Tower grounding grid multi-frequency complex wave electric field imaging and soil environment medium diagnosis method

Abstract

The invention discloses a tower grounding grid multi-frequency composite wave electric field imaging and soil environment medium diagnosis method, which comprises the following steps of: generating a composite excitation signal through superposition of a low-frequency sine wave and a high-frequency CHIRP square wave, and collecting multi-dimensional data such as voltage / current response, soil temperature and humidity, electric field intensity and the like in combination with a distributed sensor array; separating signal features by using wavelet transform and extracting impedance information of a target frequency point; a multi-grid homotopy algorithm is adopted to realize grounding grid three-dimensional resistivity distribution reconstruction, corrosion spots and abnormal areas are marked, and high-precision positioning of underground conductors is completed through UWB radar or a capacitance detector; meanwhile, based on the information entropy maximization optimization frequency band combination, the grounding grid state is comprehensively evaluated in combination with a multi-parameter fusion model. The method has the non-contact patrol capability, the unmanned aerial vehicle carries equipment to scan the grounding grid, the detection efficiency and convenience are remarkably improved, and an efficient and accurate solution is provided for guaranteeing safe operation of the power grid.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of power equipment state monitoring, and particularly relates to a tower grounding grid multi-frequency composite wave electric field imaging and soil environment medium diagnosis method. BACKGROUND

[0002] The tower grounding grid is a core component of the lightning protection system of a power transmission line, and its main function is to protect the safe operation of power equipment by reducing the grounding resistance and rapidly discharging lightning current. However, in the long-term operation process, due to the complexity of the soil environment and the influence of external factors, the grounding grid is prone to corrosion, fracture and other faults, resulting in an increase in the grounding resistance, and further causing lightning trip accidents, which seriously threaten the safety of the power grid. Therefore, it is particularly important to efficiently and accurately monitor and diagnose the state of the grounding grid.

[0003] The existing technology has the following outstanding problems in the detection of the tower grounding grid:

[0004] 1. Difficulty in separating coupled effects with a single excitation mode: Traditional grounding resistance measurement methods usually use direct current or power frequency sinusoidal waves as excitation signals, which can only reflect the overall grounding performance and cannot effectively distinguish the coupled effects between soil environment characteristics and grounding body defects. This limitation makes the diagnosis results lack of pertinence and difficult to accurately assess the actual state of the grounding grid.

[0005] 2. Low positioning accuracy and reliance on manual excavation: Conventional detection methods are mainly based on surface measurement data and cannot achieve accurate positioning of underground grounding bodies. When an anomaly is detected, manual excavation is usually required for verification, which is not only inefficient but also may cause unnecessary resource waste due to misjudgment.

[0006] 3. Insufficient frequency band coverage and lack of deep information: The existing detection methods do not fully optimize the frequency and bandwidth of the excitation signal, resulting in insufficient detection capability for deep soil characteristics and minor defects. For example, low-frequency signals have strong penetration but low resolution, while high-frequency signals can capture local defects but have limited penetration depth. This insufficient frequency band coverage limits the comprehensive diagnosis of complex grounding grid structures.

[0007] 4. Lack of multi-parameter fusion analysis capability: The characteristics of the soil environment medium have an important influence on the performance of the grounding grid, but existing technologies often only focus on a single parameter and ignore the comprehensive analysis of multi-physical field data, making it difficult to build a comprehensive health assessment system.

[0008] 5. Low efficiency of mobile detection and poor adaptability: Traditional detection methods rely on fixed point measurement and lack efficient mobile detection capability, especially in large-area grounding grid monitoring. In addition, in the face of complex environmental conditions such as vegetation coverage and heavy rainfall, the adaptability of existing technologies is poor, further limiting their application range.

[0009] In view of the above technical bottleneck, a new detection method is urgently needed to meet the following needs simultaneously: separating the coupling effect of soil environment and grounding body defects, improving the pertinence and accuracy of diagnosis; realizing high-precision three-dimensional positioning of underground grounding body, reducing the dependence on manual excavation; optimizing the frequency and bandwidth of excitation signal, improving the detection ability of deep soil information and small defects; fusing multi-physical field data, constructing a comprehensive grounding net health evaluation model, and enhancing the early warning ability; improving the patrol efficiency and environmental adaptability, supporting non-contact, large-area rapid scanning. SUMMARY

[0010] In order to solve the above problems, the present application provides a tower grounding net multi-frequency composite wave electric field imaging and soil environment medium diagnosis method, which aims to solve the above technical problems by introducing multi-frequency composite wave modulation technology, multi-dimensional signal fusion algorithm and non-contact mobile patrol device, and provides an efficient, accurate and intelligent solution for tower grounding net state monitoring.

[0011] The technical scheme adopted by the present application is:

[0012] The tower grounding net multi-frequency composite wave electric field imaging and soil environment medium diagnosis method comprises the following steps:

[0013] Step 1: Excitation signal generation and injection: generate a composite signal of low-frequency sinusoidal wave and high-frequency CHIRP square wave and inject it into the grounding net, covering the overall conductivity and local defect detection requirements;

[0014] Step 2: Response signal acquisition: arrange a distributed sensor array to synchronously acquire multi-dimensional signals of voltage / current response, soil temperature and humidity, and electric field intensity;

[0015] Step 3: Signal separation and anomaly detection: use wavelet transform to extract the characteristic frequency band, calculate the frequency response deviation based on the anomaly detection model, and label the corrosion or abnormal area;

[0016] Step 4: Three-dimensional electric field reconstruction and defect positioning: reconstruct the resistivity distribution based on the multi-grid homotopy algorithm, identify the high-resistance area, and generate an electric field distribution cloud map using UWB radar or capacitance probe, and combine the transmit-receive time delay difference to locate the underground conductor;

[0017] Step 5: Soil environment medium diagnosis: use the UWB radar echo signal to invert the soil dielectric constant distribution using the finite difference time domain (FDTD) algorithm, and diagnose the soil layering characteristics in combination with temperature and humidity data;

[0018] Step 6: Mobile patrol and anomaly marking: the unmanned aerial vehicle carrying the capacitance probe or UWB radar scans along the grounding net path, and the background system marks the electric field abnormal area in real time;

[0019] Step 7 Health status assessment and early warning: build a multi-parameter fusion model to calculate the corrosion risk index, output the health level and trigger the early warning decision.

[0020] Further, in step 1, the low-frequency sinusoidal wave is 1-100 kHz, which can evaluate the overall grounding grid conductivity; the high-frequency CHIRP square wave bandwidth is 0.1-10 MHz, which can detect local defects and deep soil information; the CHIRP square wave adopts linear frequency modulation pulse CHIRP or step frequency continuous wave SFCW modulation, the pulse width range is 10 ns-1 μs, and the peak power is ≤100 W;

[0021] The low-frequency sinusoidal wave and the high-frequency CHIRP square wave form a composite wave;

[0022] The composite wave is represented as:

[0023] I(t)=A1sin(2πf low t)+A2CHIRP low (t;f low ,f square );

[0024] In the formula, I(t) represents the composite excitation signal injected into the grounding grid, which changes with time t; A1 is the amplitude of the low-frequency sinusoidal wave, which is used to control the intensity of the low-frequency signal; sin(2πf min t) represents the low-frequency sinusoidal wave signal, the frequency is f max ; f min is the frequency of the low-frequency sinusoidal wave; A2 is the amplitude of the CHIRP square wave, which is used to control the intensity of the high-frequency signal; CHIRP max (t;f min ,f max ) represents the linear frequency modulation pulse CHIRP square wave signal, whose frequency changes linearly between f min and f max ; f j is the starting frequency of the CHIRP square wave; f j is the terminal frequency of the CHIRP square wave; t is the time variable, representing the change of the signal with time.

[0025] Further, in step 2, the multi-dimensional signal includes: arranging buried electrodes on the ground surface to form a grid-like distribution, and synchronously collecting voltage / current response signals through each electrode; obtaining soil temperature and humidity through a temperature and humidity sensor; measuring the electric field intensity through a three-dimensional electric field detector; measuring the soil's voltage / current response by arranging a distributed electrode array and injecting a composite excitation signal into the soil, using electrical impedance tomography (EIT) to reconstruct the soil's internal resistivity distribution image combined with a multi-channel data acquisition system and inversion algorithm; measuring soil moisture content through a FDR sensor; measuring soil temperature through a PT100 platinum resistance;

[0026] The collected original signal is denoised, and wavelet transform is used to extract low-frequency components for soil temperature and humidity characteristics and high-frequency components for detecting defects of the grounding body.

[0027] Further, in step 3, the response signal is decomposed using a DB4 wavelet basis to extract the target frequency band, and a hard threshold value is used to extract the target frequency band The effective signal is retained;

[0028] In the formula, T represents a hard threshold value used in the denoising process of wavelet transform in signal processing; σ represents the standard deviation of noise; N is the number of sampling points; and lnN represents a natural logarithm operation; represents a scaling factor related to the number of signal sampling points;

[0029] After the response signal is decomposed by wavelet transform, it is input into an anomaly detection model:

[0030]

[0031] In the formula, ΔE is the frequency response deviation energy; f min and f max are the lower and upper limits of the frequency range; H(f) is the measured frequency response function; H0(f) is the health state reference frequency response function; and df is the frequency differential unit;

[0032] When ΔE is greater than a threshold value, it is marked as a corrosion or abnormal area.

[0033] Further, in step 4, the multi-grid homotopy algorithm divides the grounding net into multiple layers of grids, and solves the resistivity distribution layer by layer; the coarse grid layer quickly solves the smoothing error; and the fine grid layer corrects the high-frequency component and refines to millimeter-level accuracy.

[0034] The control equation of the multi-grid homotopy algorithm is:

[0035] ▽·(σ(r)▽φ)=-Iδ(r-r0);

[0036] In the formula, ▽ is a gradient operator representing the partial derivative with respect to the spatial variable; σ(r) is the conductivity of the soil or grounding net material, which is a function of position r, indicating the conductivity characteristics of different regions; φ is the potential distribution, indicating the potential field generated in the grounding net and soil due to current injection; I is the injected current intensity, indicating the current size injected into the grounding net through the excitation source; δ(r-r0) is the Dirac delta function, used to describe the position of the point source, indicating that the current injection point is at r0, and other positions are zero; r is the spatial position vector, indicating any point in three-dimensional space; r0 is the position vector of the current injection point, indicating the specific position of the excitation signal injection;

[0037] When identifying high resistance zones and locating underground conductors, first determine whether the resistivity change exceeds the threshold value:

[0038] Δρ j = ρ j - ρ ref > 50 Ω·m;

[0039] In the formula, Δρ j is the resistivity change of the jth region, indicating the difference between the current resistivity of the region and the reference value; ρ j is the measured resistivity of the jth region, indicating the resistivity value obtained by algorithm reconstruction; ρ ref is the reference resistivity of the healthy region, indicating the normal resistivity value when no corrosion or defects occur; 50 Ω·m is the threshold resistivity change, when the resistivity change exceeds this threshold, it is determined as a high resistance zone;

[0040] Thus, the high resistance zone is labeled as a corrosion point; then, UWB radar or capacitive probe is used to generate an electric field distribution cloud map, combined with the transmission-reception time delay difference to locate the underground conductor.

[0041] Further, in step 5, the soil dielectric constant distribution is inverted using the UWB radar echo signal through the finite difference time domain (FDTD) algorithm, and the objective function is optimized as:

[0042]

[0043] In the formula, ∈r is the relative dielectric constant of the soil; σ is the conductivity of the soil; S meas (t) is the measured echo signal, indicating the reflection of electromagnetic waves after propagation in the soil; S sim (t, ∈r, σ) is the simulated echo signal, which is based on the given soil dielectric constant and conductivity, and is the theoretical echo signal calculated by the electromagnetic wave propagation model; t is the time variable;

[0044] The frequency bandwidth optimization algorithm is used, and the objective function is:

[0045] max S(f) = -∑p(f)logp(f);

[0046] Where,

[0047] In the formula, S(f) represents the information entropy, which is used to measure the information richness in the frequency band; p(f) represents the normalized frequency response power distribution, i.e., the energy proportion at frequency f; H(f) represents the frequency response of the system, reflecting the response characteristics of the system at different frequencies; ∫|H(f)| 2 df represents the total energy of the frequency response power in the entire frequency band;

[0048] The layer diagnosis logic for diagnosing the layering characteristics of the soil is as follows:

[0049] Surface clay: ∈r= 15-25, σ = 0.01-0.1 s / m;

[0050] Deep sand: ∈r= 5-10, σ = 0.001-0.01 s / m.

[0051] Further, in step 6, the background system receives the real-time electric field intensity data returned by the capacitive probe or UWB radar and the dynamic background field intensity value E bg , and runs the abnormality judgment logic formula:

[0052] if E field > (E bg + 200): tag.as.anomaly;

[0053] If the condition is met, the background system will mark the area as abnormal; and bind the electric field intensity data and the GPS coordinates to generate an electric field distribution heat map through spatial interpolation;

[0054] In the formula, E field is the current measured surface electric field intensity value, which is collected in real time by the capacitive probe or UWB radar; E bg is the background field intensity value, which is the background electric field intensity dynamically collected every 10 seconds, used to calibrate environmental interference and provide a reference; 200 is the abnormality judgment threshold, which is an empirical value based on historical data statistics; and tag.as.anomaly is the marked abnormal area.

[0055] Further, in step 7, the multi-parameter fusion model input vector X = [ρ, θ, T] or X = [E d , ρ, ξ]; in the formula, ρ is the soil resistivity, θ is the soil moisture content, T is the soil temperature, E d is the electric field distortion rate, and ξ is the soil temperature and humidity.

[0056] The corrosion risk index calculation formula is:

[0057] R = α·ln(1 / ρ) + β·θ + γ·e 0.05(T-25) ;

[0058] The output is the corrosion risk level.

[0059] The health evaluation index calculation formula is:

[0060]

[0061] In the formula, R is a corrosion risk index, used to quantify the corrosion risk level of the grounding grid; HealthIndex is a health assessment index, used to quantify the overall health status of the grounding grid; alpha, beta, gamma are weight coefficients, used to balance the contribution of each parameter to the health assessment index, and are obtained by training historical failure data; e 0.05(T-25) is a simplified form based on the Arrhenius equation, describing the nonlinear influence of temperature on the reaction rate.

[0062] Further, the corrosion risk level division rule is:

[0063] Low risk: R < R low or HealthIndex < H low ;

[0064] Medium risk: R low < R < R high or H low < HealthIndex < H high ;

[0065] High risk: R > R high or HealthIndex > H high ;

[0066] In the formula, R low , R high are the low risk and high risk thresholds of the corrosion risk index; H low , H high are the low risk and high risk thresholds of the health assessment index.

[0067] The early warning decision is to set a confidence threshold, below which manual review is triggered, and the GIS platform automatically triggers an alarm, combined with the health assessment result to guide the maintenance priority ranking.

[0068] The beneficial effects of the present application are:

[0069] The present application introduces multi-frequency composite wave modulation, multi-dimensional signal fusion and non-contact patrol testing technology, solves many problems existing in traditional grounding grid detection methods, and significantly improves the detection efficiency, accuracy and intelligent level. The specific beneficial effects are as follows:

[0070] 1. Improve detection sensitivity and resolution: by superimposing low-frequency sine wave and high-frequency CHIRP square wave, synchronous detection of shallow soil properties and deep grounding body defects is realized. Low-frequency wave has strong penetration, which can comprehensively evaluate the overall conductivity of the grounding grid. High-frequency wave has high resolution and is sensitive to local small defects. High-frequency signal can detect surface micro-cracks, significantly improving the ability to identify fine corrosion or fracture points.

[0071] 2. Achieve high-precision three-dimensional positioning: The grounding grid is divided into multiple layers of grid, and refined to millimeter-level precision layer by layer, effectively solving the problem of relying on manual excavation in traditional methods. UWB radar and capacitive detector are combined to achieve accurate positioning of underground conductors through transmission-reception time delay difference or electric field distribution cloud map. With capacitive detector or UWB radar, the ground electric field intensity map is quickly generated, and the abnormal area has high coincidence rate.

[0072] 3. Multi-parameter fusion improves diagnostic accuracy: Synchronous acquisition of soil resistivity, soil moisture content, soil temperature and other multi-dimensional data to build a comprehensive health assessment model; through corrosion risk index calculation, considering the influence of soil resistivity, soil moisture content and soil temperature on grounding grid corrosion, the warning accuracy is significantly improved; through health status evaluation index, combining electric field distortion rate, soil resistivity and temperature and humidity, a comprehensive evaluation system is built to enhance the comprehensiveness and reliability of diagnosis.

[0073] 4. Improve detection efficiency and convenience: Through unmanned aerial vehicle carrying capacitive detector or UWB radar, rapid scanning along the ground net path, efficiency is greatly improved. Based on sliding window statistical background field strength, dynamically adjust the abnormal judgment threshold, improve the adaptability in complex environment. Through precise positioning and health assessment model, reduce the proportion of invalid excavation, greatly reduce the maintenance cost.

[0074] 5. Optimize frequency band selection and data quality: Optimize the optimal frequency band combination through information entropy maximization algorithm to ensure that the signal covers the most abundant information, while reducing the amount of redundant data. Low-frequency signals cover deep soil information, and high-frequency signals capture local defects to achieve multi-level and comprehensive diagnosis.

[0075] 6. Enhance environmental adaptability: Use 50Hz power frequency notch and coherent averaging technology to suppress environmental noise interference and ensure measurement accuracy. For complex conditions such as vegetation coverage and heavy rain, develop multi-band detection mode to further improve the robustness of the detection system.

[0076] In summary, the present application realizes efficient and accurate detection of the state of the tower grounding grid through multi-frequency composite wave modulation, multi-dimensional signal fusion and non-contact patrol detection technology; improves detection sensitivity and resolution; realizes high-precision three-dimensional positioning; multi-parameter fusion improves diagnostic accuracy; improves detection efficiency and convenience; optimizes frequency band selection and data quality; enhances the adaptability of complex environment; significantly reduces detection cost and improves economic benefit; the present application has important application value and popularization prospect in the field of power equipment state monitoring. BRIEF DESCRIPTION OF DRAWINGS

[0077] In order to more clearly illustrate the technical solutions in the embodiments of the present application or the prior art, the following will briefly introduce the drawings needed to be used in the embodiments or prior art description. Obviously, the drawings in the following description are some embodiments of the present application, and other drawings can also be obtained by those skilled in the art without any creative effort.

[0078] Figure 1 The whole flowchart of the tower grounding grid multi-frequency composite wave electric field imaging and soil environment medium diagnosis method of the present application is shown in the figure.

[0079] Figure 2 The data fusion presentation process diagram of the application example 3 mobile patrol positioning is shown in the figure.

[0080] Figure 3 The flowchart of the soil environment multi-parameter fusion evaluation of the application example 4 is shown in the figure. DETAILED DESCRIPTION

[0081] The technical solutions in the embodiments of the present application will be described clearly and completely below with the drawings of the embodiments of the present application. Obviously, the described embodiments are only some embodiments of the present application, but not all the embodiments. All other embodiments obtained by those skilled in the art without any creative effort based on the embodiments in the present application belong to the protection scope of the present application.

[0082] In view of the deficiencies of the traditional grounding grid detection technology in sensitivity, positioning accuracy and multi-parameter fusion, the present embodiment provides a tower grounding grid multi-frequency composite wave electric field imaging and soil environment medium diagnosis method, which has non-contact patrol detection capability, and the unmanned aerial vehicle carries the equipment to complete the grounding grid scanning, thereby significantly improving the detection efficiency and convenience, and providing an efficient and accurate solution for ensuring the safe operation of the power grid.

[0083] The realization of the tower grounding grid multi-frequency composite wave electric field imaging and soil environment medium diagnosis method is based on a tower grounding grid multi-frequency composite wave electric field imaging and soil environment medium diagnosis system, which is composed of a composite wave excitation source, a distributed sensor array, a signal decoupling algorithm and a mobile patrol device.

[0084] The excitation source generates a composite signal of a 1-100 kHz sine wave and a CHIRP modulated square wave with a bandwidth of 0.1-10 MHz. The acquisition end measures the voltage / current response through the buried electrode, synchronously collects the soil temperature and humidity through the temperature and humidity sensor, and collects the electric field intensity through the three-dimensional electric field detector. The algorithm layer separates the response signal by using the frequency domain wavelet transform, realizes the three-dimensional reconstruction of the grounding grid by combining the electrical impedance tomography (EIT) and electromagnetic inversion. The patrol device is a simple capacitive coupled electric field detector or a special antenna echo detector UWB radar carried by the unmanned aerial vehicle or the handheld device.

[0085] As Figure 1 shown, the tower grounding grid multi-frequency composite wave electric field imaging and soil environment medium diagnosis method comprises the following steps:

[0086] Step 1: Excitation signal generation and injection:

[0087] Generate a composite signal of low-frequency sinusoidal wave and high-frequency CHIRP square wave and inject it into the grounding grid to cover the overall conductivity and local defect detection requirements.

[0088] Among them, the low-frequency sinusoidal wave is 1-100 kHz, which can evaluate the overall grounding grid conductivity; the high-frequency CHIRP square wave bandwidth is 0.1-10 MHz, which can detect local defects and deep soil information; the CHIRP square wave adopts linear frequency modulation pulse CHIRP or step frequency continuous wave SFCW modulation, the pulse width range is 10 ns-1 μs, and the peak power ≤100 W.

[0089] The low-frequency sinusoidal wave and the high-frequency CHIRP square wave form a composite wave;

[0090] The composite wave is represented as:

[0091] I(t)=A1sin(2πf low t)+A2CHIRP square (t;f min ,f max );

[0092] In the formula, I(t) represents the composite excitation signal injected into the grounding grid, which changes with time t; A1 is the amplitude of the low-frequency sinusoidal wave, used to control the intensity of the low-frequency signal; sin(2πf low t) represents the low-frequency sinusoidal wave signal, the frequency is f low ; f low is the frequency of the low-frequency sinusoidal wave; A2 is the amplitude of the CHIRP square wave for controlling the intensity of the high-frequency signal; CHIRP square (t;f min ,f max ) represents the linear frequency modulation pulse CHIRP square wave signal, whose frequency changes linearly between f min and f max ; f min is the starting frequency of the CHIRP square wave; f max is the terminal frequency of the CHIRP square wave; t is the time variable, representing the change of the signal with time.

[0093] Step 2: Response signal collection:

[0094] Distributed sensor array is arranged to synchronously collect multi-dimensional signals of voltage / current response, soil temperature and humidity, and electric field intensity.

[0095] The multi-dimensional signal includes: 1. The buried electrodes are arranged on the ground surface to form a grid distribution, and the voltage / current response signals collected synchronously through each electrode. 2. The soil temperature and humidity obtained by the temperature and humidity sensor. 3. The electric field intensity measured by the three-dimensional electric field detector. 4. The soil voltage / current response is measured by using the electrical impedance imaging (EIT) by arranging a distributed electrode array and injecting a composite excitation signal into the soil, combined with a multi-channel data acquisition system and an inversion algorithm, to reconstruct the resistivity distribution image inside the soil. 5. The soil moisture content is measured by the FDR sensor. 6. The soil temperature is measured by the PT100 platinum resistance.

[0096] The collected original signals are denoised, and the low-frequency components used for inversion of soil temperature and humidity characteristics and the high-frequency components used for detection of grounding body defects are extracted by wavelet transform.

[0097] Step 3: Signal separation and anomaly detection:

[0098] The characteristic frequency band is extracted by wavelet transform, the frequency response deviation is calculated based on the anomaly detection model, and the corrosion or abnormal area is labeled.

[0099] The characteristic frequency band is extracted by wavelet transform: the response signal is decomposed using DB4 wavelet basis, the target frequency band is extracted, and the hard threshold value The effective signal is retained.

[0100] In the formula, T represents the hard threshold value used in the denoising process of wavelet transform in signal processing; σ represents the standard deviation of noise; N is the number of sampling points; lnN represents the natural logarithm operation; represents a scaling factor related to the number of signal sampling points.

[0101] After the response signal is decomposed by wavelet transform, it is input into the anomaly detection model:

[0102]

[0103] In the formula, ΔE is the frequency response deviation energy; f min and f max are the lower and upper limits of the frequency range; H(f) is the measured frequency response function; H0(f) is the healthy state reference frequency response function; df is the frequency differential unit;

[0104] When ΔE> threshold value, it is labeled as a corrosion or abnormal area.

[0105] Step 4: Three-dimensional electric field reconstruction and defect positioning:

[0106] The resistivity distribution is reconstructed based on the multi-grid homotopy algorithm, the high-resistance area is identified, and the electric field distribution cloud map is generated by using the UWB radar or capacitive detector, combined with the transmission-reception time delay difference to locate the underground conductor.

[0107] The multi-grid homotopy algorithm divides the grounding grid into multiple layers of grids, and solves the resistivity distribution layer by layer; the coarse grid layer quickly solves the smoothing error; the fine grid layer corrects the high-frequency component and refines to millimeter-level accuracy.

[0108] The control equation of the multi-grid homotopy algorithm is:

[0109] ▽·(σ(r)▽φ)=-Iδ(r-r0);

[0110] In the formula, ▽ is the gradient operator, indicating the partial derivative with respect to the spatial variable; σ(r) is the electrical conductivity of the soil or grounding grid material, which is a function of the position r, indicating the conductivity characteristics of different regions; φ is the potential distribution, indicating the potential field generated in the grounding grid and soil due to current injection; I is the injected current intensity, indicating the current size injected into the grounding grid through the excitation source; δ(r-r0) is the Dirac delta function, used to describe the position of the point source, indicating that the current injection point is located at r0, and other positions are zero; r is the spatial position vector, indicating any point in three-dimensional space; r0 is the position vector of the current injection point, indicating the specific position of the excitation signal injection;

[0111] When identifying high-resistance areas and locating underground conductors, first determine whether the resistivity change exceeds the threshold value:

[0112] Δρ j =ρ j -ρ ref >50Ω·m;

[0113] In the formula, Δρ j is the resistivity change of the jth region, indicating the difference between the current resistivity and the reference value of the region; ρ j is the measured resistivity of the jth region, indicating the resistivity value obtained by algorithm reconstruction; ρ ref is the reference resistivity of the healthy region, indicating the normal resistivity value when no corrosion or defect occurs; 50Ω·m is the threshold resistivity change, when the resistivity change exceeds this threshold, it is determined as a high-resistance area.

[0114] Thus, the high-resistance area is labeled as a corrosion point; then, the UWB radar or capacitive probe is used to generate an electric field distribution cloud map, and the transmit-receive time delay difference is used to locate the underground conductor.

[0115] Step 5: Soil environmental medium diagnosis:

[0116] Through the UWB radar echo signal, the finite difference time domain (FDTD) algorithm is used to invert the soil dielectric constant distribution, and combined with temperature and humidity data to diagnose the soil layering characteristics.

[0117] Specifically, the soil dielectric constant distribution is inverted by the time-domain finite difference FDTD algorithm using the UWB radar echo signal, and the objective function is optimized as:

[0118]

[0119] In the formula, ∈r is the relative dielectric constant of the soil; σ is the conductivity of the soil; S meas (t) is the measured echo signal, indicating the reflection of electromagnetic waves after propagation in the soil; S sim (t, ∈r, σ) is the simulated echo signal, which is a theoretical echo signal calculated by the electromagnetic wave propagation model based on the given soil dielectric constant and conductivity; t is the time variable;

[0120] The frequency bandwidth optimization algorithm is adopted, and the objective function is:

[0121] maxS(f)=-∑p(f)logp(f);

[0122] Among them,

[0123] In the formula, S(f) represents the information entropy, which is used to measure the information richness in the frequency band; p(f) represents the normalized frequency response power distribution, that is, the energy proportion at frequency f; H(f) represents the frequency response of the system, reflecting the response characteristics of the system at different frequencies; ∫|H(f)| 2 df represents the total energy of the frequency response power in the entire frequency band range;

[0124] The layering diagnosis logic of soil layering characteristics diagnosis is as follows:

[0125] Surface clay: ∈r=15-25, σ=0.01-0.1s / m;

[0126] Deep sandstone: ∈r=5-10, σ=0.001-0.01s / m.

[0127] Step 6: Move the patrol and mark the anomaly:

[0128] The unmanned aerial vehicle carrying the capacitive probe or UWB radar scans along the ground net path, and the background system real-time marks the electric field anomaly area.

[0129] The background system receives the real-time electric field intensity data and dynamic background field intensity value E bg returned by the capacitive probe or UWB radar, and runs the anomaly judgment logic formula:

[0130] if E field >(E bg +200):tag.as.anomaly;

[0131] If the condition is met, the background system will mark the area as abnormal; and the electric field intensity data is bound with GPS coordinates, and the electric field distribution heat map is generated by spatial interpolation;

[0132] In the formula, E field is the current measured ground electric field intensity value, which is collected in real time by the capacitance detector or UWB radar; E bg is the background field intensity value, which is dynamically collected every 10 seconds, and is used to calibrate environmental interference and provide a reference; 200 is the abnormality determination threshold, which is an empirical value based on historical data statistics; tag.as.anomaly is the marked abnormal area.

[0133] Step 7 Health status evaluation and early warning:

[0134] A multi-parameter fusion model is constructed to calculate the corrosion risk index, output the health level and trigger the early warning decision.

[0135] The input vector of the multi-parameter fusion model is X=[p, q, T] or X=[E d , p, q]; in the formula, p is the soil resistivity, q is the soil moisture content, T is the soil temperature, E d is the electric field distortion rate, and q is the soil temperature and humidity.

[0136] The corrosion risk index calculation formula is:

[0137] R=a·ln(1 / p)+b·q+g·e 0.05(T-25) ;

[0138] The output is the corrosion risk level.

[0139] The health evaluation index calculation formula is:

[0140]

[0141] In the formula, R is the corrosion risk index, which is used to quantify the corrosion risk level of the grounding grid; HealthIndex is the health evaluation index, which is used to quantify the overall health status of the grounding grid; a, b, g are weight coefficients, which are used to balance the contribution of each parameter to the health evaluation index, and are obtained by training historical failure data; e 0.05(T-25) is a simplified form based on the Arrhenius equation, which describes the nonlinear effect of temperature on the reaction rate.

[0142] The corrosion risk level division rule is:

[0143] Low risk: R low or HealthIndex low ;

[0144] Medium risk: Rlow <R < R high or H low < HealthIndex < H high ;

[0145] High risk: R > R high or HealthIndex > H high ;

[0146] In the formula, R low , R high are the low-risk and high-risk thresholds of the corrosion risk index; H low , H high are the low-risk and high-risk thresholds of the health assessment index.

[0147] The early warning decision is to set a confidence threshold, below which manual review is triggered, and the GIS platform is automatically triggered to alarm, combined with the health assessment result to guide the maintenance priority ranking.

[0148] Application Example 1, composite wave modulation and electric field imaging:

[0149] In order to facilitate understanding of steps 1 to 4 of the method of multi-frequency composite wave electric field imaging and soil environment medium diagnosis of the tower grounding net, this application example realizes three-dimensional reconstruction and defect positioning of the resistivity distribution of the grounding net by injecting a composite signal of a 50 kHz low-frequency sine wave and a high-frequency CHIRP square wave with a bandwidth of 1-5 MHz into the grounding net, combined with multi-electrode arrangement and signal processing technology. The detection results show that this method can efficiently identify corrosion points and accurately locate the fracture area, with an error of ≤0.5 m, significantly improving the detection efficiency and accuracy.

[0150] The detailed steps and algorithm analysis are as follows:

[0151] Step 1, composite wave injection:

[0152] Waveform composition:

[0153] I(t) = A1sin(2π·50×10 3 t) + A2CHIRP square (t; f min = 1 MHz, f max = 5 MHz);

[0154] 50 kHz low-frequency sine wave: it has strong penetration and is used to evaluate the overall grounding net conductivity.

[0155] 1-5 MHz high-frequency CHIRP square wave: it is sensitive to local defects, especially suitable for detecting millimeter-level cracks, with a skin depth δ ≈ 0.1 mm.

[0156] Parameter definition: A1, A2 are current amplitude, typical value: 1-10A; CHIRP sweep rate: 20μs / Hz, ensure cover target frequency band.

[0157] The composite wave signal has both the global characteristics of low frequency and the local sensitivity of high frequency, meeting the synchronous detection needs of shallow soil characteristics and deep grounding body defects.

[0158] Step 2, signal acquisition and processing:

[0159] Electrode arrangement: 12 buried electrodes are arranged on the ground in a grid shape, uniformly distributed, with a spacing of ≤5m, and the voltage response V is synchronously collected i (t).

[0160] Wavelet denoising process: DB4 wavelet basis is used to decompose the signal.

[0161] Threshold processing: hard threshold

[0162] Reconstruct the signal, and retain the characteristic frequency band (50kHz±5%, 3MHz±10%).

[0163] Feature extraction: calculate the impedance amplitude and phase of each electrode at the target frequency point:

[0164]

[0165] In the formula, Z(f) is the impedance, indicating the voltage-to-current ratio at frequency f, reflecting the conductive characteristics of the grounding grid or soil at a specific frequency; F{V(t)} is the Fourier transform of the voltage signal V(t), which converts the time-domain signal V(t) into a frequency-domain signal, extracting the voltage amplitude and phase information at the target frequency f; F{I(t)} is the Fourier transform of the current signal I(t), which converts the time-domain signal I(t) into a frequency-domain signal, extracting the current amplitude and phase information at the target frequency f; V(t) is the voltage response signal, a function of time, which is collected by the buried electrode voltage response data; I(t) is the excitation current signal injected into the grounding grid, a function of time, which is generated by the composite wave signal source and injected into the grounding grid; f is the target frequency point, i.e. the specific frequency to be analyzed; F is the Fourier transform operation, which converts the time-domain signal into the frequency-domain signal, facilitating the analysis of signal characteristics at different frequencies.

[0166] Step 3, signal separation and anomaly detection:

[0167] Wavelet transform to extract characteristic frequency band: use DB4 wavelet basis to decompose the response signal, extract target frequency band 5kHz, 3MHz. Hard threshold processing: retain valid signals.

[0168] Anomaly detection model: based on frequency response deviation calculation:

[0169]

[0170] When ΔE > threshold, mark as corrosion or abnormal area.

[0171] Step 4, resistivity imaging and defect localization:

[0172] The control equation is:

[0173] ▽·(σ(r)▽φ)=-Iδ(r-r0);

[0174] The grounding grid is divided into multiple layers of grids, coarse grids are used to quickly solve the smooth error, and fine grids are used to correct the high-frequency components, and the accuracy is refined to the millimeter level.

[0175] The iteration convergence condition is: |b-Ax (k) |2<10 -6 ; where b is the target vector, representing the known quantity on the right side of the linear equation group, representing the potential field distribution generated by the excitation current injection point, i.e. observation data or measurement value; A is the coefficient matrix, representing the discretization control equation of the system, which is the coefficient matrix of the linear equation group generated by the soil or grounding grid material conductivity σ(r) and spatial grid division, describing the relationship between the potential field and the resistivity distribution; x (k) is the solution vector of the kth iteration, representing the resistivity distribution vector calculated by the kth iteration; Ax (k) represents the predicted potential field distribution according to the current resistivity distribution; |b-Ax (k) |2is the L2of the residual vector, representing the difference between the predicted potential field of the current resistivity distribution and the actual observed potential field; 10 -6 is the convergence threshold, representing the maximum error allowed.

[0176] The corrosion point criterion is: Δρ j =ρ j -ρ ref > 50Ω·m;

[0177] The defect localization uses UWB radar or capacitive probe to generate an electric field distribution cloud map, combined with the transmission-reception time delay difference to locate the underground conductor; the positioning error is ≤0.5m.

[0178] Composite wave modulation advantages: low frequency 50kHz, providing global characteristics; high frequency 3MHz, sensitive to millimeter-level cracks, skin depth δ≈0.1mm. Multiple grid algorithm efficiency: computational complexity O(N), N is the number of grids, traditional finite element method is O(N 2Wavelet adaptive denoising: better than Fourier filtering, preserving transient characteristics such as reflection peaks caused by corrosion points; because of the use of complex modulation wave, the use of low-frequency penetration, the signal has the characteristics of short-time jump quality, and the use of wavelet transform can further improve the signal-to-noise ratio.

[0179] Compared with traditional excavation detection, the advantages are as follows:

[0180] Indicator This program Traditional excavation detection Detection speed 2 hours / site 3-5 days / site Positioning accuracy ≤0.5m Depend on experience, error ≥ 2m Cost composition Equipment reuse (≥ 100 station times) Manual + mechanical ($5000 / station) Power loss No need to stop power Average 8 hours / station Implicit benefits Preventive maintenance, reduce failure rate Only after repair

[0181] In addition, in order to facilitate understanding, the following formula derivation is given:

[0182] 1. The relationship between skin depth and frequency:

[0183]

[0184] Where δ is the skin depth, which represents the penetration depth when the amplitude of electromagnetic wave or current propagating in a conductor is attenuated to 1 / e, about 37%; ρ is the resistivity, which represents the ability of a material to resist current flow, reflecting the electrical conductivity of the material; f is the frequency, which represents the frequency of the excitation signal, i.e. the number of periodic changes per second; μ is the magnetic permeability, which represents the response ability of the material to the magnetic field.

[0185] When f = 3 MHz, ρ = 10 -7 Ω·m, the skin depth δ ≈ 0.1 mm.

[0186] 2. Proof of convergence of multigrid:

[0187]

[0188] Where M represents the iteration matrix, which is usually constructed from the discrete equation Ax = b in the multigrid algorithm; ||M||2 represents the spectral norm of matrix M, i.e. the maximum singular value of the matrix; γ represents a constant between 0 and 1; e (k) k represents the error vector after the kth iteration; e (0) 0 represents the initial error vector; γ k represents the exponential decay factor of the error with the number of iterations k.

[0189] In practical applications, γ = 0.3 is taken, and 5 iterations can converge.

[0190] This application example realizes high-precision reconstruction and defect positioning of the resistivity distribution of the grounding grid through composite wave modulation, multi-electrode arrangement and multigrid algorithm. Two broken points are detected, with a positioning error ≤ 0.5 m, verifying the efficiency and reliability of the method. At the same time, through wavelet transform and frequency response deviation model, the anti-interference performance and diagnostic accuracy are improved, reflecting the technical advantages of the combination of composite wave penetration and wavelet transform.

[0191] Application Example 2, CHIRP scanning and electrical impedance imaging:

[0192] In order to facilitate the understanding of the steps of the tower grounding grid multi-frequency composite wave electric field imaging and soil environment medium diagnosis method, this application example transmits a CHIRP square wave signal with a high bandwidth ≥1GHz, combines with the UWB radar received echo signal, uses the finite difference time domain FDTD algorithm to invert the soil dielectric constant distribution, and diagnoses the soil layering characteristics combined with temperature and humidity data. The experimental results show that this method can realize soil layering accuracy >90%, grounding body burial depth positioning error ≤0.3m, and significantly improves the resolution and accuracy of non-contact detection.

[0193] The detailed steps and algorithm analysis are as follows:

[0194] Step 1, CHIRP square wave signal transmission:

[0195] Signal parameters:

[0196] The pulse width τ is 100ns, which defines the duration of a single pulse; improve the range resolution by short pulse width

[0197] In the formula, Δd is the range resolution, which represents the minimum distance that the radar system can distinguish between two targets; c is the speed of light, which is the propagation speed of electromagnetic waves in a vacuum; τ is the pulse width, which is the duration of a single pulse in the CHIRP signal; ∈r is the relative dielectric constant of the soil.

[0198] Center frequency f c is 2.4GHz, which defines the frequency center point of the CHIRP signal; its role is to penetrate the soil depth of about 0.5-2m, with an attenuation coefficient

[0199] In the formula, α is the attenuation coefficient, which represents the degree of attenuation of electromagnetic waves in the medium; f is the frequency of electromagnetic waves; σ is the conductivity of the medium.

[0200] The typical value of the bandwidth B is ≥1GHz; the derivation process is: resolution When B=1.5GHz and ∈r=10, the theoretical resolution Δd≈0.05m, and the actual error is expanded to 0.3m due to environmental noise.

[0201] Step 2: FDTD inversion of soil dielectric constant

[0202] FDTD inversion principle: update the electric field E and magnetic field H based on Maxwell's equations, preset the conductivity σ, and optimize the solution of the dielectric constant ∈r.

[0203] Control equation:

[0204] In the formula, ∇× is the curl operator, indicating the rotation characteristics of the vector field, describing the rotation intensity of the field in space; E is the electric field intensity, describing the distribution and intensity of the electric field in space; H is the magnetic field intensity, describing the distribution and intensity of the magnetic field in space; μ is the magnetic permeability, describing the response ability of the material to the magnetic field, ∈ is the dielectric constant, describing the response ability of the material to the electric field; σ is the electrical conductivity, describing the electrical conductivity of the material, reflecting the conduction performance of the current in the material; is the time derivative, indicating the rate of change of the field with time; t is the time, indicating the time variable, used to describe the dynamic change of the field with time.

[0205] Inversion algorithm objective function: comparison of measured echo signal S meas (t) and simulated echo signal S sim (t, ∈r, σ), optimize the objective function:

[0206] Step 3, temperature and humidity correction and soil layering diagnosis:

[0207] Dielectric model correction: Topp formula is used to correct the dielectric constant: ∈r = a + b·θ + c·θ 2 +d·θ 3 ;

[0208] In the formula, parameters a, b, c, d are calibrated by soil type clay / sandstone; θ is the volumetric water content.

[0209] The layering diagnosis logic is as follows:

[0210] Soil type Dielectric constant range Conductivity range Surface clay r ≈15-25]]> ​ σ≈0.01-0.1 S / m Deep sand r ​​ σ≈0.001-0.01 S / m

[0211] Determine the interface depth by the ∈r(z) profile mutation point.

[0212] The main error source analysis is as follows:

[0213] Error type Influence Suppression method Time delay measurement error Insufficient sampling rate Use ≥20GSPS ADC Dielectric model deviation Inaccurate temperature and humidity measurement Multi-point temperature and humidity sensor fusion Multiple reflections Interlayer multiple reflections Wiener filter deconvolution

[0214] In the above process: high resolution: GHz level bandwidth to realize centimeter level resolution; multi-physical field fusion: combine electromagnetic characteristics and temperature and humidity data to improve model robustness; non-destructive detection: use UWB radar antenna reflection wave technology to avoid the destruction of traditional drilling to the soil structure.

[0215] The key parameters are summarized as follows:

[0216] Parameter Value / Range Physical meaning Pulse width (τ) 100 ns Distance resolution control center frequency (f c )]]> 2.4 GHz Penetration depth and attenuation balance Bandwidth (B) ≥1 GHz Theoretical resolution Δd∝1 / B dielectric constant (ε r )]]> Clay: 15-25, sand: 5-10 Layering determination basis Temperature and humidity correction Real-time sensor data Reduce model error

[0217] Application scenarios expansion: For power systems: corrosion diagnosis of tower grounding grids and substation grounding grids, indication of corrosion due to changes in burial depth. For agriculture: soil moisture stratification monitoring to provide a basis for precision irrigation. For archaeology: non-destructive exploration of underground relics to avoid damage to the site caused by traditional drilling.

[0218] This embodiment achieves accurate inversion of soil dielectric constant distribution by transmitting high-bandwidth CHIRP square wave signals, combining UWB radar and FDTD algorithm, and significantly improves soil stratification accuracy to >90% and grounding body burial depth positioning accuracy with an error ≤0.3m through temperature and humidity correction and stratification diagnosis logic. This method has technical advantages such as non-contact, high resolution, and multi-physical field fusion, and can be widely applied in power systems, agriculture, and archaeology, etc., demonstrating its efficiency and practicality.

[0219] Application Example 3, mobile patrol positioning:

[0220] To facilitate understanding of Step 6 of the tower grounding grid multi-frequency composite wave electric field imaging and soil environment medium diagnosis method, this application example uses a drone carrying a 20cm plate spacing capacitance probe or UWB radar to fly along the grounding grid path, real-time collects ground electric field intensity data, and combines GPS coordinate labeling of abnormal areas. The background system generates an electric field distribution heat map through dynamic threshold algorithm and spatial interpolation technology, significantly improving detection efficiency and accuracy. Experimental results show that 100m of grounding grid scanning can be completed within 10 minutes, and the abnormal area coincidence rate is 95%, verifying the efficiency and reliability of the method.

[0221] Detailed steps and algorithm analysis are as follows:

[0222] Step 1, UAV patrol system setup:

[0223] The system components are as follows:

[0224]

[0225] The plate spacing d is 20cm, which determines the detection sensitivity and spatial resolution. A 20cm preset 1.5m defect detection depth is provided. The abnormal threshold is 200V / m, which is based on historical data statistics and is the critical electric field intensity for grounding grid damage. The scanning speed is 0.17m / s, ensuring a sampling point spacing ≤10cm to ensure data coverage density.

[0226] Step 2, real-time analysis and anomaly labeling:

[0227] For electric field intensity measurement, the probe outputs an original voltage signal V, which is converted to electric field intensity through a calibration formula: In the formula, E is the electric field intensity, and k is the sensor coefficient.

[0228] Dynamic baseline calibration: background field intensity value E is collected every 10 seconds bg , eliminating environmental interference.

[0229] Abnormality determination logic: the determination condition is:

[0230] if E field > (E bg + 200): tag.as.anomaly (GPS_coord);

[0231] 200V / m is the abnormality determination threshold, which is obtained based on historical data statistics; the background system marks the abnormal area as Anomaly and generates an electric field distribution heat map.

[0232] Dynamic threshold algorithm: based on sliding window statistics background field intensity: E th = μ bg + 3σ bg + 200;

[0233] In the formula, μ bg is the background mean; σ bg is the background standard deviation.

[0234] Regarding the definition of key parameters, the following explanations are given:

[0235]

[0236]

[0237] Step 3, error processing mechanism:

[0238] For spatial positioning error: take ±5cm GPS drift and UAV attitude fluctuation.

[0239] The compensation formula is: Δ x = Δ GPS + v·tan(θ)·Δt;

[0240] In the formula, v is the speed of the UAV; θ is the pitch angle of the UAV; Δt is the sampling interval.

[0241] Measure: IMU inertial navigation fusion positioning, position error ≤8cm

[0242] For electric field measurement error, as shown in the following table:

[0243] Error source Correction method Temperature drift Polynomial compensation model: E cal = E / (1 + a Delta T) Electromagnetic interference Digital band-pass filter (50-1000 Hz) Height change Height sensor real-time correction of detection gain

[0244] In the above process: high efficiency: 100m grounding net scanning only needs 3 minutes, traditional manual detection needs 30 minutes-1 hour; high precision: abnormal area coincidence rate reaches 95%, compared with excavation verification results; high sensitivity: can detect ≥5cm2 Conductor fracture defects; Dynamic threshold algorithm: based on sliding window statistics background field strength, adapt to complex environment; Three-dimensional path planning: preset flight height 1.5 m, that is, the optimal detection height, automatic obstacle avoidance radar to ensure safety distance.

[0245] The data fusion presentation process is shown in Figure 2 , which can be improved by further perfecting the regionalized electric field annotation and combining spatial interpolation to improve image resolution.

[0246] Application limitations and improvement directions:

[0247] Current limitations: detection accuracy decreases when vegetation coverage is > 30 cm; cannot operate in heavy rain.

[0248] Optimization direction: develop multi-band detection mode 1-1000 kHz to enhance adaptability in complex environments.

[0249] This application example realizes large-area ground surface electric field strength scanning with the help of a UAV, and the core lies in regionalized electric field analysis and dynamic anomaly annotation. By mobile scanning to obtain electric field differences and combining spatial interpolation technology to generate electric field gradient distribution map, the detection efficiency and accuracy are significantly improved. This method is suitable for wideband detection under composite wave excitation, can reflect various micro-defects, and has high sensitivity and practicality. Although there are limitations of vegetation coverage and weather, through optimization of multi-band detection mode, its application range can be further expanded.

[0250] Application Example 4: Soil Environment Multi-Parameter Fusion Evaluation

[0251] To facilitate understanding of the steps of the tower grounding grid multi-frequency composite wave electric field imaging and soil environment medium diagnosis method 7; as shown in Figure 3 , this application example synchronously collects multi-dimensional data such as soil resistivity, soil moisture content, and soil temperature, constructs a health evaluation model, calculates the corrosion risk index and outputs the health grade, and triggers the early warning decision. The experimental results show that the early warning accuracy of this method reaches 88.7%, the invalid excavation rate is reduced by 90%, and the efficiency and accuracy of grounding grid state monitoring are significantly improved.

[0252] Detailed steps and algorithm analysis are as follows:

[0253] Step 1: Synchronous acquisition of multiple parameters

[0254] Parameter definition and measurement principle are as follows:

[0255]

[0256] Acquisition method:

[0257] A GPS timing module is used with an accuracy of ±1 ms to ensure simultaneous collection of three parameters at the same location.

[0258] For spatial heterogeneity, high-density EIT scanning points are arranged, ≥5 points / m 2 , combined with FDR and PT100 array distribution, to improve data coverage.

[0259] Step 2, health assessment model construction:

[0260] Model input and output:

[0261] Input vector: X = [p, q, T]; Output: Corrosion risk level: low / medium / high.

[0262] Classification logic is:

[0263] Written by python: Based on ternary decision boundary to divide risk level:

[0264] if p > 50 ohm and q < 15%: risk = "low" elif 20 <= p <= 50

[0265] ohm and 15% <= q <= 30%: risk = "medium" else: risk = "high".

[0266] Formula derivation:

[0267] Corrosion risk index R calculation formula: R = a * ln (1 / p) + b * q + g * e 0.05(T-25) ;

[0268] Risk level is divided according to R value segmentation:

[0269] Low risk: R < R low ; Medium risk: R low < R < R high ; High risk: R > R high .

[0270] Comprehensive quantification of the overall health status of grounding grid:

[0271] Step 3, error processing mechanism:

[0272] Error source Processing scheme Sensor noise Kalman filter real-time noise reduction Spatial heterogeneity EIT multilayer scanning + FDR / PT100 array plot (density ≥ 5 points / m 2 ) Temperature drift PT100 dynamic temperature compensation algorithm Model misjudgment Confidence threshold setting (<85% confidence triggers manual review)

[0273] Step 4, early warning decision:

[0274] Dynamic early warning rule: R > R high or health assessment index HealthIndex > H high ; Automatic trigger GIS platform alarm.

[0275] The maintenance priority calculation formula is:

[0276] Long-term trend analysis: Establish soil parameter degradation model to predict the corrosion evolution trend in the next 5 years.

[0277] In the above process: Multi-source data space-time synchronization: Adopt GPS timing module with an accuracy of ±1ms to ensure the simultaneous collection of soil resistivity, soil moisture content, soil temperature and other parameters. Data fusion framework: Adopt data fusion formula: In the formula, Data fusion is the data fusion result, EIT 3D is the three-dimensional electrical impedance tomography data, FDR avg is the average moisture content measured by the frequency domain reflectometry sensor, T grad is the temperature gradient or temperature change trend, indicates wavelet transform coupling. Machine learning model optimization: Introduce high-frequency and low-frequency related parameters to deeply analyze multi-dimensional data through compound wave modulation, and improve the comprehensiveness and accuracy of the health assessment model.

[0278] The technical and economic advantages are as follows:

[0279] Indicator Traditional method This program Promotion effect Early warning accuracy 62-75% (single parameter) 88.7％ +25％ Invalid excavation rate 45-60％ 4.5％ Reduced by 90% Single-point detection cost $1200 (manual excavation) $200 (sensor) Cost down 83% Detection efficiency 2 man-days / 100m Real-time continuous monitoring Efficiency improved 100 times

[0280] Application scenario expansion:

[0281] Dynamic early warning for high-risk areas: When R>R high , automatically trigger GIS platform alarm.

[0282] Maintenance strategy optimization: According to the maintenance priority formula to guide resource allocation.

[0283] Long-term trend analysis: Establish soil parameter degradation model to predict the corrosion evolution trend in the next 5 years, and take preventive measures in advance.

[0284] This application example synchronously collects multi-dimensional data: soil resistivity, soil moisture content, soil temperature, etc., and realizes comprehensive monitoring and accurate early warning of grounding grid state by combining health assessment model. Its core lies in multi-source data space-time synchronization, data fusion framework and machine learning model optimization, which significantly improves the early warning accuracy to 88.7% and detection efficiency, and reduces the invalid excavation rate by 90%. This method is not only suitable for power system grounding grid monitoring, but also can be extended to agricultural soil condition monitoring, underground facility protection and other fields, and has wide application prospects.

[0285] The above-described embodiments are only used to illustrate the technical solutions of the present application, and are not intended to limit the present application; although the present application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that: it can still modify the technical solutions recorded in the foregoing embodiments, or make equivalent replacement for part of the technical features; and these modifications or replacements do not make the essence of the corresponding technical solutions deviate from the spirit and scope of the technical solutions of the embodiments of the present application, and should be included in the protection scope of the present application.

Claims

1. A method for tower grounding grid multi-frequency composite wave electric field imaging and soil environmental medium diagnosis, characterized in that, The method comprises the following steps: Step 1: Excitation signal generation and injection: generate a composite signal of low-frequency sinusoidal wave and high-frequency CHIRP square wave and inject it into the grounding grid to cover the overall conductivity and local defect detection requirements; Step 2: Response signal acquisition: arrange a distributed sensor array to synchronously acquire multi-dimensional signals of voltage / current response, soil temperature and humidity, and electric field intensity; Step 3: Signal separation and anomaly detection: use wavelet transform to extract characteristic frequency bands, calculate frequency response deviation based on an anomaly detection model, and label corrosion or abnormal areas; Step 4: Three-dimensional electric field reconstruction and defect positioning: reconstruct the resistivity distribution based on the multi-grid homotopy algorithm, identify high-resistance areas, and generate an electric field distribution cloud map using UWB radar or a capacitance probe, combined with the transmission-reception time delay difference to locate underground conductors; Step 5: Soil environment medium diagnosis: use the time-domain finite difference FDTD algorithm to invert the soil dielectric constant distribution based on the UWB radar echo signal, and diagnose the soil layering characteristics in combination with temperature and humidity data; Step 6: Mobile patrol and anomaly marking: a drone equipped with a capacitance probe or UWB radar scans along the path of the grounding grid, and the background system labels the electric field abnormal area in real time; Step 7: Health status evaluation and early warning: build a multi-parameter fusion model to calculate the corrosion risk index, output the health level, and trigger the early warning decision.

2. The method according to claim 1, wherein the method is characterized in that: In step 1, the low-frequency sinusoidal wave is 1-100 kHz, which can evaluate the overall grounding grid conductivity; the high-frequency CHIRP square wave has a bandwidth of 0.1-10 MHz, which can detect local defects and deep soil information; the CHIRP square wave adopts linear frequency modulation pulse CHIRP or step frequency continuous wave SFCW modulation, with a pulse width range of 10 ns-1 μs and a peak power ≤100 W; The low-frequency sinusoidal wave and the high-frequency CHIRP square wave form a composite wave; The composite wave is represented as: I(t) = A1sin(2πf low t) + A2CHIRP square (t; f min ,f max ) where I(t) represents the complex excitation signal injected into the grounding grid, which varies with time t; A1 is the amplitude of the low-frequency sinusoidal wave, used to control the strength of the low-frequency signal; sin(2πf low t) represents the low-frequency sinusoidal wave signal, with a frequency of f low ; f low is the frequency of the low-frequency sinusoidal wave; A2 is the amplitude of the CHIRP square wave, used to control the strength of the high-frequency signal; CHIRP square (t; f min , f max ) represents the linear frequency modulation pulse CHIRP square wave signal, with a linearly varying frequency between f min and f max ; f min is the starting frequency of the CHIRP square wave; f max is the terminal frequency of the CHIRP square wave; t is the time variable, representing the change of the signal with time.

3. The method according to claim 1, wherein the method is characterized by: In step 2, the multi-dimensional signals include: arranging buried electrodes on the ground to form a grid-like distribution, acquiring voltage / current response signals synchronously through each electrode; acquiring soil temperature and humidity through temperature and humidity sensors; measuring electric field intensity through a three-dimensional electric field probe; measuring soil voltage / current response using electrical impedance tomography EIT by arranging a distributed electrode array and injecting a composite excitation signal into the soil, combining a multi-channel data acquisition system and an inversion algorithm to reconstruct the soil internal resistivity distribution image; measuring soil water content through FDR sensors; measuring soil temperature through PT100 platinum resistance; The collected original signals are denoised, and wavelet transform is used to extract low-frequency components for soil temperature and humidity characteristics and high-frequency components for grounding body defect detection.

4. The method according to claim 1, wherein the method is characterized in that: In step 3, the response signal is decomposed using a DB4 wavelet base, the target frequency band is extracted, and a hard threshold is used The effective signal is reserved; In the formula, T represents a hard threshold value used in a wavelet transform denoising process in signal processing; σ represents a standard deviation of noise; N is a number of sampling points; lnN represents a natural logarithm operation; represents a scaling factor related to the number of signal sampling points; After the response signal is decomposed by wavelet transform, it is input into the anomaly detection model: wherein ΔE is the frequency response deviation energy; f min and f max are the lower and upper limits of the frequency range; H(f) is the measured frequency response function; H0(f) is the baseline frequency response function for the healthy state; and df is the frequency differential unit. When ΔE is greater than the threshold value, it is labeled as a corrosion or abnormal area.

5. The method according to claim 1, wherein the method is characterized by: In step 4, the multi-grid homotopy algorithm divides the grounding grid into multiple layers of grids, and solves the resistivity distribution layer by layer; the coarse grid layer quickly solves the smoothing error; the fine grid layer corrects the high-frequency component and refines to millimeter-level accuracy; The control equation of the multi-grid homotopy algorithm is: ​ Where ▽ is the gradient operator, representing the partial derivative of the spatial variable; σ(r) is the conductivity of the soil or grounding grid material, which is a function of position r, representing the conductive characteristics of different regions; φ is the potential distribution, representing the potential field generated in the grounding grid and soil due to current injection; I is the injected current intensity, representing the size of the current injected into the grounding grid by the excitation source; δ(r-r0) is the Dirac delta function, used to describe the position of the point source, indicating that the current injection point is at r0, and other positions are zero; r is the spatial position vector, representing any point in three-dimensional space; r0 is the position vector of the current injection point, representing the specific position of the excitation signal injection; When identifying high resistance areas and locating underground conductors, first determine whether the resistivity change exceeds the threshold: Δρ j = p j - p ref > 50 Ω-m; wherein Δρ j is the resistivity variation of the jth region, representing the difference between the current resistivity of the region and the reference value; ρ j is the measured resistivity of the jth region, representing the resistivity value obtained by algorithm reconstruction; ρ ref is the reference resistivity of the healthy region, representing the normal resistivity value when no corrosion or defect occurs; 50 Ω·m is the threshold resistivity variation, when the resistivity variation exceeds this threshold, it is determined as a high resistance region; Thus, the high resistance area is labeled as a corrosion point; Then, use UWB radar or capacitive probes to generate an electric field distribution cloud map, combined with the transmit-receive time delay difference to locate underground conductors.

6. The method according to claim 1, wherein the method is characterized by: In step 5, the UWB radar echo signal is used to invert the soil dielectric constant distribution through the finite difference time domain (FDTD) algorithm, and the objective function is optimized as: where ∈ris the relative dielectric constant of the soil; σ is the electrical conductivity of the soil; S meas (t) is the measured echo signal, representing the reflection of the electromagnetic wave after propagating in the soil;S sim (t, ∈r, σ) is the simulated echo signal, which is the theoretical echo signal calculated by the electromagnetic wave propagation model based on the given dielectric constant and electrical conductivity of the soil; t is the time variable; The frequency bandwidth optimization algorithm is used, and the objective function is: max S(f) = -∑p(f) log p(f); wherein, In the formula, S(f) represents the information entropy, which is used to measure the information richness in the frequency band; p(f) represents the normalized frequency response power distribution, that is, the proportion of energy at frequency f; H(f) represents the frequency response of the system, reflecting the response characteristics of the system at different frequencies; ∫|H(f)|2df represents the total energy of the frequency response power in the entire frequency band range; and f represents the frequency. 2 df represents the total energy of the frequency response power in the entire frequency band range; and f represents the frequency. The layering diagnosis logic for soil layering characteristics diagnosis is: Surface clay: ∈r=15-25, σ=0.01-0.1 s / m; Deep sandstone: ∈r=5-10, σ=0.001-0.01 s / m.

7. The method according to claim 1, wherein the method is characterized by: In step 6, the background system receives real-time electric field intensity data and dynamic background field intensity value E returned by the capacitive probe or UWB radar bg , and runs the abnormality determination logic formula: if E field >(E bg +200):tag.as.anomaly; If the conditions are met, the background system will mark the area as abnormal; Bind the electric field intensity data with the GPS coordinates, and generate an electric field distribution heat map through spatial interpolation; In the formula, E field E is the current measured ground field intensity value, which is collected in real time by a capacitive probe or UWB radar; E bg E is the background field intensity value, which is dynamically collected every 10 seconds, and is used for calibrating environmental interference and providing a reference; 200 is an anomaly determination threshold value, which is an empirical value based on historical data statistics; tag.as.anomaly is the area marked as abnormal.

8. The method according to claim 1, wherein the method is a method for tower grounding network multi-frequency composite wave electric field imaging and soil environmental medium diagnosis. In step 7, the multi-parameter fusion model input vector X = [p, q, T] or X = [E d , p, q] is input; wherein, p is the soil resistivity, q is the soil moisture content, T is the soil temperature, E d is the electric field distortion rate, and q is the soil temperature and humidity. The corrosion risk index calculation formula is: R = a - ln(l / p) + β - θ + γ - e 0.05(T-25) ; The output is the corrosion risk level; The health assessment index calculation formula is: Where R is the corrosion risk index, used to quantify the corrosion risk level of the grounding grid; HealthIndex is a health evaluation index for quantifying the overall health state of the grounding grid; a, b, g are weight coefficients for balancing the contribution of each parameter to the health evaluation index, which are obtained by training historical failure data; e 0.05(T-25) is a simplified form of the Arrhenius equation, which describes the nonlinear effect of temperature on the rate of chemical reaction.

9. The method according to claim 8, wherein the method is characterized by: The corrosion risk level division rule is: Low risk: R < R low or HealthIndex < H low ; Medium risk: R low <R < R high or H low < Health Index < H high ; High risk: R > R high or Health Index > H high ; wherein R low high are the low and high risk thresholds for the corrosion risk index; H low high are the low and high risk thresholds for the health assessment index;​​ The early warning decision is to set a confidence threshold, below which manual review is triggered, and the GIS platform automatically triggers an alarm, combined with the health assessment results to guide maintenance priority sorting.

Images