Method for evaluating state of large grounding grid of power grid

Abstract

The invention provides a power grid large-scale grounding grid state evaluation method, which belongs to the technical field of grounding grid detection, and comprises the following steps of: extracting a resistance component and an inductance component through frequency domain impedance analysis to construct a sparse impedance characteristic matrix, and establishing an aggregation grid distribution matrix by adopting a self-adaptive grid strategy; potential distribution and current density distribution are calculated through a regional decomposition parallel algorithm to obtain a conversion equivalent circuit parameter vector, a solving strategy is dynamically adjusted through nuclear space and image space decomposition, and the conversion equivalent circuit parameter vector and a standard grounding grid parameter vector are compared to calculate an abnormal state evaluation index vector; a corrosion area is marked to generate a corrosion position distribution matrix, a grounding grid state evaluation report is output in combination with an adaptive measurement strategy and an automatic correction mechanism, and the technical problems that the detection cost is high, the workload is large and comprehensive detection cannot be achieved due to the fact that excavation verification is needed for large grounding grid corrosion state detection are solved.

Description

TECHNICAL FIELD

[0001] The present application belongs to the technical field of grounding grid detection, and in particular relates to a method for evaluating the state of a large grounding grid of a power grid. BACKGROUND

[0002] The grounding grid of a power system is an important facility for ensuring the safe operation of equipment and personal safety, and its state detection traditionally adopts the power frequency grounding resistance measurement method and the excavation verification method. The power frequency grounding resistance measurement method can only obtain the overall impedance value of the grounding grid, and cannot locate the specific corrosion position, while the excavation verification method can directly observe the conductor corrosion condition, but requires large-area excavation, which destroys the site and has a long working cycle. In the prior art, due to the large number of nodes and complex spatial distribution of a large grounding grid, the traditional excavation method needs to consume a large amount of manpower and material resources for comprehensive corrosion detection, and the excavation process may damage the structure of the grounding grid, while relying only on power frequency measurement cannot identify local corrosion and poor contact of the welding points and other hidden faults. That is to say, in the prior art, there is a technical problem that the corrosion state detection of a large grounding grid requires excavation verification, resulting in high detection cost, large workload and inability to achieve comprehensive detection. SUMMARY

[0003] Therefore, the present application provides a method for evaluating the state of a large grounding grid of a power grid, which can solve the technical problem in the prior art that the corrosion state detection of a large grounding grid requires excavation verification, resulting in high detection cost, large workload and inability to achieve comprehensive detection.

[0004] The present application is implemented in the following manner: The present application provides a method for evaluating the state of a large grounding grid of a power grid, which injects a multi-frequency high-frequency test current at each node of the grounding grid and collects voltage response signals to construct an original impedance response data set; performs frequency domain analysis on the original impedance response data set to extract resistance components and inductance components and calculate a sparse impedance feature matrix; divides the grounding grid into key current dispersion areas and non-key current dispersion areas and establishes an aggregated grid distribution matrix using different grid densities; calculates the potential distribution and current density distribution using a region decomposition parallel algorithm according to the sparse impedance feature matrix and the aggregated grid distribution matrix to obtain a transformed equivalent circuit parameter vector; decomposes the transformed equivalent circuit parameter vector into kernel space and image space, and dynamically adjusts the region decomposition boundary conditions according to the kernel space dimension ratio parameter; compares the transformed equivalent circuit parameter vector with a standard grounding grid parameter vector to calculate an abnormal state evaluation index vector and generate a corrosion position distribution matrix; adjusts the sampling frequency according to the repeatability error index of the welding point contact resistance measurement; executes a current injection zero point automatic correction program according to the zero point offset; generates a grounding grid state evaluation report and outputs the corrosion position distribution matrix and the grounding impedance change trend curve.

[0005] The multi-frequency high-frequency test current is a sinusoidal current signal with a frequency range of 1 kHz to 100 kHz, and the amplitude is set to 10 A to 50 A.

[0006] Wherein, the establishment of the sparse impedance feature matrix adopts Fourier transform to calculate complex impedance, separates the real part to obtain resistance component, and separates the imaginary part and divides by the angular frequency to obtain inductance component.

[0007] Wherein, the key current dispersion area is an area in which the current density gradient is greater than the current density gradient threshold in the grounding grid, and the non-key current dispersion area is an area in which the current density gradient is less than the current density gradient threshold.

[0008] Wherein, the establishment of the aggregated mesh distribution matrix adopts fine mesh density for the key current dispersion area and coarse mesh density for the non-key current dispersion area.

[0009] Wherein, the region decomposition parallel algorithm adopts a geometric decomposition strategy to divide the spatial domain into non-overlapping sub-regions, and realizes information transmission between sub-regions through interface boundary conditions, and adopts Schwarz alternating method for iterative solution.

[0010] Wherein, the establishment of the transformed equivalent circuit parameter vector adopts finite element method to establish linear equations, and adopts multilevel fast multipole algorithm to accelerate matrix vector multiplication operation.

[0011] Wherein, the kernel space and image space decomposition adopts singular value decomposition to calculate the kernel space and the image space, and the kernel space dimension reflects the uncertainty degree of the linear equations.

[0012] Wherein, the kernel space dimension ratio parameter is the ratio of the current kernel space dimension increment to the initial kernel space dimension, and the region decomposition boundary condition is adjusted when the kernel space dimension ratio parameter is greater than 0.15.

[0013] Wherein, the establishment of the abnormal state evaluation index vector is by calculating the resistance deviation index and the inductance deviation index, and defining the corrosion index component as the arithmetic mean of the resistance deviation index and the inductance deviation index.

[0014] Further, when the corrosion index component is greater than the corrosion determination threshold, the corresponding conductor segment is marked as a corrosion area, and the corrosion determination threshold is 0.3. Wherein, the corrosion position distribution matrix arranges the corrosion index component according to the grounding grid plane coordinate position to form a two-dimensional matrix, which is used for visual display of the corrosion distribution of the grounding grid. Wherein, the welding point contact resistance measurement repeatability error index is represented by relative standard deviation, and when the welding point contact resistance measurement repeatability error index is greater than the first error allowed value, the sampling number is increased to the enhanced sampling number.

[0015] The first error allowable value is 8%, the enhanced sampling number is 12 times, and the standard measurement sampling number is 5 times. The zero point offset is the non-zero current value displayed by the measurement loop of the current injection device in the state of no current output. When the zero point offset is greater than the second error allowable value, the current injection zero point automatic correction program is executed. The second error allowable value is 5%, and the current injection zero point automatic correction program adjusts the bias voltage of the current source step by step to make the absolute value of the average zero point current less than 0.01A.

[0016] The present application uses the physical characteristics of increased inductance component after conductor corrosion to perform trenchless diagnosis by combining multi-frequency high-frequency current injection method with frequency domain impedance analysis, avoids the destructive detection of traditional excavation verification, and realizes comprehensive evaluation of the corrosion state of the grounding net. The present application uses adaptive grid strategy and regional decomposition parallel algorithm, performs fine calculation on the key current dispersion area, and uses coarse grid on the non-key area, which not only ensures the calculation accuracy but also greatly reduces the calculation amount, solving the problem of low efficiency of large-scale grounding net fine numerical calculation. In summary, the present application realizes quantitative evaluation of the grounding net state and accurate positioning of the corrosion position by establishing a sparse impedance characteristic matrix and transforming an equivalent circuit parameter vector, combining the dynamic analysis mechanism of kernel space and image space decomposition, solving the technical problems of high detection cost, large workload and inability to realize comprehensive detection caused by the need for excavation verification in the background technology. BRIEF DESCRIPTION OF DRAWINGS

[0017] Figure 1 The flowchart of the method of the present application.

[0018] Figure 2 The corrosion index component distribution curve along the conductor number in the embodiment.

[0019] Figure 3 The grounding impedance curve with test frequency in the embodiment. DETAILED DESCRIPTION

[0020] In order to make the purpose, technical scheme and advantages of the embodiments of the present application clearer, the technical scheme in the embodiments of the present application will be described clearly and completely below with reference to the drawings in the embodiments of the present application.

[0021] As Figure 1 shown is the flowchart of the grounding net state evaluation method provided by the present application, and the method comprises the following steps:

[0022] S01, inject multi-frequency high-frequency test current at each node of the grounding net, collect voltage response signals of each accessible contact, and construct an original impedance response data set;

[0023] S02. Perform frequency domain analysis on the original impedance response dataset, extract the resistance and inductance components of each measurement point, and calculate the sparse impedance characteristic matrix.

[0024] S03. Divide the grounding grid into critical current dissipation areas and non-critical current dissipation areas. Apply a fine grid density to the critical current dissipation areas and a coarse grid density to the non-critical current dissipation areas to establish an aggregated grid distribution matrix.

[0025] S04. Based on the sparse impedance characteristic matrix and the aggregated grid distribution matrix, the potential distribution and current density distribution of each node are calculated using a region decomposition parallel algorithm to obtain the equivalent circuit parameter vector.

[0026] S05. Perform kernel space and image space decomposition on the converted equivalent circuit parameter vector. When the kernel space dimension ratio parameter belongs to the first dimension ratio interval, adjust the region decomposition boundary conditions and recalculate the potential distribution. When the kernel space dimension ratio parameter belongs to the second dimension ratio interval, maintain the current decomposition strategy and continue calculation.

[0027] S06. Compare and analyze the equivalent circuit parameter vector with the standard grounding grid parameter vector, calculate the abnormal state evaluation index vector, and when the corrosion index component in the abnormal state evaluation index vector is greater than the corrosion judgment threshold, mark the corresponding conductor segment as a corrosion area and generate a corrosion location distribution matrix.

[0028] S07. If the repeatability error index of the welding point contact resistance measurement is greater than the first allowable error value, the number of multi-frequency test samplings is increased to the enhanced sampling number to improve the reliability of the results. When the repeatability error index of the welding point contact resistance measurement is less than or equal to the first allowable error value, the standard measurement sampling number is used to save resources.

[0029] S08. When the zero-point offset is greater than the second allowable error value, immediately execute the current injection zero-point automatic correction program to restore the reference accuracy. If the zero-point offset is less than or equal to the second allowable error value, maintain the current reference calibration state and continue to operate.

[0030] S09. Generate a grounding grid status assessment report based on the abnormal state assessment index vector and the corrosion location distribution matrix, and output the corrosion location distribution matrix and the grounding impedance change trend curve.

[0031] The multi-frequency high-frequency test current is a sinusoidal current signal with a frequency range of 1kHz to 100kHz and an amplitude set to 10A to 50A. It is injected into the grounding grid node through a current injection device. The frequency range is selected based on the fact that the inductive effect of the grounding conductor is significant and the capacitance effect is negligible within this frequency range. The amplitude setting range is selected based on the fact that it can ensure the strength of the measurement signal without interfering with the normal operation of the grounding grid.

[0032] The original impedance response dataset is a collection of data containing voltage response signals and corresponding frequency information for each measurement point. Each element in the dataset contains four fields: measurement point number, frequency value, voltage amplitude, and phase angle. The dataset dimension is the number of measurement points multiplied by the number of frequency sampling points.

[0033] The sparse impedance characteristic matrix is ​​a matrix describing the impedance relationship between nodes of the grounding grid, established through frequency domain impedance analysis. The establishment process is as follows: for each measuring point... voltage response signal and the multi-frequency high-frequency test current Perform a Fourier transform to the frequency domain and calculate the complex impedance. The complex impedance The calculation formula is as follows: The complex impedance is equal to the frequency domain representation of the voltage response signal divided by the frequency domain representation of the multi-frequency high-frequency test current, and the resistance component is obtained by separating the real part. Separate the imaginary part and divide by the angular frequency. Obtain the inductance component The angular frequency Equal to twice pi multiplied by frequency For a circular conductor, the inductance value is calculated using the following formula: the inductance component is equal to the permeability of free space multiplied by the conductor length, divided by twice pi, and then multiplied by the natural logarithm of the ratio of the conductor length to the conductor radius. Values When the conductor becomes thinner due to corrosion, the conductor radius decreases, leading to an increase in the inductance component. (Matrix elements) Represents a node With nodes The normalized impedance values ​​are obtained by dividing the measured impedance value by the standard impedance value. The standard impedance value The value is 100Ω. The standard impedance value is based on the median value of the typical impedance range of a large grounding grid. The normalized impedance value is calculated as follows: The normalized impedance value is equal to the complex sum of the products of the resistance component and the angular frequency and inductance component divided by the standard impedance value. Since the number of nodes in a large grounding grid is huge but the number of directly electrically connected node pairs is limited, most elements in the sparse impedance characteristic matrix are zero, exhibiting sparse characteristics.

[0034] The critical current-dissipating region is the area in the grounding grid where the current density gradient is greater than a current density gradient threshold. Values The current density gradient threshold is determined by the 75th percentile of the current density gradient distribution through simulation analysis of the current distribution of the grounding grid in a typical substation. The key current dissipation area includes the area around the grounding down conductor connection point and the area around the grounding conductor cross welding point.

[0035] The non-critical current dissipation region is the region in the grounding grid where the current density gradient is less than the current density gradient threshold. The non-critical current dissipation region includes the edge conductor region far from the injection point and the region around the deeply buried grounding body.

[0036] The mesh size corresponding to the fine mesh density is 0.1m, and the mesh size corresponding to the coarse mesh density is 0.5m. The mesh size corresponding to the fine mesh density is selected to ensure that the calculation accuracy error of the potential gradient in the key area is less than 5%, and the mesh size corresponding to the coarse mesh density is selected to maximize the mesh size to reduce the amount of computation while ensuring the convergence of the calculation.

[0037] The aggregated grid distribution matrix is ​​a matrix describing the spatial discretization distribution of the grounding grid, established using an adaptive grid strategy. The establishment process is as follows: First, an initial uniform grid is divided based on the grounding grid topology and the estimated current density. Regions with current density gradients greater than the current density gradient threshold are marked as critical current-dissipating regions, and regions with current density gradients less than the current density gradient threshold are marked as non-critical current-dissipating regions. The refined grid density is applied to the critical current-dissipating regions, and the coarser grid density is applied to the non-critical current-dissipating regions. Matrix elements... Indicates spatial location The mesh density level is defined as follows: a mesh density level of 1 indicates a fine mesh, and a mesh density level of 0 indicates a coarse mesh. The aggregation process is achieved by merging adjacent meshes of the same type into mesh blocks. Each mesh block employs a uniform numerical solution precision. The dimension of the aggregated mesh distribution matrix is ​​[dimension missing]. ,in and They are respectively direction and Number of grid blocks in the direction.

[0038] The described domain decomposition parallel algorithm is a numerical method that decomposes a large-scale computational problem into multiple independent sub-problems for parallel solution. For the grounding grid problem, a geometric decomposition strategy is employed, dividing the spatial domain into non-overlapping sub-regions. Each sub-region is solved on an independent computational unit. Information is transferred between sub-regions through interface boundary conditions. The Schwarz alternation method is used for iterative solution. Subregion during the next iteration The interface boundary conditions are determined by the adjacent sub-regions. The results of this iteration are provided when the solutions for all said subregions converge to a relative error of less than [a certain value]. Stop iterating when the time comes.

[0039] The potential distribution is a distribution function formed by arranging the potential values ​​of each node of the grounding grid according to their spatial positions. It is obtained by solving a system of linear equations within the sub-region. The form of the system of linear equations is that the stiffness matrix multiplied by the node potential vector equals the node current vector.

[0040] The current density distribution is a distribution function formed by arranging the current density values ​​at various spatial locations of the grounding grid according to spatial coordinates, and is calculated using Ohm's law based on the potential distribution.

[0041] The equivalent circuit parameter vector is the set of parameters used to transform the distributed grounding grid physical model into a lumped parameter circuit model. The process involves: decomposing the grounding grid into regions based on the aggregated grid distribution matrix, dividing the overall network into... Each of the aforementioned sub-regions is independently computed using the region decomposition parallel algorithm, and the sub-regions are coupled through the interface boundary conditions. The nodes within the system of linear equations are established using the finite element method, and the matrix-vector multiplication operation is accelerated using a multilevel fast multipole algorithm, reducing the computational load from... Reduce to The potentials of each node are obtained by solving. and current density Then, for each conductor segment Extracting equivalent resistance and equivalent inductance The equivalent resistance The calculation formula is as follows: the equivalent resistance is equal to the voltage drop across the conductor divided by the current flowing through it, and the equivalent inductance is... The calculation formula is as follows: the equivalent inductance is equal to the change in magnetic flux divided by the current flowing through it. The equivalent resistance and equivalent inductance of all conductor segments are arranged in order to form the converted equivalent circuit parameter vector, and the dimension of the converted equivalent circuit parameter vector is... , This represents the total number of conductor segments.

[0042] The described multilevel fast multipole algorithm is a numerical technique for accelerating matrix-vector multiplication operations. It divides the spatial region into multiple layers of grids in a tree structure. Long-range interactions are approximated through multipole expansion and local expansion, while short-range interactions are calculated directly and accurately. This reduces the computational burden. Reduce to ,in This represents the total number of nodes.

[0043] The kernel space and image space decomposition are mathematical methods for linearly decomposing the parameter vectors of the transformed equivalent circuit. This is used to dynamically adjust the numerical solution strategy. For the linear equation system, the kernel space is defined as the space consisting of all solution vectors that satisfy the condition that the product of the stiffness matrix and the node potential vector equals zero. The image space is defined as the space consisting of linear combinations of all column vectors of the stiffness matrix. The kernel space dimension reflects the degree of uncertainty of the linear equation system. When electrical islands or connection failures exist in the grounding grid, the kernel space dimension increases. The image space dimension reflects the number of effective constraints in the linear equation system. Singular value decomposition is used to calculate the kernel space and the image space. Singular values ​​less than the singular value threshold are considered valid. The corresponding singular vectors constitute the kernel space basis vectors, and the corresponding singular vectors with singular values ​​greater than the singular value threshold constitute the image space basis vectors. Values The singular value threshold is determined based on the typical lower limit of machine precision in numerical computation.

[0044] The initial kernel space dimension is denoted as , is the kernel space dimension obtained when the kernel space and image space decomposition is first executed, and the changes in the kernel space dimension are monitored in real time during the calculation process.

[0045] The kernel space dimension ratio parameter is the ratio of the current kernel space dimension increment to the initial kernel space dimension, and the kernel space dimension increment is the difference between the current kernel space dimension and the initial kernel space dimension.

[0046] The first dimension ratio interval is the range of values ​​where the kernel space dimension ratio parameter is greater than 0.15. The second dimension ratio interval is the range of values ​​where the kernel space dimension ratio parameter is greater than or equal to 0 and less than or equal to 0.15. The lower bound of the first dimension ratio interval, 0.15, is determined by statistical analysis of measured data from the grounding grid. When the kernel space dimension ratio parameter exceeds 0.15, the deviation of the calculation result exceeds 10%, and the solution strategy needs to be adjusted.

[0047] The region decomposition boundary conditions are the interface boundary conditions used for information transmission between the sub-regions, including potential continuity conditions and current continuity conditions. Adjusting the region decomposition boundary conditions is achieved by re-dividing the sub-regions or modifying the coupling method of the interface boundary conditions.

[0048] The standard grounding grid parameter vector is a set of equivalent circuit parameters under the design or healthy state of the grounding grid, derived from the grounding grid design parameters or historical measurement data. The dimension of the standard grounding grid parameter vector is the same as the dimension of the transformed equivalent circuit parameter vector.

[0049] The abnormal state assessment index vector is a comprehensive set of indicators used to quantify the degree to which the grounding grid deviates from the normal state. The establishment process is as follows: the equivalent circuit parameter vector is compared element by element with the standard grounding grid parameter vector to calculate the resistance deviation index. and inductance deviation index The resistance deviation index The calculation formula is as follows: the resistance deviation index is equal to the difference between the equivalent resistance and the standard equivalent resistance divided by the standard equivalent resistance; the inductance deviation index... The calculation formula is expressed as follows: the inductance deviation index is equal to the difference between the equivalent inductance and the standard equivalent inductance divided by the standard equivalent inductance, and the corrosion index component is defined. The contact resistance deviation of the weld point is calculated as the arithmetic mean of the resistance deviation index and the inductance deviation index. The contact resistance deviation of the welding point The calculation formula is as follows: The deviation of the weld point contact resistance is equal to the difference between the measured weld point contact resistance and the standard weld point contact resistance divided by the standard weld point contact resistance. When the deviation values ​​are ∈ (0.5, +∞), the welding point is considered to have poor contact. The deviation indices of each conductor segment and welding point are arranged according to their spatial position to form the abnormal state evaluation index vector. The dimension of the abnormal state evaluation index vector is... , This represents the total number of welding points.

[0050] The corrosion index component is an element in the abnormal state assessment index vector used to characterize the degree of conductor corrosion; the larger the value, the more severe the corrosion.

[0051] The corrosion judgment threshold is set to 0.3. The value of the corrosion judgment threshold is based on the statistical analysis of historical corrosion data of grounding grids of multiple substations. When the corrosion index component exceeds 0.3, the conductor cross-sectional area loss exceeds 20% and the grounding resistance increases by more than 30%, which has a significant impact on the safety performance of the grounding grid.

[0052] The corrosion location distribution matrix is ​​a two-dimensional matrix formed by arranging the corrosion index components according to the plane coordinates of the grounding grid. The matrix element value is the corrosion index component of the conductor segment at the corresponding position. When the corrosion index component is greater than the corrosion judgment threshold, the corresponding matrix element is marked as a corrosion state. The corrosion location distribution matrix is ​​used to visually display the corrosion distribution of the grounding grid.

[0053] The first allowable error value is the upper limit of the repeatability error index of the welding point contact resistance measurement, which is 8%, and is an empirical value.

[0054] The enhanced sampling number is 12 times, and the standard measurement sampling number is 5 times. The values ​​of the enhanced sampling number and the standard measurement sampling number are based on statistical principles. When the number of measurements reaches 12 times, the confidence level of the relative standard deviation can reach 95%. The standard measurement sampling number of 5 times can meet the measurement accuracy requirements under normal circumstances.

[0055] The repeatability error index for measuring the contact resistance of the welding point is the degree of dispersion of the resistance values ​​obtained from multiple measurements of the same welding point, characterized by the relative standard deviation. The calculation method involves performing measurements on the same welding point... The sequence of weld point contact resistance values ​​was obtained from the measurements, and the average weld point contact resistance value was calculated. and standard deviation The average contact resistance value of the weld point The calculation formula is as follows: The average weld point contact resistance value is equal to the sum of the contact resistance values ​​of each weld point in the weld point contact resistance value sequence divided by the number of measurements. The standard deviation The calculation formula is expressed as follows: The standard deviation is equal to the square root of the sum of the squares of the differences between the contact resistance values ​​of each welding point in the sequence of welding point contact resistance values ​​and the average welding point contact resistance value, divided by the number of measurements minus 1. The calculation formula for the repeatability error index of welding point contact resistance measurement is expressed as follows: The repeatability error index of welding point contact resistance measurement is equal to the standard deviation divided by the average welding point contact resistance value multiplied by 100%.

[0056] The second allowable error value is the upper limit of the zero-point offset, which is 5%. The second allowable error value is determined based on the allowable error requirements for zero-point drift in current measurement in the Verification Procedure for Measuring Instruments of the Power Industry DL / T596.

[0057] The zero-point offset is the non-zero current value displayed by the measurement circuit of the current injection device in the state of no current output, reflecting the reference accuracy of the measurement system. It is calculated by adjusting the measurement circuit with the current source off. The zero-point current reading sequence was obtained from the sampling, and the average zero-point current was calculated. The average zero-point current The calculation formula is as follows: The average zero-point current is equal to the sum of the zero-point current readings in the zero-point current reading sequence divided by the number of samplings. The zero-point offset is calculated using the following formula: the zero-point offset equals the absolute value of the average zero-point current divided by the rated injection current value. Multiply by 100%, the rated injection current value The value is 30A, and the rated injection current value is based on the median value of the typical operating current range of the high-frequency current injection test of the grounding grid.

[0058] The current injection zero-point automatic correction program is an automated program that adjusts the current source output to bring the measurement circuit reading to zero. The execution process involves gradually adjusting the current source bias voltage until the absolute value of the average zero-point current is less than 0.01A.

[0059] The reference calibration state is the operating state in which the measurement system maintains the zero-point offset within the allowable error range.

[0060] The grounding grid condition assessment report is a comprehensive report document containing the grounding grid detection and analysis results. The content includes a table of grounding impedance measurement results, the corrosion location distribution matrix, the distribution curve of the corrosion index components along the conductor length, the grounding impedance change trend curve, and maintenance suggestion text. The grounding grid condition assessment report is output and archived in electronic document form.

[0061] The grounding impedance change trend curve is a function curve of the grounding impedance value changing with time. The horizontal axis is the time axis and the vertical axis is the grounding impedance value. The curve data comes from the grounding impedance records of multiple historical measurements and the grounding impedance value measured in this measurement.

[0062] This invention achieves trenchless detection of grounding grid corrosion status through a high-frequency current injection method. It utilizes the physical characteristic of increased inductance after conductor corrosion for diagnosis, significantly reducing detection costs and workload compared to traditional excavation verification methods. The invention employs adaptive meshing and a parallel algorithm for region decomposition to address the problem of low efficiency in refined numerical calculations for large-scale grounding grids. Rapid non-destructive testing of weld contact resistance is achieved through multi-frequency impedance spectroscopy analysis. The established sparse impedance characteristic matrix, aggregated mesh distribution matrix, transformed equivalent circuit parameter vector, and abnormal state evaluation index vector provide a systematic data foundation for the quantitative assessment of grounding grid status. The dynamic analysis mechanism of kernel space and image space decomposition ensures the robustness and accuracy of numerical solutions. The adaptive measurement strategy and automatic correction mechanism improve the reliability of detection results, providing a scientific basis for the status assessment and maintenance decisions of power system grounding grids.

[0063] The present invention also provides a computer-based method for forming a large-scale power grid grounding network status assessment system, wherein the computer is equipped with a readable storage medium that stores program instructions, which execute the above-described method when the computer is run.

[0064] The specific implementation methods of the above steps are described in detail below.

[0065] The specific implementation of step S01 is as follows: First, a current injection device is configured and connected to the main node of the grounding grid. A sinusoidal current signal with multiple discrete frequency points between 1kHz and 100kHz is generated by a signal generator. The current amplitude at each frequency point is set between 10A and 50A to ensure the measurement signal-to-noise ratio. Test currents of different frequencies are injected into the grounding grid sequentially using a frequency step scanning method. At the same time as the current injection, the voltage response of each accessible point of the grounding grid is synchronously acquired using a high-speed data acquisition card. The sampling frequency is set to more than 20 times the test signal frequency to meet the requirements of the Nyquist sampling theorem. The acquired voltage amplitude, phase angle information, and corresponding frequency values ​​and measurement point numbers are recorded together to form the original impedance response dataset. The purpose of this step is to obtain the electrical response characteristics of the grounding grid under different frequency excitations, providing basic data for subsequent frequency domain analysis. The reason for choosing a high-frequency test current is that the inductive effect of the grounding conductor is significant while the capacitance effect is negligible in this frequency range, which can effectively reflect the change in impedance characteristics caused by changes in conductor geometry.

[0066] The specific implementation of step S02 is as follows: Fast Fourier Transform is performed on the voltage response signal and injected current signal of each measurement point in the original impedance response dataset to convert the time-domain signal to the frequency domain. The ratio of the voltage frequency domain representation to the current frequency domain representation is calculated to obtain the complex impedance value of each measurement point at different frequencies. The real part of the complex impedance is extracted as the resistance component, and the imaginary part of the complex impedance is divided by the angular frequency to obtain the inductance component. A model relating the inductance value to the conductor radius is established based on the inductance calculation theory of a circular cross-section conductor. When the conductor corrodes, causing a decrease in cross-sectional area, the inductance value will increase accordingly. This physical characteristic is used to identify corroded conductors. The impedance relationship between any two nodes in the grounding grid is normalized and then filled into the corresponding position of the sparse impedance characteristic matrix. The normalization process uses a standard impedance value of 100Ω as a reference to eliminate the influence of dimensions. Since there are many nodes in a large grounding grid but only a limited number of directly connected node pairs, the generated matrix exhibits obvious sparsity characteristics. The purpose of this step is to convert the measured time-domain response signal into frequency-domain impedance parameters and establish an impedance correlation matrix between nodes, providing a quantitative parameter basis for subsequent conductor corrosion diagnosis and equivalent circuit modeling.

[0067] The specific implementation of step S03 is as follows: based on the grounding grid topology and current distribution prediction results, a preliminary grid division is performed on the entire grounding grid area. Then, the current density gradient value at each location is calculated and compared with the current density gradient threshold of 100. By comparing the gradient values, regions with gradient values ​​greater than a threshold are marked as critical current dissipation regions, while regions with gradient values ​​less than a threshold are marked as non-critical current dissipation regions. Critical current dissipation regions mainly include current-converging locations such as around grounding down conductor connection points and conductor cross-welding points. Non-critical current dissipation regions mainly include locations with gentle current density changes, such as edge conductors far from the injection point and around deeply buried grounding bodies. A fine mesh with a grid size of 0.1m is used for critical current dissipation regions to ensure that the potential gradient calculation accuracy error is less than 5%, while a coarse mesh with a grid size of 0.5m is used for non-critical current dissipation regions to reduce the number of computational nodes and improve solution efficiency. Adjacent similar meshes are aggregated to form mesh blocks, and the mesh density level of each location is marked in the aggregated mesh distribution matrix. The purpose of this step is to establish an adaptive spatial discretization scheme to balance computational accuracy and computational efficiency. By densifying the mesh in complex current distribution regions and sparsening the mesh in gentle current distribution regions, the optimal allocation of limited computational resources is achieved.

[0068] The specific implementation of step S04 is as follows: The entire grounding grid computational domain is divided into several non-overlapping sub-regions based on the aggregated grid distribution matrix. Each sub-region is solved using a region decomposition parallel algorithm on an independent computational unit. Potential and current information are transmitted between sub-regions through interface boundary conditions to achieve coupled computation. For each node within a sub-region, a system of linear equations is established using the finite element method, consisting of a stiffness matrix and node current vectors. A multi-level fast multipole algorithm is introduced to accelerate matrix-vector multiplication, reducing the computational complexity from the square of the number of nodes to the logarithm of the number of nodes. The Schwarz alternation method is used for iterative solving until the solutions for all sub-regions converge to a relative error less than [value missing]. The equivalent resistance is calculated by extracting the voltage drop and current flowing through each conductor segment from the obtained node potential distribution and current density distribution. The equivalent inductance is calculated by the ratio of the change in magnetic flux to the current. The equivalent resistance and equivalent inductance of all conductor segments are arranged in order to form a transformed equivalent circuit parameter vector. The purpose of this step is to achieve efficient numerical solution of large-scale grounding grids through parallel computing technology and fast algorithms, and to transform the distributed physical model into a lumped parameter circuit model that is easy to analyze.

[0069] The specific implementation of step S05 is to perform singular value decomposition on the stiffness matrix of the linear equation system to obtain a sequence of singular values ​​and the corresponding singular vectors, and to select singular values ​​less than a threshold. The singular vectors are used to form a set of basis vectors for the kernel space to calculate the kernel space dimension. Singular vectors with singular values ​​greater than a threshold are used to form a set of basis vectors for the image space to calculate the image space dimension. The initial kernel space dimension obtained during the first decomposition is recorded as the baseline value. During the calculation, the changes in the kernel space dimension are monitored in real time, and the difference between the current kernel space dimension and the initial kernel space dimension is calculated to obtain the kernel space dimension increment. The kernel space dimension increment is divided by the initial kernel space dimension to obtain the kernel space dimension ratio parameter. When the ratio parameter is greater than 0.15, it indicates that there is a large uncertainty in the linear equation system, which may be due to electrical islanding or connection failure. At this time, the sub-region partitioning method or the coupling strategy of the interface boundary conditions is modified, and the potential distribution is recalculated. When the ratio parameter is in the range of 0 to 0.15, it indicates that the solution strategy is reasonable and the current calculation process can continue. The purpose of this step is to dynamically monitor the stability of the numerical solution through the linear space decomposition theory and adaptively adjust the solution strategy according to the changes in the kernel space dimension to ensure the robustness and accuracy of the calculation results.

[0070] The specific implementation of step S06 is as follows: The equivalent circuit parameter vector is compared element-by-element with the standard grounding grid parameter vector to calculate the resistance deviation index and inductance deviation index of each conductor segment. The resistance deviation index is obtained by dividing the difference between the equivalent resistance and the standard equivalent resistance by the standard equivalent resistance. The inductance deviation index is obtained by dividing the difference between the equivalent inductance and the standard equivalent inductance by the standard equivalent inductance. The arithmetic mean of the resistance deviation index and the inductance deviation index is defined as the corrosion index component. Simultaneously, the contact resistance deviation of the welding point is calculated, and the contact state is determined. The deviation indices of each conductor segment and welding point are arranged according to their spatial position. An abnormal state assessment index vector is formed. All corrosion index components in the abnormal state assessment index vector are traversed and compared with the corrosion judgment threshold of 0.3. When the corrosion index component of a conductor segment is greater than 0.3, it indicates that the cross-sectional area loss of the conductor exceeds 20% and the grounding resistance increases by more than 30%. The conductor segment is marked as a corrosion area and the corrosion index component value is recorded at the corresponding position in the corrosion location distribution matrix. The purpose of this step is to identify corroded conductors and poorly contacted welding points in the grounding grid through the deviation analysis of equivalent circuit parameters, establish a quantitative abnormal state assessment system, and generate an intuitive corrosion location distribution map.

[0071] The specific implementation of step S07 is as follows: multiple contact resistance measurements are performed on the same welding point to obtain a sequence of measurement values. The average value and standard deviation of the sequence are calculated to obtain the relative standard deviation as the repeatability error index of the welding point contact resistance measurement. This error index is compared with the first allowable error value of 8%. When the error index is greater than 8%, it indicates that the measurement result has large dispersion and insufficient reliability. At this time, the number of measurement samplings is increased to 12 to improve the statistical confidence. When the error index is less than or equal to 8%, it indicates that the current measurement result meets the accuracy requirements. The measurement task can be completed by using the standard number of measurement samplings of 5, saving test time and resource consumption. The purpose of this step is to adaptively adjust the sampling strategy according to the repeatability error of the measurement result and optimize the test efficiency while ensuring the reliability of the measurement.

[0072] The specific implementation of step S08 is as follows: with the current source off, the measurement circuit is sampled multiple times to obtain a sequence of zero-point current readings. The average value of this sequence is calculated to obtain the average zero-point current. The absolute value of this average zero-point current is divided by the rated injection current value of 30A and multiplied by 100% to obtain the zero-point offset. The zero-point offset is compared with the second allowable error value of 5%. When the zero-point offset is greater than 5%, it indicates that the reference of the measurement system has drifted significantly, affecting the measurement accuracy. The current injection zero-point automatic correction program is immediately started. By gradually adjusting the current source bias voltage, the absolute value of the average zero-point current is reduced to below 0.01A to restore the reference accuracy. When the zero-point offset is less than or equal to 5%, it indicates that the measurement system is in a good reference calibration state and can continue to operate normally. The purpose of this step is to monitor and correct the zero-point drift of the measurement system to ensure the reference accuracy of current injection and voltage measurement and avoid the impact of system errors on the grounding grid status assessment results.

[0073] The specific implementation of step S09 is to integrate the abnormal state assessment index vector and corrosion location distribution matrix obtained in the previous steps to generate a grounding grid status assessment report. The report includes a tabular summary of the grounding impedance measurement results, a two-dimensional heatmap visualization of the corrosion location distribution matrix, a distribution curve of corrosion index components along the conductor length, a grounding impedance change trend curve, and maintenance recommendation text based on the abnormal state assessment results. The grounding impedance change trend curve shows the evolution law of grounding grid performance by plotting the functional relationship between historical multiple measurement records and the grounding impedance value of the current measurement over time. The generated assessment report is output in electronic document form and archived. The purpose of this step is to systematically organize and visualize the various results of grounding grid detection and analysis, providing comprehensive and accurate technical basis for the maintenance decisions of power system operation and maintenance personnel.

[0074] It should be noted that the key technical ideas of this invention include a multi-frequency high-frequency current injection and inductance effect diagnosis mechanism, an adaptive mesh and domain decomposition parallel algorithm, a dynamic solution strategy for kernel space and image space decomposition, and an adaptive measurement and automatic correction mechanism. The multi-frequency high-frequency current injection and inductance effect diagnosis mechanism achieves trenchless detection of grounding grid corrosion by injecting test current in the high-frequency range and analyzing changes in inductance components. It utilizes the physical law that conductor corrosion leads to a decrease in cross-sectional area and an increase in inductance to establish a corrosion diagnosis model. Compared with traditional excavation verification methods, this significantly reduces detection costs and workload while avoiding destructive impacts on the grounding grid structure. The adaptive mesh and domain decomposition parallel algorithm dynamically adjusts the mesh density according to the current density gradient and uses a domain decomposition strategy to decompose a large-scale computational problem into multiple sub-problems for parallel solution. It densifies the mesh in key current-dissipating regions to ensure computational accuracy while sparsening the mesh in non-critical regions to reduce computational load. Combined with a multi-layer fast multipole algorithm to accelerate matrix operations, it effectively solves the problem of low efficiency in refined numerical calculations of large-scale grounding grids. The dynamic solution strategy for kernel space and image space decomposition adaptively adjusts the solution strategy by monitoring changes in the kernel space dimension of the linear equation system in real time. When the kernel space dimension increases abnormally, the boundary conditions of the domain decomposition are adjusted in a timely manner to avoid numerical instability, ensuring reliable calculation results even when there are local connection failures or complex topologies in the grounding grid. The adaptive measurement and automatic correction mechanism dynamically adjusts the number of samplings based on measurement repeatability errors and automatically triggers the correction program by monitoring zero-point offset. This optimizes test efficiency and eliminates the impact of systematic errors on the evaluation results while ensuring the reliability of measurement results. The synergistic effect of the above key technical ideas is to build a complete technical system from data acquisition, numerical calculation, condition diagnosis to result output. The combination of physical diagnosis mechanism and efficient numerical algorithm enables accurate identification of grounding grid corrosion status. The synergy between dynamic solution strategy and adaptive measurement mechanism ensures the robustness and accuracy of evaluation results. Compared with traditional methods, it achieves significant improvements in detection efficiency, calculation accuracy, and result reliability, providing a systematic technical solution for the condition assessment and maintenance decision-making of power system grounding grids.

[0075] It should be noted that this invention also solves the following technical problem: In the numerical calculation of large-scale grounding grids, due to the large number of nodes and complex spatial distribution, the traditional finite element method requires establishing a global stiffness matrix and solving a large-scale linear equation system. The computational load increases quadratically with the number of nodes, leading to excessively long computation time or even the inability to complete the calculation. This invention decomposes the large-scale computational problem into multiple independent sub-problems by employing a domain decomposition parallel algorithm. Each sub-region is solved in parallel on an independent computational unit. Information is transferred between sub-regions through interface boundary conditions. Combined with the multilayer fast multipole algorithm to accelerate matrix-vector multiplication operations, the computational load is reduced from O(N²) to O(NlogN). Furthermore, a kernel space and image space decomposition mechanism is introduced to dynamically monitor the solution process. When the kernel space dimension ratio parameter exceeds 0.15, the domain decomposition boundary conditions are automatically adjusted to ensure the convergence and accuracy of the numerical solution, thereby achieving efficient and refined calculation of large-scale grounding grids.

[0076] Furthermore, this invention also solves the technical problems of poor repeatability in grounding grid weld point contact resistance measurement and the impact of measurement system zero-point drift on the reliability of test results. This invention establishes an adaptive measurement strategy to monitor the repeatability error index of weld point contact resistance measurement in real time. When this index exceeds the first allowable error value of 8%, the number of multi-frequency test samplings is automatically increased to 12 to improve statistical confidence. When the index is less than or equal to 8%, the standard measurement sampling number of 5 times is used to save resources. Simultaneously, this invention monitors the zero-point offset in real time. When the zero-point offset exceeds the second allowable error value of 5%, an automatic current injection zero-point correction procedure is immediately executed, gradually adjusting the current source bias voltage until the absolute value of the average zero-point current is less than 0.01A, ensuring the accuracy of the measurement system's reference, thereby improving the reliability and consistency of grounding grid condition assessment results.

[0077] Specifically, the principle of this invention is as follows: Based on the electromagnetic principle that the cross-sectional area of ​​a conductor decreases after corrosion, leading to an increase in the inductive component, this invention injects a multi-frequency test current from 1kHz to 100kHz into the grounding grid. Within this frequency range, the inductive effect is significant while the capacitance effect is negligible, enabling accurate extraction of the resistance and inductive components. This invention performs a Fourier transform on the acquired voltage response signal to calculate the complex impedance in the frequency domain, separating the real and imaginary parts to obtain the resistance and inductive components respectively. The corrosion index component is calculated by comparing it with standard parameters, and when the corrosion index component exceeds the corrosion judgment threshold, it is marked as a corrosion area. This invention employs an adaptive mesh strategy to divide the grounding grid into critical and non-critical current-dissipating regions. A fine mesh is used for critical regions with large current density gradients, while a coarse mesh is used for non-critical regions with small current density gradients. Combining a parallel domain decomposition algorithm and a multi-layer fast multipole algorithm, the computational load is reduced from O(N²) to O(NlogN), significantly improving computational efficiency while maintaining accuracy. This makes comprehensive and refined calculation of large grounding grids possible, thereby achieving accurate assessment and location of the grounding grid corrosion state under trenchless conditions.

[0078] The following provides a specific embodiment 1 of the present invention, and the specific implementation of each step in this embodiment 1 is described in detail below.

[0079] The specific implementation of step S01 is as follows: First, a current injection device is configured and connected to the main node of the grounding grid. Then, a signal generator is used to generate a frequency range of 1 to 100 Hz. A sinusoidal current signal with multiple discrete frequency points between 10 and 50, where the current amplitude at each frequency point is set between 10 and 50. To ensure the signal-to-noise ratio of the measurement, a frequency step scanning method is used to sequentially inject test currents of different frequencies into the grounding grid. Simultaneously, a high-speed data acquisition card is used to synchronously acquire the voltage response of each accessible contact of the grounding grid. The sampling frequency is set to more than 20 times the test signal frequency to meet the requirements of the Nyquist sampling theorem. The acquired voltage amplitude, phase angle information, and corresponding frequency values ​​and measurement point numbers are recorded together to form the original impedance response dataset. The purpose of this step is to obtain the electrical response characteristics of the grounding grid under different frequency excitations, providing basic data for subsequent frequency domain analysis. The reason for choosing a high-frequency test current is that the inductive effect of the grounding conductor is significant while the capacitance effect is negligible in this frequency range, which can effectively reflect the change in impedance characteristics caused by changes in conductor geometry.

[0080] The specific implementation of step S02 is as follows: Fast Fourier Transform is performed on the voltage response signal and injected current signal at each measurement point in the original impedance response dataset to convert the time-domain signal to the frequency domain. The complex impedance value at each measurement point at different frequencies is obtained by calculating the ratio of the voltage frequency domain representation to the current frequency domain representation. The calculation formula is expressed as follows:

[0081] ;

[0082] In the formula, For measuring points In frequency The complex impedance is given in units of . Among them, measuring points The numbering range is 1 to , This represents the total number of measurable points accessible to the grounding grid. For measuring points The voltage response signal is represented in the frequency domain, with units of . ; This represents the frequency domain of the multi-frequency high-frequency test current, in units of... ; Test frequency, unit: .

[0083] Extract the real part of the complex impedance into the resistance component. Dividing the imaginary part of the complex impedance by the angular frequency yields the inductance component. angular frequency The calculation formula is expressed as follows:

[0084] ;

[0085] In the formula, Angular frequency, unit: ; Pi, with a value of 3.14159.

[0086] For a conductor with a circular cross-section, the inductive component The calculation formula is expressed as follows:

[0087] ;

[0088] In the formula, For measuring points The inductance component of the corresponding conductor, in units of ; Let be the vacuum permeability, with a value of . ; For measuring points The length of the corresponding conductor, in units of Obtained through grounding grid design drawings; For measuring points The radius of the corresponding conductor, in units of , obtained through actual measurement or design parameters.

[0089] Normalized impedance value The calculation formula is expressed as follows:

[0090] ;

[0091] In the formula, For nodes With nodes The normalized impedance values ​​between, dimensionless, where the nodes The numbering range is 1 to ; For measuring points The resistance component, in units of ; The imaginary unit satisfies ; For measuring points The inductance component, in units of ; This is the standard impedance value, taken as 100. .

[0092] The purpose of this step is to convert the measured time-domain response signal into frequency-domain impedance parameters and establish an impedance correlation matrix between nodes, providing a quantitative parameter basis for subsequent conductor corrosion diagnosis and equivalent circuit modeling.

[0093] The specific implementation of step S03 is as follows: based on the grounding grid topology and current distribution prediction results, a preliminary grid division is performed on the entire grounding grid area. Then, the current density gradient value at each location is calculated and compared with the current density gradient threshold of 100. By comparing the gradient values, regions with gradient values ​​greater than a threshold are marked as critical current dissipation regions, while regions with gradient values ​​less than a threshold are marked as non-critical current dissipation regions. Critical current dissipation regions mainly include current-concentrating locations such as around grounding down conductor connection points and conductor cross-welding points. Non-critical current dissipation regions mainly include locations with gradual current density changes, such as edge conductors far from the injection point and around deeply buried grounding bodies. A grid size of 0.1 is used for critical current dissipation regions. A finer mesh was used to ensure that the accuracy error of the potential gradient calculation was less than 5%, and a mesh size of 0.5 was used for non-critical current-diffusing regions. The coarsening of the mesh reduces the number of computational nodes and improves solution efficiency. Adjacent meshes of the same type are aggregated to form mesh blocks, and the mesh density level at each location is marked in the aggregated mesh distribution matrix. The elements of the aggregated mesh distribution matrix... The expression is as follows:

[0094] ;

[0095] In the formula, For spatial location The grid density level at that location is dimensionless. For position Current density gradient at point, in units of ; This is the current density gradient threshold, with a value of 100. ; for Directional grid block index, with values ​​ranging from 1 to ; for Directional grid block index, with values ​​ranging from 1 to ; for Number of grid blocks in the direction; for Number of grid blocks in the direction.

[0096] The purpose of this step is to establish an adaptive spatial discretization scheme that balances computational accuracy and efficiency. This is achieved by refining the mesh in regions with complex current distribution and sparsening the mesh in regions with smooth current distribution, thus optimizing the allocation of limited computational resources.

[0097] The specific implementation of step S04 is to divide the entire grounding grid calculation domain into sections based on the aggregated grid distribution matrix. Non-overlapping sub-regions, among which The number of sub-regions is empirically set to 4 to 16. Each sub-region is solved using a parallel domain decomposition algorithm on an independent computational unit. Potential and current information are transferred between sub-regions through interface boundary conditions to achieve coupled computation. For the nodes within each sub-region, a system of linear equations is established using the finite element method, consisting of a stiffness matrix and nodal current vectors. The form of the linear equations is as follows:

[0098] ;

[0099] In the formula, Here is the stiffness matrix, with dimension 1. ; Calculate the total number of nodes in the grounding grid; The node potential vector, in units of , dimension ; The node current vector, in units of , dimension .

[0100] Introducing a multilevel fast multipole algorithm to accelerate matrix-vector multiplication operations reduces computational complexity from... Reduce to The Schwarz alternation method is used for iterative solution until the solutions for all subregions converge to a relative error of less than 1. The equivalent resistance is calculated by extracting the voltage drop across each conductor segment and the current flowing through it from the obtained node potential distribution and current density distribution. Current density distribution With potential distribution The relationship is expressed as follows:

[0101] ;

[0102] In the formula, For spatial location Current density at the point, in units of ; The conductivity of a conductor, in units of It is determined by the properties of the conductor material; For spatial location Potential gradient at point, in units of ; This represents the coordinates of any position in the grounding grid space.

[0103] The formula for calculating the equivalent resistance is as follows:

[0104] ;

[0105] In the formula, For conductor segment The equivalent resistance, in units of The conductor segment The numbering range is 1 to ; For conductor segment The pressure drop, in units of ; For the conductor segment The current, in units of .

[0106] Equivalent Inductance The calculation formula is expressed as follows:

[0107] ;

[0108] In the formula, For conductor segment The equivalent inductance, in units of ; For conductor segment The change in magnetic flux linkage, in units of It is obtained by numerical integration of the magnetic field around the conductor segment.

[0109] The equivalent resistance and equivalent inductance of all conductor segments are arranged in order to form a transformed equivalent circuit parameter vector. The dimension of the transformed equivalent circuit parameter vector is... ,in The total number of conductor segments is the target of this step. The purpose of this step is to achieve efficient numerical solutions for large-scale grounding grids through parallel computing techniques and fast algorithms, and to transform the distributed physical model into a lumped parameter circuit model that is easy to analyze.

[0110] The specific implementation of step S05 is to perform singular value decomposition on the stiffness matrix of the linear equation system to obtain a sequence of singular values ​​and the corresponding singular vectors, and to select singular values ​​less than a threshold. The singular vectors are used to construct the kernel space basis vector set, and the kernel space dimension is calculated. Singular vectors with singular values ​​greater than a threshold are used to construct the image space basis vector set, and the image space dimension is calculated. The initial kernel space dimension obtained during the first decomposition is recorded. As a baseline, the kernel space dimension is monitored in real time during the calculation process, and the difference between the current kernel space dimension and the initial kernel space dimension is calculated to obtain the kernel space dimension increment, and the kernel space dimension ratio parameter. The calculation formula is expressed as follows:

[0111] ;

[0112] In the formula, The ratio parameter of the kernel space dimension is dimensionless. The current kernel space dimension is dimensionless; Let be the initial kernel space dimension, which is dimensionless.

[0113] When the ratio parameter is greater than 0.15, it indicates that there is a large uncertainty in the linear equation system, which may be due to electrical islanding or connection failure. In this case, the potential distribution should be recalculated after adjusting the sub-region division method or modifying the coupling strategy of the interface boundary conditions. When the ratio parameter is in the range of 0 to 0.15, it indicates that the solution strategy is reasonable and the current calculation process can continue. The purpose of this step is to dynamically monitor the stability of the numerical solution through the linear space decomposition theory and adaptively adjust the solution strategy according to the change of the kernel space dimension to ensure the robustness and accuracy of the calculation results.

[0114] The specific implementation of step S06 is to compare the equivalent circuit parameter vector with the standard grounding grid parameter vector element by element to calculate the resistance deviation index and inductance deviation index of each conductor segment. The calculation formula is expressed as follows:

[0115] ;

[0116] In the formula, For conductor segment The resistance deviation index is dimensionless; For conductor segment The equivalent resistance, in units of ; For conductor segment Standard equivalent resistance, unit: This information is derived from grounding grid design parameters or historical measurement data.

[0117] Inductance deviation index The calculation formula is expressed as follows:

[0118] ;

[0119] In the formula, For conductor segment The inductance deviation index is dimensionless. For conductor segment The equivalent inductance, in units of ; For conductor segment Standard equivalent inductance, unit: This information is derived from grounding grid design parameters or historical measurement data.

[0120] Corrosion index components The calculation formula is expressed as follows:

[0121] ;

[0122] In the formula, For conductor segment The corrosion index component is dimensionless.

[0123] Welding point contact resistance deviation The calculation formula is expressed as follows:

[0124] ;

[0125] In the formula, The contact resistance deviation at the welding point is dimensionless. To measure the contact resistance of the weld point, the unit is... ; The standard solder joint contact resistance is expressed in units of 1000 ppm. .

[0126] The deviation indices of each conductor segment and weld point are arranged according to their spatial location to form an abnormal state assessment index vector. The dimension of the abnormal state assessment index vector is... ,in The total number of conductor segments. Given the total number of weld points, iterate through all corrosion index components in the abnormal state assessment index vector and compare them with the corrosion judgment threshold of 0.3. When the corrosion index component of a conductor segment is greater than 0.3, it indicates that the cross-sectional area loss of the conductor exceeds 20% and the grounding resistance increases by more than 30%. Mark the conductor segment as a corrosion area and record the corrosion index component value at the corresponding position in the corrosion location distribution matrix. The elements of the corrosion location distribution matrix... The expression is as follows:

[0127] ;

[0128] In the formula, Planar coordinate position The elements of the corrosion location distribution matrix are dimensionless. This represents the corrosion index component of the conductor segment corresponding to that location; For the grounding grid plane Coordinates, in units ; For the grounding grid plane Coordinates, in units .

[0129] The purpose of this step is to identify corroded conductors and poorly connected solder joints in the grounding grid through deviation analysis of equivalent circuit parameters, establish a quantitative abnormal state assessment system, and generate an intuitive corrosion location distribution map.

[0130] The specific implementation of step S07 is to perform multiple contact resistance measurements on the same welding point to obtain a sequence of measurement values, and then average the contact resistance value of the welding point. The calculation formula is expressed as follows:

[0131] ;

[0132] In the formula, The average contact resistance value at the weld point, in units of ; The number of measurements is 5 or 12. For the first The measured contact resistance value of the weld point, in units of... .

[0133] Standard deviation The calculation formula is expressed as follows:

[0134] ;

[0135] In the formula, The standard deviation is expressed in units of 1000 ppm. .

[0136] Repeatability error index of weld point contact resistance measurement The calculation formula is expressed as follows:

[0137] ;

[0138] In the formula, This is a repeatability error index for measuring the contact resistance of welding points, expressed as a percentage.

[0139] The error index is compared with the first allowable error value of 8%. When the error index is greater than 8%, it indicates that the measurement result has large dispersion and insufficient reliability. At this time, the number of measurement samplings is increased to 12 to improve the statistical confidence. When the error index is less than or equal to 8%, it indicates that the current measurement result meets the accuracy requirements. The measurement task can be completed by using the standard number of measurement samplings of 5, saving test time and resource consumption. The purpose of this step is to adaptively adjust the sampling strategy according to the repeatability error of the measurement result, and optimize the test efficiency while ensuring the reliability of the measurement.

[0140] The specific implementation of step S08 is to obtain a zero-point current reading sequence by performing multiple zero-point current samplings on the measurement circuit while the current source is off, and then averaging the zero-point current. The calculation formula is expressed as follows:

[0141] ;

[0142] In the formula, The average zero-point current is expressed in units of 1. ; This refers to the number of samples, with an empirical value of 10 to 20. For the first The zero-point current reading of the sampled unit is... .

[0143] Zero offset The calculation formula is expressed as follows:

[0144] ;

[0145] In the formula, This is the zero-point offset, expressed as a percentage. The rated injection current value is 30. .

[0146] The zero-point offset is compared with the second allowable error value of 5%. If the zero-point offset is greater than 5%, it indicates that the measurement system reference has drifted significantly, affecting measurement accuracy. The automatic zero-point calibration program using current injection is immediately initiated, gradually adjusting the current source bias voltage to reduce the absolute value of the average zero-point current to 0.01. The following steps restore the reference accuracy. When the zero-point offset is less than or equal to 5%, it indicates that the measurement system is in a good reference calibration state and can continue to operate normally. The purpose of this step is to monitor and correct the zero-point drift of the measurement system to ensure the reference accuracy of current injection and voltage measurement, and to avoid the impact of system errors on the grounding grid status assessment results.

[0147] The specific implementation method of step S09 is the same as described above, and will not be repeated in detail here.

[0148] It needs to be explained that the formula for the kernel space dimension ratio parameter... Based on the theory that the kernel space dimension in linear algebra reflects the uncertainty of a system of linear equations, the stability of the numerical solution is judged by monitoring the increment ratio of the kernel space dimension relative to the initial value. When there are electrical islands or connection failures in the grounding grid, the rank of the stiffness matrix will decrease, leading to an increase in the kernel space dimension. This formula realizes the adaptive identification of abnormal solution states and triggers the solution strategy adjustment mechanism, ensuring the robustness and accuracy of large-scale grounding grid calculations.

[0149] Corrosion index component formula Taking into account the physical characteristics that conductor corrosion causes both increased resistance and increased inductance, the resistance deviation index is used... and inductance deviation index A more robust corrosion assessment index is obtained by performing an arithmetic average. This formula avoids the misjudgment that may occur with evaluation of a single parameter and improves the reliability of corrosion diagnosis.

[0150] Formula for repeatability error index of weld point contact resistance measurement The relative standard deviation is used as a quantitative indicator of the dispersion of measurement results. This formula provides a basis for judgment of the adaptive sampling strategy. When the measurement repeatability is poor, the number of sampling times is increased to 12 to improve the confidence of the results. When the measurement repeatability is good, the standard measurement sampling time of 5 times is used to save resources, thus achieving a dynamic balance between measurement efficiency and reliability.

[0151] Zero offset formula Normalizing the zero-point drift of the measurement system as a percentage relative to the rated current provides a trigger condition for automatic calibration of the measurement reference, ensuring the accuracy of the grounding grid impedance measurement reference and avoiding the impact of system errors on corrosion diagnosis and condition assessment results.

[0152] Formula for aggregated grid distribution matrix elements An adaptive mesh generation criterion was established based on the physical characteristics of the current density gradient. By comparing the local current density gradient with a threshold, intelligent mesh density allocation was achieved. This formula automatically adopts a fine mesh in critical current dissipation regions with complex current distribution and a coarse mesh in non-critical current dissipation regions with gentle current distribution, achieving the optimal balance between computational accuracy and computational efficiency, and significantly reducing the resource consumption of large-scale grounding grid numerical calculations.

[0153] Corrosion location distribution matrix element formula By setting a corrosion judgment threshold, the binary identification and quantitative labeling of the corrosion area are realized. The formula maps the corrosion index components to a two-dimensional matrix according to the plane coordinate position of the grounding grid, forming an intuitive corrosion distribution heat map, which provides visual technical support for grounding grid maintenance decisions.

[0154] It should be noted that the variables involved in this invention are explained in detail in Tables 1 and 2.

[0155] Table 1. Variable Explanation Table (Part 1)

[0156]

[0157] Table 2. Variable Explanation Table (Part Two)

[0158]

[0159] To better understand and implement this invention, the following is a specific application scenario of the invention, Example 2: A technical team is conducting a condition assessment of the grounding grid of a large substation that has been in operation for 15 years. The grounding grid consists of 252 nodes and 187 conductor segments, with a total area of ​​approximately 8500 square meters. The designed grounding resistance value is 0.42Ω. The technical team first injected multi-frequency high-frequency test currents into each node of the grounding grid, selecting six frequency points: 1kHz, 5kHz, 10kHz, 20kHz, 50kHz, and 100kHz. The injected current amplitude was set to 30A, and test signals were injected into the 24 accessible nodes of the grounding grid using a high-precision current injection device. The acquisition system simultaneously recorded the voltage response signals at each measurement point, performing 32 repeated samplings at each frequency for each measurement point. The sampling frequency was set to 500kHz, constructing an original impedance response dataset with 24×6=144 sets of measurement data.

[0160] The technical team performed frequency domain analysis on the original impedance response dataset, applying Fast Fourier Transform (FFT) to the voltage response and injected current signals at each measurement point to calculate the complex impedance at different frequencies. At 10kHz, the resistance component at measurement point A12 was 2.35Ω and the inductance component was 15.8μH, while the resistance component at measurement point B07 was 3.67Ω and the inductance component was 28.3μH. Through frequency domain impedance analysis, a sparse impedance characteristic matrix was established with a dimension of 252×252, containing only 1478 non-zero elements, achieving a sparsity of 97.7%. The team then normalized the measured impedance values ​​by dividing them by the standard impedance value of 100Ω, forming a normalized impedance characteristic matrix.

[0161] Based on the predicted current distribution of the grounding grid, the technical team divided the grounding grid into critical current dissipation areas and non-critical current dissipation areas. Through finite element pre-analysis, the current density gradient distribution was calculated, identifying areas with current density gradients greater than 100. The area marked as the critical current dissipation area mainly includes the area around 18 grounding down conductor connection points and the area around 46 conductor cross welding points, with a total area of ​​approximately 1320 square meters. The remaining areas are marked as non-critical current-dissipating areas, including the edge conductor area and the area around the deeply buried grounding electrode, covering an area of ​​approximately 7180 square meters. A fine mesh density with a mesh size of 0.1m is used for critical flow dispersion areas, while a coarser mesh density with a mesh size of 0.5m is used for non-critical flow dispersion areas. An aggregated mesh distribution matrix is ​​then established with a dimension of 850×100.

[0162] Based on the sparse impedance characteristic matrix and the aggregated mesh distribution matrix, the technical team employed a domain decomposition parallel algorithm to calculate the potential and current density distributions at each node. The spatial domain was divided into eight non-overlapping sub-regions, each solved in parallel on an independent computational core. Information was transferred between sub-regions via interface boundary conditions. An iterative solution using the alternating Schwarz method was employed. After 127 iterations, solutions for all sub-regions converged, with a relative error reaching [value missing]. The calculated potential values ​​for each node range from 0V to 76.5V, with the maximum current density occurring near the downlead connection point, and the value being... The grounding grid potential distribution exhibits significant non-uniformity, with the largest potential gradient observed in the down conductor connection area. Multilevel fast multipole algorithms are used to accelerate matrix-vector multiplication operations, reducing computational complexity from... Reduce to For each conductor segment, the equivalent resistance and equivalent inductance are extracted to form a vector of equivalent circuit parameters with a dimension of 374.

[0163] The technical team performed kernel space and image space decomposition on the equivalent circuit parameter vector, and used singular value decomposition to calculate the singular values ​​of the stiffness matrix. Initial kernel space dimension. The initial kernel space dimension was 3. During the calculation, the kernel space dimension was monitored in real time. At the 45th iteration, the kernel space dimension increased to 4, with an increment of 1. The kernel space dimension ratio parameter was 0.33, exceeding the lower bound of the first dimension ratio interval (0.15), indicating a local connection anomaly in the grounding grid. The technical team immediately adjusted the region decomposition boundary conditions, re-divided the sub-region boundaries, increasing the original 8 sub-regions to 12, modified the coupling method of the interface boundary conditions, and recalculated the potential distribution. After the adjustment, the kernel space dimension decreased to 3, the kernel space dimension ratio parameter returned to 0, remaining within the second dimension ratio interval, and the calculation continued normally.

[0164] The technical team compared and analyzed the converted equivalent circuit parameter vector with the standard grounding grid parameter vector, which was derived from the grounding grid's design parameters and historical measurement data from five years ago. By comparing and calculating the resistance and inductance deviation indices element by element, the results showed that 23 out of 187 conductor segments had corrosion index components greater than the corrosion judgment threshold of 0.3. (See Table 2.)

[0165] Table 3 Corrosion Status Assessment Results of Partial Conductor Sections

[0166]

[0167] The corrosion index components of conductor segments C-053, C-089, C-112, and C-145 all exceeded 0.3, and were marked as corrosion areas. For example... Figure 2 As shown, the corrosion index components exhibit a clear local concentration characteristic along the conductor, with severely corroded areas mainly distributed in the northeast and southwest corners of the grounding grid. The technical team generated a corrosion location distribution matrix with dimensions of 85×100, arranging the corrosion index components according to the plane coordinates of the grounding grid to visualize the corrosion distribution of the grounding grid.

[0168] For the measurement of weld contact resistance, the technical team performed standard measurements on 64 weld points, with each weld point measured five times. The five measurement results for weld point W-027 were 0.0032Ω, 0.0029Ω, 0.0035Ω, 0.0031Ω, and 0.0033Ω, respectively, with an average weld contact resistance of 0.0032Ω and a standard deviation of [missing value]. The repeatability error of the contact resistance measurement at the weld point was 6.8%, which is less than the first allowable error value of 8%. The standard measurement sampling of 5 times met the accuracy requirements. However, the 5 measurements of weld point W-041 showed significant fluctuations, with a repeatability error reaching 11.3%, exceeding the first allowable error value of 8%. The technical team immediately increased the measurement sampling to an enhanced 12 times and re-measured, obtaining more stable results. The repeatability error was reduced to 7.2%, improving the reliability of the results.

[0169] Before testing, the technical team performed a zero-point calibration check on the current injection equipment. With the current source off, 20 samples were taken from the measurement circuit to obtain a zero-point current reading sequence. The calculated average zero-point current was 0.52A, the rated injection current was 30A, and the zero-point offset was 1.73%, less than the second allowable error value of 5%. The system continued operation while maintaining the current benchmark calibration. During testing, at the 58th measurement point, the zero-point offset increased to 6.8%, exceeding the second allowable error value of 5%. The technical team immediately executed the automatic zero-point correction program for current injection, gradually adjusting the current source bias voltage. After 8 adjustments, the absolute value of the average zero-point current decreased to 0.008A, the zero-point offset recovered to 0.027%, and the benchmark accuracy was restored.

[0170] The technical team generates a grounding grid status assessment report based on the abnormal state assessment index vector and corrosion location distribution matrix. The report includes a table of grounding impedance measurement results, a visualization of the corrosion location distribution matrix, distribution curves of corrosion index components along the conductor length, and a grounding impedance variation trend curve. Figure 3 As shown, the grounding impedance decreases with frequency, reaching 0.58Ω at 1kHz and dropping to 0.36Ω at 100kHz. Compared to historical measurements from five years ago, the grounding impedance in the low-frequency range increases significantly, reflecting the increased resistance due to conductor corrosion. The report proposes targeted maintenance recommendations: conductor sections C-053 and C-112 with corrosion index components exceeding 0.5 are recommended for priority replacement; conductor sections with corrosion index components between 0.3 and 0.5 are recommended for increased monitoring frequency. The assessment report will be output and archived electronically.

[0171] This invention achieves trenchless detection of grounding grid corrosion compared to traditional excavation verification methods. Its core principle lies in utilizing high-frequency current injection technology combined with frequency domain impedance analysis. By measuring the impedance response of the conductor at different frequencies, the resistance and inductance components are extracted. When corrosion occurs, the cross-sectional area decreases, leading to an increase in equivalent resistance; simultaneously, the conductor radius decreases, increasing the inductance. This change in physical characteristics can be accurately captured through high-frequency impedance spectroscopy measurements. Traditional methods require excavating the grounding grid conductor for direct observation and measurement, which is not only labor-intensive and costly but also damages the grounding grid structure. This invention, however, completes the detection by injecting test current at accessible points and collecting voltage response signals, avoiding physical damage to the grounding grid. The application of adaptive mesh technology and domain decomposition parallel algorithms solves the problem of low computational efficiency in fine-grained numerical calculations of large grounding grids. By using fine meshes in critical current-dissipating regions and coarse meshes in non-critical regions, the computational load is significantly reduced while maintaining accuracy. The domain decomposition parallel algorithm decomposes large-scale computational problems into multiple sub-problems for parallel solution, further improving computational efficiency. The dynamic analysis mechanism for kernel space and image space decomposition can monitor anomalies in the numerical solution process in real time. When an abnormal increase in kernel space dimension is detected, the solution strategy is automatically adjusted to ensure the robustness and accuracy of the calculation results. This adaptive capability is not available in traditional static solution methods. The adaptive measurement strategy dynamically adjusts the number of samplings based on the measurement repeatability error index, saving testing time and resources while ensuring measurement accuracy. The automatic zero-point correction mechanism monitors and corrects the zero-point drift of the measurement system in real time, ensuring the long-term stability of the measurement benchmark. These intelligent quality control measures significantly improve the reliability and engineering applicability of the test results.

[0172] The above description is merely a specific embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any changes or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in the present invention should be included within the scope of protection of the present invention.

Claims

1. A method for evaluating the state of a large grounding grid of a power grid, characterized by, The multi-frequency high-frequency test current is injected into each node of the grounding grid, and a voltage response signal is collected to construct an original impedance response data set; the original impedance response data set is analyzed in the frequency domain to extract resistance components and inductance components, and a sparse impedance characteristic matrix is calculated; The grounding grid is divided into a key current dispersion area and a non-key current dispersion area, and an aggregated grid distribution matrix is established using different grid densities; the sparse impedance characteristic matrix and the aggregated grid distribution matrix are used to calculate the potential distribution and the current density distribution using a domain decomposition parallel algorithm to obtain a transformed equivalent circuit parameter vector; the transformed equivalent circuit parameter vector is decomposed in the kernel space and the image space, and the domain decomposition boundary conditions are dynamically adjusted according to the dimension ratio parameter of the kernel space; the transformed equivalent circuit parameter vector is compared with a standard grounding grid parameter vector to calculate an abnormal state evaluation index vector and generate a corrosion position distribution matrix; the sampling frequency is adjusted according to the repeatability error index of the contact resistance measurement of the welding point; a current injection zero-point automatic correction program is executed according to the zero-point offset; and a grounding grid state evaluation report is generated to output the corrosion position distribution matrix and the grounding impedance trend curve.

2. The method of claim 1, wherein, The multi-frequency high-frequency test current is a sinusoidal current signal with a frequency range of 1 kHz to 100 kHz, and the amplitude is set to 10 A to 50 A.

3. The method of claim 2, wherein, The sparse impedance characteristic matrix is established by calculating the complex impedance using the Fourier transform, separating the real part to obtain the resistance component, and separating the imaginary part and dividing by the angular frequency to obtain the inductance component.

4. The method of claim 3, wherein, The key current dispersion area is an area in the grounding grid where the current density gradient is greater than a current density gradient threshold, and the non-key current dispersion area is an area where the current density gradient is less than the current density gradient threshold.

5. The method of claim 4, wherein, The aggregated grid distribution matrix is established by using a refined grid density for the key current dispersion area and a coarse grid density for the non-key current dispersion area.

6. The method of claim 5, wherein, The domain decomposition parallel algorithm uses a geometric decomposition strategy to divide the spatial domain into non-overlapping sub-regions, and information transmission is achieved between the sub-regions through interface boundary conditions, and the Schwarz alternating method is used for iterative solution.

7. The method of claim 6, wherein, The transformed equivalent circuit parameter vector is established by using the finite element method to establish a linear equation system and using the multilevel fast multipole algorithm to accelerate the matrix-vector multiplication operation.

8. The method of claim 7, wherein, The kernel space and image space decomposition uses singular value decomposition to calculate the kernel space and image space, and the kernel space dimension reflects the uncertainty degree of the linear equation system.

9. The method of claim 8, wherein, The kernel space dimension ratio parameter is the ratio of the current kernel space dimension increment to the initial kernel space dimension, and when the kernel space dimension ratio parameter is greater than 0.15, the domain decomposition boundary conditions are adjusted.

10. The method of claim 9, wherein, The abnormal state evaluation index vector is established by calculating the resistance deviation index and the inductance deviation index, and the corrosion index component is defined as the arithmetic mean of the resistance deviation index and the inductance deviation index.

Images