Ground grid corrosion branch positioning method and system based on node voltage difference

Abstract

The invention relates to the technical field of earth screen corrosion branch positioning, and particularly discloses an earth screen corrosion branch positioning method and system based on node voltage difference. Comprising the steps of multi-frequency composite signal generation, injection signal acquisition response, multi-frequency characteristic parameter extraction, real-time environment parameter acquisition, branch corrosion probability generation, criterion threshold dynamic generation and ground grid topology thermodynamic diagram generation. The method comprises the following steps: injecting a multi-frequency composite signal containing low-frequency and intermediate-frequency components to a ground screen, synchronously acquiring node voltage difference data, extracting multi-frequency characteristic parameters, combining real-time environment parameters, generating a branch corrosion probability, dynamically generating a self-adaptive corrosion criterion threshold value, and calculating the corrosion probability of the branch according to the self-adaptive corrosion criterion threshold value. Accurately positioning a corrosion branch and generating an earth screen topology thermodynamic diagram containing a corrosion level; according to the invention, the technical span from the judgment of the overall performance of the earth screen to the positioning of the specific corrosion branch is realized, the positioning accuracy and reliability are improved, the environment self-adaptive capability is realized, and the maintenance efficiency and scientificity of the earth screen are remarkably improved.

Description

TECHNICAL FIELD

[0001] The application belongs to the technical field of grounding grid corrosion branch positioning, and relates to a grounding grid corrosion branch positioning method and system based on node voltage difference. BACKGROUND

[0002] The grounding grid is an indispensable safety guarantee part in the power system, substation and large industrial facilities, which is composed of underground buried conductors connected to each other, and is mainly used for safely dissipating fault current and stabilizing system potential. The integrity of the grounding grid is directly related to the safety of equipment and human body. However, due to long-term burial in complex soil environment, the grounding grid conductor is inevitably subjected to electrochemical corrosion, resulting in reduction of the conductive cross-sectional area and increase of the resistivity, thereby weakening the electrical performance thereof.

[0003] At present, for the detection of grounding grid corrosion, the commonly used methods include periodic measurement of the overall grounding resistance of the grounding grid, local potential distribution measurement by using a large current source, or visual inspection by local excavation. Among them, the overall grounding resistance measurement is the most commonly used means, which measures the overall grounding resistance value of the entire grounding grid by applying a direct current or low-frequency alternating current to evaluate the overall performance thereof. In addition, some technologies also use single-frequency alternating impedance method to try to obtain more information.

[0004] However, the above existing technical means has significant defects and limitations. The overall grounding resistance measurement method can only reflect the overall performance state of the grounding grid macroscopically, and once the measurement value exceeds the standard range, it cannot accurately locate the specific branch position where the corrosion occurs, which leads to lack of pertinence, great blindness and low efficiency in the maintenance work, and it is difficult to quickly solve the potential safety hazards. The local excavation inspection method is not only costly, time-consuming and labor-intensive, but also destructive, causing irreversible damage to the grounding grid structure, and can only check a limited area, leaving a large detection blind area and being unable to comprehensively evaluate the health status of the grounding grid. The measurement method based on a single electrical parameter is easily disturbed by complex environmental factors such as soil resistivity variation and stray current, resulting in large fluctuations in the diagnosis results, significantly reducing the accuracy and reliability, and being unable to provide stable and reliable decision basis for the grounding grid maintenance. SUMMARY

[0005] In view of this, in order to solve the problems raised in the background art, the grounding grid corrosion branch positioning method and system based on node voltage difference are proposed.

[0006] The purpose of the application can be achieved by the following technical solutions: the first aspect of the application provides a grounding grid corrosion branch positioning method based on node voltage difference, which comprises the following steps: S1, generating a multi-frequency composite signal containing low-frequency components and medium-frequency components.

[0007] S2. Injection signal acquisition response: The multi-frequency composite signal is injected into the ground grid as an excitation source, and the voltage response of multiple key nodes in the ground grid is acquired simultaneously to obtain node voltage difference data.

[0008] S3. Extraction of multi-frequency feature parameters: Perform frequency domain separation on the node voltage difference data and extract multi-frequency feature parameters from the separated frequency components.

[0009] S4. Real-time environmental parameter acquisition: Acquire real-time environmental parameters that characterize the state of the medium in which the ground network is located.

[0010] S5. Branch Corrosion Probability Generation: Input the multi-frequency feature parameters and the real-time environmental parameters into a trained intelligent diagnostic model to generate a corrosion probability characterizing the possibility of branch corrosion.

[0011] S6. Dynamic generation of criterion threshold: Based on the corrosion probability and the real-time environmental parameters, an adaptive corrosion criterion threshold is dynamically generated.

[0012] S7. Generating a ground network topology heat map: Based on the adaptive corrosion criterion threshold, the multi-frequency characteristic parameters are judged to locate the corrosion branches and generate a ground network topology heat map containing the corrosion level.

[0013] The second aspect of the present invention provides a ground grid corrosion branch location system based on node voltage difference, comprising: a multi-frequency composite signal generation module, which generates a multi-frequency composite signal containing low-frequency components and mid-frequency components.

[0014] The injection signal acquisition and response module injects the multi-frequency composite signal as an excitation source into the ground grid and simultaneously acquires the voltage response of multiple key nodes in the ground grid to obtain node voltage difference data.

[0015] The multi-frequency feature parameter extraction module performs frequency domain separation on the node voltage difference data and extracts multi-frequency feature parameters from the separated frequency components.

[0016] The real-time environmental parameter acquisition module acquires real-time environmental parameters that characterize the state of the medium in which the ground network is located.

[0017] The branch corrosion probability generation module inputs the multi-frequency feature parameters and the real-time environmental parameters into a trained intelligent diagnostic model to generate a corrosion probability characterizing the possibility of branch corrosion.

[0018] The dynamic generation module for the criterion threshold dynamically generates an adaptive corrosion criterion threshold based on the corrosion probability and the real-time environmental parameters.

[0019] The ground network topology heat map generation module makes a judgment on the multi-frequency characteristic parameters based on the adaptive corrosion criterion threshold, so as to locate the corrosion branches and generate a ground network topology heat map containing the corrosion level.

[0020] Compared with the prior art, the embodiments of the present invention have at least the following advantages or beneficial effects: (1) By injecting multi-frequency composite signals and extracting multi-frequency characteristic parameters including phase, amplitude and frequency response slope, the present invention can comprehensively evaluate the branch status from multiple dimensions, effectively distinguish different types of corrosion modes, thereby realizing the technical leap from judging the overall performance degradation of the ground network to accurately locating specific corrosion branches, and significantly improving the accuracy of positioning.

[0021] (2) By introducing real-time environmental parameters and combining historical data with intelligent diagnostic models to dynamically generate adaptive corrosion criterion thresholds, this invention can effectively eliminate measurement interference caused by changes in environmental factors such as soil moisture and temperature, enabling the diagnostic system to have environmental adaptability, greatly improving the stability and reliability of diagnostic results under different working conditions, and avoiding false alarms and missed alarms that are easy to be generated by traditional fixed threshold methods.

[0022] (3) This invention transforms complex diagnostic data into an intuitive ground network topology heat map and automatically generates a maintenance report containing priorities and operation suggestions based on the importance of branches. This achieves a closed loop from data analysis to decision support, enabling maintenance personnel to quickly and comprehensively grasp the health status of the ground network and take targeted measures, which significantly improves the efficiency and scientific nature of ground network maintenance. Attached Figure Description

[0023] To more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0024] Figure 1 This is a schematic diagram of the method steps of the present invention.

[0025] Figure 2 This is a schematic diagram of the system structure connection of the present invention.

[0026] Figure 3 This is a schematic diagram of the ground network topology heat map of the present invention. Detailed Implementation

[0027] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0028] Please see Figure 1 The first aspect of the present invention provides a method for locating corrosion branches of ground grid based on node voltage difference, including: S1, generating a multi-frequency composite signal: generating a multi-frequency composite signal containing low-frequency components and mid-frequency components.

[0029] In a specific embodiment of the present invention, the step of generating a multi-frequency composite signal containing low-frequency and mid-frequency components includes: using a programmable current source, and based on the scale of the ground grid and the estimated soil resistivity, setting the current amplitude of the low-frequency and mid-frequency components.

[0030] It should be noted that the specific process of setting the current amplitude of the low-frequency and mid-frequency components includes: 1) assessing the scale of the grounding grid and soil resistivity: First, it is necessary to assess the scale of the target grounding grid, including the coverage area, conductor length, connection method, etc. At the same time, the soil resistivity of the area where the grounding grid is located is estimated, which can be done through geological exploration, soil sampling tests or by referring to existing geological data. 2) Principles for setting current amplitude: Setting the current amplitude of low-frequency components: Low-frequency electromagnetic waves have stronger penetration ability and less attenuation in conductive media, making them suitable for detecting the macroscopic corrosion status of deep buried geogrid main branches. Therefore, the current amplitude of low-frequency components should be set relatively large to ensure that the signal has enough energy to penetrate deep soil and obtain the electrical response of the main branches. Setting the current amplitude of mid-frequency components: Mid-frequency signals are more sensitive to impedance changes on the conductor surface and in the near-field region, and can accurately reflect the local corrosion characteristics caused by the reduction of cross-sectional area in shallow branches. The current amplitude of mid-frequency components can be set according to the level of detail of the shallow branches to be detected, ensuring signal sensitivity while avoiding excessive amplitude that could lead to signal distortion or interference. 3) Specific setting steps: Use a programmable current source as a signal generator. The current source can receive digital instructions and output a preset current signal. Based on the estimated scale of the ground grid and the soil resistivity, the current amplitude of the low-frequency and medium-frequency components output by the programmable current source is set by programming. When setting, you can refer to the experience data of similar ground grid projects, or determine the appropriate amplitude range through preliminary tests.

[0031] The programmable current source is controlled to output a current signal that is a linear superposition of the low-frequency and mid-frequency components.

[0032] The current signal is used as the multi-frequency composite signal.

[0033] It should be noted that, in order to generate an excitation source capable of comprehensively detecting different depths of corrosion in the ground network, this method employs a programmable current source. This programmable current source receives digital instructions and generates a multi-frequency composite signal composed of a linear superposition of low-frequency and mid-frequency components. The instantaneous current of the multi-frequency composite signal can be expressed by the following function: ,in, Represents a point in time At that time, the instantaneous current value of the multi-frequency composite signal output by the programmable current source. and These represent the current amplitudes for the low-frequency and mid-frequency components, respectively. The frequency of the low-frequency component is set. The frequency of the intermediate frequency component is set. By programming a controllable current source to accurately output a current that conforms to the above function waveform, a multi-frequency composite signal for subsequent injection steps is obtained.

[0034] In one specific embodiment of the present invention, the frequency of the low-frequency component can be set to 5Hz. This is because, in common ground network detection scenarios, this frequency, considering the general burial depth of the ground network and the medium soil resistivity, can reach the main branches of the deeply buried ground network with the strong penetrating power of low-frequency electromagnetic waves, while maintaining good characteristics in signal processing and other aspects, effectively detecting its macroscopic corrosion status. The frequency of the medium-frequency component can be set to 500Hz, because this frequency is highly sensitive to the impedance changes of the conductor surface and near-field region. When facing most shallow branch structures, it can accurately reflect the local corrosion characteristics caused by the reduction of cross-sectional area, meeting the accuracy requirements for shallow branch corrosion detection.

[0035] This invention, through the generation of a multi-frequency composite signal containing low-frequency and mid-frequency components, enables the simultaneous acquisition of both the macroscopic overall health status and microscopic local corrosion details of the ground network in a single measurement. The low-frequency components provide detection capabilities for the deep layers and main structure of the ground network, while the mid-frequency components focus on the detailed diagnosis of shallow layers and terminal branches. The synergistic effect of these two frequency components allows the system to construct a more three-dimensional and comprehensive image of the ground network corrosion, effectively overcoming the technical limitations of single-frequency detection methods that cannot simultaneously achieve both detection depth and accuracy. It also avoids misjudgments or omissions caused by a single information dimension, providing a rich and reliable raw data foundation for subsequent precise location of corroded branches.

[0036] S2. Injection signal acquisition response: The multi-frequency composite signal is injected into the ground grid as an excitation source, and the voltage response of multiple key nodes in the ground grid is acquired simultaneously to obtain node voltage difference data.

[0037] In a specific embodiment of the present invention, the step of synchronously acquiring the voltage response of multiple key nodes in the grounding network to obtain node voltage difference data includes: configuring a timing module for receiving standard time signals for each voltage sensor deployed at the key nodes.

[0038] The timing module generates a synchronization reference to unify the sampling clocks of each voltage sensor.

[0039] The voltage response of each node is collected based on the unified sampling clock, and the difference in voltage response between adjacent nodes is calculated to obtain the node voltage difference data.

[0040] It should be noted that after injecting the multi-frequency composite signal into the grounding network, high-precision voltage sensors need to be deployed at pre-planned key nodes of the grounding network to accurately capture the potential changes generated as the signal propagates through the underground branch network. Key nodes of the grounding network refer to the main conductor junctions, the connection points between equipment grounding leads and the main network, and representative nodes at the edge of the grounding network. Because these nodes are widely distributed, traditional data acquisition methods cannot guarantee that all sensors sample voltage at the same time. Small time deviations can produce significant phase errors when processing high-frequency signals, thus distorting the true voltage difference. To solve this problem, this method integrates a GPS timing module into each voltage sensor. This GPS timing module receives standard time signals broadcast from Global Positioning System satellites and decodes them into a high-precision pulses per second (PPS) signal. This PPS signal is used as a global synchronization reference, triggering the analog-to-digital converters of each independent voltage sensor to sample at exactly the same time, thus forming a unified sampling clock. Under the control of this unified sampling clock, each sensor synchronously records the instantaneous voltage response of its respective node. The collected voltage response data from each node is then appended with a precise timestamp and transmitted to the data processing unit. The data processing unit, based on the topology of the ground network, processes data from two adjacent nodes. and At the same sampling time The voltage responses are subtracted to obtain the node voltage difference data at both ends of the branch. , ,in, It is a node and Between moments The node voltage difference, and These are nodes and At the same synchronous sampling time The voltage response. By performing this operation on all adjacent node pairs, time-seriesd node voltage difference data covering the entire monitoring area is finally generated.

[0041] This invention employs a GPS timing module to force synchronous acquisition of all measurement points, fundamentally eliminating measurement errors caused by inconsistent sampling times. This high-precision synchronization mechanism ensures that the obtained node voltage difference data accurately reflects the true phase delay and amplitude attenuation generated when multi-frequency composite signals propagate along different branches, guaranteeing the integrity and fidelity of the original data. This technical effect is a necessary prerequisite for subsequent precise multi-frequency response separation and feature extraction steps. It significantly improves the signal-to-noise ratio and reliability of the entire positioning method, making it possible to identify subtle feature changes caused by corrosion from complex voltage signals, thus laying a solid data foundation for ultimately achieving accurate corrosion branch positioning.

[0042] S3. Extraction of multi-frequency feature parameters: Perform frequency domain separation on the node voltage difference data and extract multi-frequency feature parameters from the separated frequency components.

[0043] In a specific embodiment of the present invention, the step of extracting multi-frequency feature parameters from the separated frequency components includes: performing a Fourier transform on the node voltage difference data to obtain a frequency domain complex spectrum.

[0044] It should be noted that the specific process of performing a Fourier transform on the node voltage difference data to obtain a frequency domain complex spectrum is as follows: after obtaining the node voltage difference data containing timestamps, the fast Fourier transform algorithm is used to process the time series of node voltage difference data for each branch. This algorithm converts time-domain signals Convert to frequency domain complex spectrum This complex spectrum can clearly show the amplitude and phase information of the signal at different frequency points. For example, it can clearly show the response of the low-frequency and mid-frequency components in the multi-frequency composite signal injected into the ground network, thus providing basic data for the subsequent extraction of characteristic parameters such as low-frequency phase offset and mid-frequency amplitude attenuation rate. The Fourier transform algorithm is a mature existing technology and will not be elaborated here.

[0045] The responses corresponding to the low-frequency and mid-frequency components are identified from the complex spectrum in the frequency domain, and their low-frequency phase shift and mid-frequency amplitude attenuation rate are extracted respectively.

[0046] It should be noted that, since the injected multi-frequency composite signal contains known low-frequency components, and the frequency of the mid-frequency component We focus on the responses at these two specific frequency points in the frequency domain spectrum. First, we query the frequency domain spectrum... and The value of is used to separate the low-frequency and mid-frequency components. Then, feature extraction is performed.

[0047] The first characteristic is the low-frequency phase shift. It consists of low-frequency components. The phase angle is determined. ,in This parameter represents the phase angle operation when taking complex numbers. It primarily reflects the total time delay caused by changes in the macroscopic resistance and reactance characteristics of the conductor as current flows through the branch, and is sensitive to impedance changes caused by overall corrosion.

[0048] The second characteristic is the mid-frequency amplitude attenuation rate, which is determined by the mid-frequency component. amplitude Characterization. This amplitude is directly related to the degree of signal energy attenuation at that frequency. Due to the skin effect of mid-frequency current, its attenuation in the mid-frequency range is particularly sensitive to the reduction of conductor cross-sectional area.

[0049] A frequency response curve slope is calculated based on the response amplitudes of the low-frequency and mid-frequency components, and the low-frequency phase offset, mid-frequency amplitude attenuation rate, and frequency response curve slope are combined to form the multi-frequency characteristic parameters.

[0050] It should be noted that the slope of the frequency response curve quantifies the trend of branch impedance changing with frequency. The slope of the frequency response curve can be calculated using the known amplitude responses at the low and mid-frequency points. , ,in and These represent the response amplitudes of the mid-frequency and low-frequency components, respectively. The slope reveals the uniformity of the corrosion distribution; uniform corrosion and localized corrosion have significantly different effects on the frequency response curve shape. Finally, these three independent but complementary physical quantities—low-frequency phase shift, mid-frequency amplitude attenuation rate, and frequency response curve slope—are combined into a multi-dimensional vector to form the aforementioned multi-frequency characteristic parameters, serving as a comprehensive quantitative description of the corrosion state of this branch.

[0051] This invention extracts a set of multi-frequency characteristic parameters with clear physical meaning from raw, mixed node voltage difference data. This process is not merely data dimensionality reduction, but also a deep mining and synergistic enhancement of information. By combining low-frequency phase shift, mid-frequency amplitude attenuation rate, and frequency response curve slope, a three-dimensional corrosion status evaluation system is constructed, achieving a leap from a simple judgment of "whether corrosion exists" to a refined characterization of "corrosion degree and type." This multi-feature fusion analysis method can effectively distinguish interference caused by non-corrosion factors such as changes in soil environment. Its robustness and diagnostic accuracy far exceed traditional methods that rely on a single frequency or single parameter, providing a highly reliable input basis for subsequent adaptive threshold correction and precise positioning.

[0052] S4. Real-time environmental parameter acquisition: Acquire real-time environmental parameters that characterize the state of the medium in which the ground network is located.

[0053] In a specific embodiment of the present invention, the step of obtaining real-time environmental parameters characterizing the state of the medium in which the ground grid is located includes: deploying soil moisture sensors and temperature sensors within the ground grid coverage area.

[0054] The soil moisture sensor and temperature sensor respectively collect the volumetric water content and temperature readings of the soil.

[0055] The volumetric moisture content and temperature readings are combined to form the real-time environmental parameters.

[0056] It should be noted that, in order to eliminate the interference of changes in soil physical properties on electrical measurement results, this method simultaneously collects key environmental factors affecting soil conductivity while acquiring the grounding grid's electrical response. Specifically, soil moisture sensors and temperature sensors are strategically buried at multiple representative locations within the grounding grid's coverage area. The placement and depth of these sensors are carefully considered to reflect the average environmental state of the soil layer where the grounding grid is located. The soil moisture sensor accurately inverts the soil's volumetric water content by measuring the propagation speed of electromagnetic waves in the soil medium. The temperature sensor directly measures the soil temperature at the corresponding location. These sensors are connected to a data recording unit synchronized with the voltage acquisition system, ensuring that real-time soil moisture and temperature values ​​are acquired at the moment of each electrical measurement. The acquired raw sensor signals are processed and calibrated by internal circuitry, converted into standard physical quantity units. This quantified and timestamped set of soil moisture and temperature data constitutes the aforementioned environmental parameters.

[0057] S5. Branch Corrosion Probability Generation: Input the multi-frequency feature parameters and the real-time environmental parameters into a trained intelligent diagnostic model to generate a corrosion probability characterizing the possibility of branch corrosion.

[0058] In a specific embodiment of the present invention, the step of generating the corrosion probability characterizing the possibility of branch corrosion includes: combining the newly collected multi-frequency feature parameters and real-time environmental parameters into an input vector during actual operation; The input vector is fed into a trained intelligent diagnostic model, which then performs complex reasoning on the input vector based on the parameters and structure it has learned internally. Finally, the output layer outputs a value between 0 and 1 as the corrosion probability, representing the possibility of branch corrosion. Here, 0 indicates no possibility of corrosion and 1 indicates that corrosion is certain to occur.

[0059] It should be noted that the detailed explanation of the intelligent diagnostic model is as follows: 1) Structural composition: Input layer: Receives multi-frequency feature parameters and real-time environmental parameters as input signals. The number of neurons in the input layer depends on the number of input parameters. Hidden layer: Located between the input layer and the output layer, there can be one or more layers. Neurons in the hidden layer perform complex nonlinear transformations on the input signals. Each neuron is connected to the neurons in the previous layer through weights and processes the input signals through an activation function. Output layer: Outputs the final corrosion probability. The number of neurons in the output layer is usually 1. Its output value is processed by an activation function to ensure it is between 0 and 1. 2) Working principle: When input data enters the input layer, the signal is transmitted to the hidden layer through weights. The neurons in the hidden layer perform weighted summation on the input signal and perform nonlinear transformation through an activation function. After processing by the hidden layer, the signal is transmitted to the output layer. The neurons in the output layer also perform weighted summation and activation function processing, finally outputting the corrosion probability.

[0060] It should also be noted that the training process of the intelligent diagnostic model is as follows: 1) Data collection and labeling: Collect a large amount of historical data, including multi-frequency feature parameters, real-time environmental parameters, and corresponding true labels of branch corrosion status, such as whether corrosion has occurred. This data can be obtained from actual ground network detection projects or generated through simulation experiments; label the collected data, clearly marking the corrosion status of each sample as the target for model training. 2) Data preprocessing: Clean the collected data, removing outliers and noise; normalize or standardize the data to ensure that the numerical ranges of different features are within similar intervals, avoiding excessive influence of certain features on model training. 3) Model initialization: Select a suitable neural network structure and determine the number of neurons in the input layer, hidden layer, and output layer; initialize the weights and biases in the model, usually using random initialization. 4) Forward propagation: Input the preprocessed training data into the model, calculate the model's output prediction value according to the model's structure and weights. 5) Loss calculation: Calculate the value of the loss function based on the model's output prediction value and the true labels. Commonly used loss functions include the mean squared error loss function, which measures the difference between the model's predicted value and the true value. 6) Backpropagation and parameter update: The backpropagation algorithm calculates the gradient of the loss function with respect to each weight and bias in the model. Based on the calculated gradient, optimization algorithms, such as stochastic gradient descent, are used to update the model's weights and biases to reduce the value of the loss function. 7) Iterative training: The process of forward propagation, loss calculation, backpropagation, and parameter update is repeated until the model's loss function value on the training data reaches a small threshold or the preset number of iterations is reached. Through the above training process, the intelligent diagnostic model can learn the complex relationship between multi-frequency feature parameters, real-time environmental parameters, and branch erosion probability, thereby accurately generating the erosion probability of branches in practical applications.

[0061] S6. Dynamic generation of criterion threshold: Based on the corrosion probability and the real-time environmental parameters, an adaptive corrosion criterion threshold is dynamically generated.

[0062] In a specific embodiment of the present invention, the specific steps of dynamically generating an adaptive corrosion criterion threshold include: retrieving a subset of historical data from a historical database that matches the real-time environmental parameters.

[0063] It should be noted that, in order to adapt the corrosion diagnosis to the complex and ever-changing soil environment of the grounding grid, this method does not use a fixed threshold, but instead generates an adaptive corrosion threshold that changes with the environment through an adaptive threshold dynamic correction algorithm. First, the system maintains a threshold containing recent data... The system uses a historical database of measurement results. Each record in the database contains a set of multi-frequency characteristic parameters and their corresponding environmental parameters at the time of acquisition. When a new corrosion detection is performed, the system first acquires the current real-time environmental parameters. Subsequently, a sliding window algorithm automatically searches the historical database and selects a subset of historical records that are similar to the current real-time environmental parameters.

[0064] A statistical baseline representing normal fluctuations is calculated based on the aforementioned subset of historical data.

[0065] It should be noted that, based on this selected subset of historical data representing similar environmental conditions, the statistical baseline of the multi-frequency characteristic parameters under this specific environment is calculated, which is usually the mean and standard deviation.

[0066] The corrosion probability is used as an adjustment factor to correct the statistical baseline in order to generate the adaptive corrosion criterion threshold.

[0067] In a specific embodiment of the present invention, the specific steps for correcting the statistical baseline include: determining a sensitivity coefficient based on the corrosion probability.

[0068] It should be noted that the specific method for determining a sensitivity coefficient based on the corrosion probability is as follows: the corrosion probability is compared with the corrosion probability intervals corresponding to each sensitivity coefficient stored in the database. If the corrosion probability is within the corrosion probability interval corresponding to a certain sensitivity coefficient, then that sensitivity coefficient is used as the sensitivity coefficient determined by the corrosion probability.

[0069] The sensitivity coefficient is applied to the standard deviation term in the statistical baseline.

[0070] The corrected standard deviation term is added to the mean term in the statistical baseline to generate the adaptive corrosion criterion threshold.

[0071] It should be noted that the adaptive corrosion criterion threshold for this measurement is dynamically generated based on the statistical baseline. The adaptive corrosion criterion threshold can be determined by the following functional relationship: In this relation, That is, the generated adaptive corrosion criterion threshold. It is the statistical average of multi-frequency characteristic parameters under historical conditions similar to the current environment. It is the corresponding statistical standard deviation, which reflects the normal fluctuation range of the parameter under the environmental conditions. It is a sensitivity coefficient determined based on the corrosion probability.

[0072] This invention endows the entire diagnostic system with environmental awareness and self-learning capabilities, achieving a qualitative leap from static diagnosis to dynamic adaptive diagnosis. By establishing a dynamic correlation between electrical response and environmental parameters, the system can intelligently distinguish between characteristic parameter anomalies caused by branch corrosion and normal parameter drifts caused by environmental fluctuations. This capability significantly reduces false alarms and missed alarms caused by environmental changes, substantially improving the accuracy of diagnostic results and the stability of long-term monitoring. This method enables the system to adapt to various harsh and variable field conditions, ensuring the consistency and comparability of diagnostic conclusions across different times and seasons, thus providing a highly reliable decision-making basis for long-term health status assessment and preventative maintenance of the grounding network.

[0073] Please see Figure 3 S7. Generating a ground network topology heat map: Based on the adaptive corrosion criterion threshold, the multi-frequency feature parameters are judged to locate the corrosion branches and generate a ground network topology heat map containing the corrosion level.

[0074] In a specific embodiment of the present invention, the step of locating corroded branches and generating a ground network topology heatmap containing corrosion levels includes: comparing the multi-frequency feature parameters extracted in real time with an adaptive corrosion criterion threshold; if any feature parameter in a branch is greater than the adaptive corrosion criterion threshold, the branch is determined to have corrosion risk and is defined as a corroded branch.

[0075] The corrosion probability is mapped to a color gradient scale to generate the ground network topology heatmap.

[0076] Based on the comparison between the corrosion probability and a set of corrosion level classification thresholds, a corrosion level is determined for each branch in the grounding network.

[0077] Branches with high corrosion levels are highlighted on the ground network topology heat map.

[0078] It's important to note that the goal of this step is to transform the abstract corrosion probability distribution calculated in the previous steps into an intuitive and instructive operational view. First, the system loads a pre-built digital topology model of the ground network, which graphically and accurately describes the physical layout and connections of all branches in the network. Then, the system matches the obtained corrosion probability distribution with this topology model, assigning a corresponding corrosion probability value to each branch in the model. Based on this, the system uses color mapping technology to generate a ground network topology heatmap. A preset color gradient scale is established, mapping the continuous range of corrosion probability values ​​to a set of visually distinguishable colors, such as a smooth transition from low-probability cool tones to high-probability warm tones. The system iterates through all branches in the topology model, rendering the graphical elements of each branch using the corresponding color from the color gradient scale based on its corrosion probability value.

[0079] Meanwhile, to provide more specific treatment recommendations, the system discretizes the continuous corrosion probability values ​​into specific corrosion levels. This is achieved through a set of preset corrosion level classification thresholds. ,in, Represents the corrosion level. It is the corrosion probability of the branch. It is a piecewise function, if Below the first threshold The corrosion level is... Defined as mild; if Between the first threshold Second threshold If it falls between these values, the level is moderate; if... Above the second threshold If the corrosion level is high, the corrosion level is classified as severe. Finally, on the generated ground network topology heat map, the system highlights or adds specific graphic markers to all high-probability corrosion branches that are judged to be of moderate or severe corrosion level, along with textual annotations of their corrosion level.

[0080] In a specific embodiment of the present invention, the first threshold It can be set to 0.3, the second threshold. It can be set to 0.6. The value is based on a combination of extensive actual data and engineering experience from ground network corrosion detection, considering the probability of branch corrosion. Below 0.3, the likelihood of corrosion is low, defined as a mild corrosion level that can reasonably distinguish situations with no significant corrosion risk; when When the value is between 0.3 and 0.6, the corrosion risk is at a moderate level, and setting it to a moderate corrosion level is consistent with the judgment of the actual corrosion situation; when... When the value is above 0.6, the probability of corrosion is relatively high. Defining it as a severe corrosion level helps to identify branches with high corrosion risk in a timely manner, so that targeted maintenance measures can be taken to ensure the safe and stable operation of the grounding network.

[0081] This invention successfully transforms complex, discrete diagnostic data into a global, visualized situational awareness image. The ground network topology heatmap allows maintenance personnel to quickly and clearly grasp the overall health status of the ground network, identify concentrated areas of corrosion and their spread trends. By further annotating the locations of high-probability corrosion branches and their corrosion levels, this method provides direct, explicit, and actionable maintenance instructions, greatly improving the targeting and efficiency of troubleshooting and maintenance work.

[0082] Reference Figure 2The second aspect of the present invention provides a ground grid corrosion branch location system based on node voltage difference, comprising: a multi-frequency composite signal generation module, an injection signal acquisition and response module, a multi-frequency feature parameter extraction module, a real-time environmental parameter acquisition module, a branch corrosion probability generation module, a criterion threshold dynamic generation module, and a ground grid topology heat map generation module.

[0083] It should be noted that the present invention also includes a database for storing the corrosion probability ranges corresponding to each sensitivity coefficient.

[0084] The multi-frequency composite signal generation module is connected to the injection signal acquisition and response module. The injection signal acquisition and response module is connected to the multi-frequency feature parameter extraction module. Both the multi-frequency feature parameter extraction module and the real-time environmental parameter acquisition module are connected to the branch corrosion probability generation module. Both the real-time environmental parameter acquisition module and the branch corrosion probability generation module are connected to the criterion threshold dynamic generation module. Both the multi-frequency feature parameter extraction module and the criterion threshold dynamic generation module are connected to the ground network topology heat map generation module. The criterion threshold dynamic generation module is connected to the database.

[0085] The multi-frequency composite signal generation module generates a multi-frequency composite signal containing low-frequency and mid-frequency components.

[0086] The injected signal acquisition and response module injects the multi-frequency composite signal as an excitation source into the ground grid and simultaneously acquires the voltage response of multiple key nodes in the ground grid to obtain node voltage difference data.

[0087] The multi-frequency feature parameter extraction module performs frequency domain separation on the node voltage difference data and extracts multi-frequency feature parameters from the separated frequency components.

[0088] The real-time environmental parameter acquisition module acquires real-time environmental parameters that characterize the state of the medium in which the ground network is located.

[0089] The branch corrosion probability generation module inputs the multi-frequency feature parameters and the real-time environmental parameters into a trained intelligent diagnostic model to generate a corrosion probability characterizing the possibility of branch corrosion.

[0090] The dynamic generation module for the criterion threshold dynamically generates an adaptive corrosion criterion threshold based on the corrosion probability and the real-time environmental parameters.

[0091] The ground network topology heat map generation module makes a judgment on the multi-frequency feature parameters based on the adaptive corrosion criterion threshold in order to locate the corrosion branches and generate a ground network topology heat map containing the corrosion level.

[0092] The above content is merely an example and illustration of the concept of the present invention. Those skilled in the art can make various modifications or additions to the specific embodiments described, or use similar methods to replace them, as long as they do not deviate from the concept of the invention or exceed the scope defined by the present invention, and all such modifications and additions should fall within the protection scope of the present invention.

Claims

1. A method for locating corroded branches in a grounding grid based on node voltage differences, characterized in that, include: S1. Multi-frequency composite signal generation: Generates a multi-frequency composite signal containing low-frequency and mid-frequency components; S2, Injection Signal Acquisition Response: The multi-frequency composite signal is injected into the ground grid as an excitation source, and the voltage response of multiple key nodes in the ground grid is acquired simultaneously to obtain node voltage difference data. S3. Multi-frequency feature parameter extraction: The node voltage difference data is separated in the frequency domain, and multi-frequency feature parameters are extracted from the separated frequency components. S4. Real-time environmental parameter acquisition: Acquire real-time environmental parameters that characterize the state of the medium in which the ground network is located; S5. Branch corrosion probability generation: Input the multi-frequency feature parameters and the real-time environmental parameters into a trained intelligent diagnostic model to generate a corrosion probability characterizing the possibility of branch corrosion. S6. Dynamic generation of criterion threshold: Based on the corrosion probability and the real-time environmental parameters, an adaptive corrosion criterion threshold is dynamically generated; S7. Generating a ground network topology heat map: Based on the adaptive corrosion criterion threshold, the multi-frequency characteristic parameters are judged to locate the corrosion branches and generate a ground network topology heat map containing the corrosion level.

2. The method for locating ground grid corrosion branches based on node voltage difference according to claim 1, characterized in that: The step of generating a multi-frequency composite signal containing low-frequency and mid-frequency components includes: Using a programmable current source, the current amplitudes of low-frequency and medium-frequency components are set based on the scale of the ground grid and the estimated soil resistivity. Control the programmable current source to output a current signal that is a linear superposition of the low-frequency component and the mid-frequency component; The current signal is used as the multi-frequency composite signal.

3. The method for locating ground grid corrosion branches based on node voltage difference according to claim 1, characterized in that: The step of synchronously acquiring the voltage response of multiple key nodes within the grounding network to obtain node voltage difference data includes: Configure a timing module for receiving standard time signals for each voltage sensor deployed at the critical node; The timing module generates a synchronization reference to unify the sampling clocks of each voltage sensor. The voltage response of each node is collected based on the unified sampling clock, and the difference in voltage response between adjacent nodes is calculated to obtain the node voltage difference data.

4. The method for locating ground grid corrosion branches based on node voltage difference according to claim 1, characterized in that: The step of extracting multi-frequency feature parameters from the separated frequency components includes: Perform a Fourier transform on the node voltage difference data to obtain a frequency domain complex spectrum; The responses corresponding to the low-frequency and mid-frequency components are identified from the complex spectrum in the frequency domain, and their low-frequency phase shift and mid-frequency amplitude attenuation rate are extracted respectively. A frequency response curve slope is calculated based on the response amplitudes of the low-frequency and mid-frequency components, and the low-frequency phase offset, mid-frequency amplitude attenuation rate, and frequency response curve slope are combined to form the multi-frequency characteristic parameters.

5. The method for locating ground grid corrosion branches based on node voltage difference according to claim 1, characterized in that: The steps for obtaining real-time environmental parameters characterizing the state of the medium in which the grounding grid is located include: Soil moisture sensors and temperature sensors are deployed within the area covered by the ground grid. The soil volumetric water content and temperature readings are collected by the soil moisture sensor and temperature sensor, respectively. The volumetric moisture content and temperature readings are combined to form the real-time environmental parameters.

6. The method for locating ground grid corrosion branches based on node voltage difference according to claim 1, characterized in that: The step of generating the corrosion probability characterizing the likelihood of branch corrosion includes: In actual operation, the newly acquired multi-frequency feature parameters and real-time environmental parameters are combined into an input vector; The input vector is fed into a trained intelligent diagnostic model, which then performs complex reasoning on the input vector based on the parameters and structure it has learned internally. Finally, the output layer outputs a value between 0 and 1 as the corrosion probability, representing the possibility of branch corrosion. Here, 0 indicates no possibility of corrosion and 1 indicates that corrosion is certain to occur.

7. The method for locating ground grid corrosion branches based on node voltage difference according to claim 1, characterized in that: The specific steps for dynamically generating an adaptive corrosion criterion threshold include: Retrieve a subset of historical data that matches the real-time environmental parameters from a historical database; A statistical baseline representing normal fluctuations is calculated based on the aforementioned subset of historical data; The corrosion probability is used as an adjustment factor to correct the statistical baseline in order to generate the adaptive corrosion criterion threshold.

8. The method for locating ground grid corrosion branches based on node voltage difference according to claim 7, characterized in that: The specific steps for correcting the statistical baseline include: A sensitivity coefficient is determined based on the corrosion probability; The sensitivity coefficient is applied to the standard deviation term in the statistical baseline; The corrected standard deviation term is added to the mean term in the statistical baseline to generate the adaptive corrosion criterion threshold.

9. The method for locating ground grid corrosion branches based on node voltage difference according to claim 1, characterized in that: The steps of locating eroded branches and generating a ground network topology heatmap containing erosion levels include: The multi-frequency feature parameters extracted in real time are compared with the adaptive corrosion criterion threshold. If any feature parameter in a branch is greater than the adaptive corrosion criterion threshold, the branch is determined to have corrosion risk and is defined as a corrosion branch. The corrosion probability is mapped to a color gradient scale to generate the ground network topology heatmap; Based on the comparison results between the corrosion probability and a set of corrosion level classification thresholds, a corrosion level is determined for each branch in the grounding network; Branches with high corrosion levels are highlighted on the ground network topology heat map.

10. A ground grid corrosion branch location system based on node voltage difference, characterized in that, include: The multi-frequency composite signal generation module generates a multi-frequency composite signal containing low-frequency and mid-frequency components. The injection signal acquisition and response module injects the multi-frequency composite signal as an excitation source into the ground grid and simultaneously acquires the voltage response of multiple key nodes in the ground grid to obtain node voltage difference data. The multi-frequency feature parameter extraction module performs frequency domain separation on the node voltage difference data and extracts multi-frequency feature parameters from the separated frequency components. The real-time environmental parameter acquisition module acquires real-time environmental parameters that characterize the state of the medium in which the ground network is located. The branch corrosion probability generation module inputs the multi-frequency feature parameters and the real-time environmental parameters into a trained intelligent diagnostic model to generate a corrosion probability characterizing the possibility of branch corrosion. The dynamic generation module for the criterion threshold dynamically generates an adaptive corrosion criterion threshold based on the corrosion probability and the real-time environmental parameters. The ground network topology heat map generation module makes a judgment on the multi-frequency characteristic parameters based on the adaptive corrosion criterion threshold, so as to locate the corrosion branches and generate a ground network topology heat map containing the corrosion level.

Images