatial position information of the grounding grid nodes, and establishes a topological structure diagram of the grounding grid according to the connection relationship of the nodes. This topological structure diagram not only contains the geometric coordinates and type information of the nodes, but also records parameters such as the type, length, and burial depth of each conductor segment, laying the foundation for subsequent corrosion analysis.

[0112] Secondly, the system measures the electromagnetic signal transmission characteristics between the nodes of the grounding grid by using the method of carrier signal injection and multi-channel signal acquisition. By establishing an electromagnetic transmission line theoretical model, the signal transmission integrity and packet loss rate between each node can be calculated, reflecting the overall signal transmission quality of the grounding grid. This information provides an important basis for evaluating the corrosion condition.

[0113] At the same time, the system also collects parameters such as voltage and current data, soil resistivity, pH value, water content, and temperature of each node of the grounding grid. These parameters are closely related to the electrochemical reactions during the metal corrosion process. By establishing a corrosion kinetics model, the actual corrosion degree can be calculated.

[0114] By analyzing the difference between the actual corrosion condition and the theoretical expectation, the system extracts the time-varying characteristics, spatial distribution characteristics, and amplitude characteristics of the corrosion degree, and constructs a corrosion feature vector. Based on these feature vectors, the system uses deep learning methods to train a corrosion prediction model, which can accurately predict the future corrosion trend.

[0115] Finally, the system classifies the prediction results according to the preset corrosion degree threshold, generates corrosion warning information of different levels, and provides decision-making support for the maintenance and management of the grounding grid.

[0116] The following provides an embodiment of a specific application scenario of the present invention: A power enterprise is responsible for managing the grounding grid system of a large substation. The substation is located in an industrial park with a complex surrounding environment and diverse soil types. The substation covers an area of about 10 hectares, and the grounding grid consists of tens of kilometers of grounding conductors, connected with hundreds of grounding electrodes, forming a complex network structure.

[0117] In order to comprehensively evaluate the corrosion condition of the grounding grid, the power enterprise decides to use the large grounding grid corrosion condition prediction system proposed by the present invention for monitoring. First, the system uses the GPS positioning module to obtain the geographical coordinate information of each node of the grounding grid. By analyzing the connection relationship between the nodes, a grounding grid topological structure model as shown in Table - 21 is established.

[0118] Table 1 Node Table of the Grounding Grid Topological Structure Model

[0119] Node number Coordinates (longitude, latitude) Node type 1 (113.2501,22.5401) Grounding electrode 2 (113.2502,22.5402) Normal connection point 3 (113.2503,22.5403) Test point … … … 102 (113.2599,22.5499) Grounding electrode

[0120] Table 2 Edge Point Table of the Grounding Grid Topology Structure Model

[0121] Edge number Starting node Ending node Conductor type Length (m) Burial depth (m) 1 1 2 Horizontal 50 0.8 2 2 3 Horizontal 80 0.6 3 3 4 Vertical 30 1.2 … … … … … … 201 101 102 Horizontal 120 0.9

[0122] Based on this topology structure model, the system divides the entire grounding grid into 16 detection units, providing a basis for subsequent corrosion assessment and prediction.

[0123] Next, the system activates the carrier signal injection and multi-channel signal acquisition module. The carrier signal generation module generates a sine wave signal with a frequency of 50 kHz and an amplitude of 10 V. After power amplification and impedance matching, it is injected through the conductors in the grounding grid. The multi-channel signal acquisition module uses a 16-bit AD converter with a sampling rate of 100 kHz to synchronously acquire the voltage signals between nodes in all detection units.

[0124] By analyzing the acquired voltage signals, the system calculates parameters such as the transfer impedance, transfer coefficient, and power transfer ratio between nodes, forming a node transfer matrix as shown in Table 3.

[0125] Table 3 Node Transfer Matrix

[0126] Node number 1 2 3 … 102 1 1 0.98 0.92 … 0.85 2 0.98 1 0.96 … 0.88 3 0.92 0.96 1 … 0.91 … … … … … … 102 0.85 0.88 0.91 … 1

[0127] As can be seen from Table 3, the transfer coefficients between Node 1 and other nodes are generally low, indicating that there may be relatively serious corrosion problems in this area.

[0128] Meanwhile, the system also acquires the voltage and current data, soil resistivity data, pH data, water content data, and temperature data of each node in the grounding grid. The specific results are shown in Table 4.

[0129] Table 4 Measurement Results of Key Parameters of the Grounding Grid

[0130] Node number Voltage (V) Current (A) Resistivity (Ω·m) pH value Water content (%) Temperature (℃) 1 5.12 2.13 150 6.8 25 22 2 5.08 2.09 160 6.5 28 23 3 5.04 2.05 180 6.2 30 24 … … … … … … … 102 4.92 1.98 220 5.9 35 26

[0131] Based on these data and combined with the node transfer matrix obtained previously, the system calculates the actual corrosion matrix using the corrosion state assessment function, as shown in Table 5.

[0132] Table 5 Actual Corrosion Matrix

[0133]

[0134]

[0135] As can be seen from Table 4, the corrosion degree near Node 1 is relatively serious, reaching 32%, while the corrosion near Node 102 is relatively light, only 24%. This is consistent with the analysis results of the previous node transfer matrix.

[0136] To determine the differences between these actual corrosion conditions and the theoretical expectations, the system also established the following theoretical corrosion calculation model:

[0137]

[0138] Among them, M ij is the metal mass loss at position (i, j), and I ij is the corrosion current density, which is calculated by the formula η ij = E ij - E eq,ij is the overpotential, is the equilibrium potential. Based on the soil parameter data in Table 3, the system calculated the theoretical corrosion matrix shown in Table 6.

[0139] Table 6 Theoretical Corrosion Matrix

[0140] Node number 1 2 3 … 102 1 0 0.08 0.18 … 0.28 2 0.08 0 0.12 … 0.23 3 0.18 0.12 0 … 0.20 … … … … … … 102 0.28 0.23 0.20 … 0

[0141] By comparing the actual corrosion matrix (Table 4) and the theoretical corrosion matrix (Table 5), the system calculated the difference matrix between the two, as shown in Table 7.

[0142] Table 7 Difference Matrix

[0143] Node number 1 2 3 … 102 1 0 0.04 0.04 … 0.04 2 0.04 0 0.04 … 0.04 3 0.04 0.04 0 … 0.04 … … … … … … 102 0.04 0.04 0.04 … 0

[0144] As can be seen from Table 6, there are significant differences between the actual corrosion situation near Node 1 and the theoretical expectations, reaching 4 percentage points. The differences in other nodes are relatively small. This indicates that there may be some special corrosion factors in the area where Node 1 is located and further analysis is required.

[0145] Based on the above analysis results, the system extracted the time-varying characteristics, spatial distribution characteristics, and amplitude characteristics of corrosion, forming the following corrosion feature vectors:

[0146]

[0147] By inputting these feature vectors into a pre-trained deep learning network, the system obtained the corrosion prediction results for each node of the grounding grid in the next 12 months, as Figure 3 shown. It can be seen that the corrosion degree near Node 1 will continue to increase in the next 12 months, reaching and exceeding the warning threshold of 50%. The corrosion prediction results for other nodes are relatively stable and still within the acceptable range.

[0148] As described above, it is only the specific implementation manner of the present invention, but the protection scope of the present invention is not limited thereto. Any person skilled in the art within the technical scope disclosed by the present invention can easily think of changes or substitutions, which should all be covered by the protection scope of the present invention.

Claims

1. A large-scale power grid grounding electrode corrosion situation prediction system, characterized in that, Including: A carrier signal generation module for generating a carrier signal of a specific frequency; A signal injection module including a power amplifier and an impedance matching circuit for injecting the carrier signal into the grounding network; A multi-channel signal acquisition module including at least 8 sampling channels, each channel equipped with a 16-bit AD converter and a sampling rate of not less than 100 kHz; A signal conditioning module including a preamplifier, an anti-aliasing filter and a band-pass filter; A voltage / current measurement module for measuring the voltage and current between the nodes of the grounding network; A soil parameter detection module including a resistivity sensor, a pH value sensor and a water content sensor; A temperature detection module for detecting the soil temperature around the grounding network; A GPS positioning module for determining the precise geographical location of each detection point; A wireless communication module supporting 4G / 5G networks for data transmission; And a 32-bit control chip, which is electrically connected to all the above modules for data acquisition, processing and corrosion prediction calculation. The corrosion prediction calculation is performed by using a deep learning network to establish a corrosion prediction model based on the spatio-temporal distribution change; wherein, the carrier signal generation module is connected in series with the signal injection module; the signal injection module is connected in parallel with the multi-channel signal acquisition module; the multi-channel signal acquisition module is connected to the control chip through the signal conditioning module; the voltage / current measurement module, the soil parameter detection module, the temperature detection module and the GPS positioning module are all connected in parallel with the control chip; the wireless communication module is connected in series with the control chip.

2. The large-scale power grid grounding body corrosion condition prediction system according to claim 1, characterized in that, The number of sampling channels of the multi-channel signal acquisition module is 8 - 16, and each channel is equipped with an independent signal conditioning circuit, including: A preamplifier with an adjustable gain range of 20 dB - 60 dB; A low-pass filter with a cut-off frequency adjustable in the range of 1 kHz - 50 kHz; A band-pass filter with a center frequency the same as the carrier frequency and a bandwidth of ±5% of the carrier frequency; A 16-bit AD converter with a sampling rate adjustable in the range of 50 kHz - 200 kHz.

3. A large-scale power grid grounding body corrosion situation prediction system according to claim 2, characterized in that, The soil parameter detection module includes: A sensor array for measuring soil resistivity by the four-electrode method, with a measurement range of 0.1 - 10000 Ω·m; A glass electrode pH sensor with a measurement range of 0 - 14 pH and an accuracy of ±0.1 pH; A capacitive soil water content sensor with a measurement range of 0 - 100% and an accuracy of ±2%.

4. A large-scale power grid grounding body corrosion condition prediction system according to claim 1, characterized in that, A corrosion prediction software module is set in the control chip. The corrosion prediction calculation is performed by using a deep learning network to establish a corrosion prediction model based on the spatio-temporal distribution change for performing the following steps: S10. Obtain the grounding network node position data collected by the GPS positioning module, establish a node connection relationship matrix according to the grounding network node position data, generate a grounding network topology structure diagram, and divide the grounding network topology structure diagram into multiple detection units; S20. Based on the carrier signal generated by the carrier signal generation module, within the sampling period of the multi-channel signal acquisition module, calculate the signal transmission integrity between each node in the detection unit to form a node transmission matrix, and calculate the real-time packet loss rate matrix according to the node transmission matrix; Establish a network transmission loss evaluation model according to the real-time packet loss rate matrix to determine the signal attenuation degree of each detection unit of the grounding grid; S30. Collect the node voltage data and current data measured by the voltage and current measurement module, obtain the soil data measured by the soil parameter detection module, and the soil temperature data measured by the temperature detection module. Combine the signal attenuation degree, and input the node voltage data, the soil data, and the soil temperature data into the corrosion state evaluation function to generate an actual corrosion matrix; Arrange the actual corrosion matrix in the order of sampling time to form an actual corrosion matrix sequence; S40. Establish a theoretical corrosion calculation model according to the geometric dimension parameters and material characteristic parameters of the detection unit, input the soil data and the soil temperature data into the theoretical corrosion calculation model to generate a theoretical corrosion matrix; arrange the theoretical corrosion matrix in the order of sampling time to form a theoretical corrosion matrix sequence; S50. Calculate the difference degree between the actual corrosion matrix sequence and the theoretical corrosion matrix sequence, extract the time variation characteristics, spatial distribution characteristics, and amplitude characteristics of the difference degree, and combine the time variation characteristics, the spatial distribution characteristics, and the amplitude characteristics to form a corrosion feature vector; S60. Use a deep learning network to construct a corrosion prediction model, use the corrosion feature vector to construct a training data set, input the training data set into the deep learning network, and train the deep learning network through the backpropagation algorithm to obtain an initial corrosion prediction model; S70. Use the sliding time window method to obtain multiple groups of verification data, input the verification data into the initial corrosion prediction model for verification, calculate the prediction error, optimize the network parameters of the initial corrosion prediction model according to the prediction error, and obtain an optimized corrosion prediction model; repeat this step multiple times, select the network parameters with the smallest prediction error as the final model parameters, and generate a final corrosion prediction model; S80. Input the newly collected corrosion feature vector into the final corrosion prediction model to obtain the corrosion state prediction result of the grounding grid node, and classify the corrosion state prediction result according to the preset corrosion degree threshold to generate corrosion warning level information.

5. The large-scale power grid grounding electrode corrosion situation prediction system according to claim 4, characterized in that, The nodes of the grounding grid topology structure diagram represent the conductor connection points and the positions of the grounding electrodes in the grounding grid. The nodes include node numbers, geographical coordinate information, and node type identifiers, where the node type identifiers are used to distinguish ordinary connection points, test points, and grounding electrodes; the edges of the grounding grid topology structure diagram represent the conductor connection paths in the grounding grid. The edges include the starting node number, the ending node number, the conductor type, the conductor length, and the burial depth information, where the conductor type is used to distinguish horizontally laid conductors and spiral vertical grounding electrodes.

6. The large-scale power grid grounding electrode corrosion situation prediction system according to claim 4, characterized in that, The soil data includes soil resistivity data, pH data, and water content data.

7. The large-scale power grid grounding body corrosion situation prediction system according to claim 2, characterized in that, The multi-channel signal acquisition module further includes a multiplexer for channel switching and a clock circuit for signal synchronization.

8. A large power grid grounding electrode corrosion situation prediction system according to claim 3, characterized in that The soil parameter detection module further includes a signal acquisition circuit and a data processing circuit for real-time acquisition and processing of sensor data, and the sampling period is set between 1 minute and 24 hours.

9. A large power grid grounding body corrosion condition prediction system according to claim 4, characterized in that, The sampling period is set to 1 minute.

10. A large power grid grounding body corrosion situation prediction system according to claim 4, characterized in that, The preset corrosion degree thresholds include 10%, 30%, and 50%, corresponding to low, medium, and high-level corrosion warnings respectively.

Images