Photovoltaic system and power distribution network state cooperative monitoring method based on Internet of Things

Abstract

The invention provides a photovoltaic system and power distribution network state cooperative monitoring method based on the Internet of Things, and the method comprises the steps: deploying a multi-point collection device in a key region of a photovoltaic power station through a sensor network, and obtaining the soil humidity change and soil conductivity at regular time; denoising and standardizing the acquired soil humidity change and soil conductivity data to generate a standardized environmental parameter matrix; according to the dynamic change trend of grounding grid corrosion, identifying the mapping relation between the soil environment parameters and the grounding resistance value, and combining the real-time environment parameters and the corrosion trend to predict the resistance change; and comparing the predicted resistance change with the current grounding resistance value in real time, if the current grounding resistance value is detected to exceed a safety threshold range, triggering an abnormal early warning signal, and determining a potential area with high electric shock risk according to the resistance deviation degree and the corrosion rate.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of information technology, and in particular to a photovoltaic system and power distribution network state collaborative monitoring method based on Internet of Things. BACKGROUND

[0002] The safe operation of photovoltaic systems is an important issue in the field of new energy, directly related to energy supply stability and personal safety, and its research has an important strategic value. With the widespread application of photovoltaic power stations, the reliability of the grounding system has become a key link to ensure equipment and personal safety, and any decline in grounding performance can lead to serious consequences. However, current solutions to grounding system problems focus on post-repair or regular detection, lacking the ability to adapt to dynamic changes in the environment, especially in complex natural conditions, making it difficult to predict potential risks in advance, leading to long-term latent safety hazards. In this field, the corrosion problem of the grounding grid is particularly critical, especially in the context of changing soil environments, corrosion can significantly affect the stability of the grounding resistance. Soil moisture and conductivity, as important factors affecting corrosion rate, their fluctuations often accelerate the loss of grounding grid materials, and thus lead to an increase in grounding resistance. This abnormal change in resistance not only can raise the potential of the equipment shell to the ground, increasing the risk of electric shock, but also can interfere with the normal operation of the entire system, creating safety hazards. More seriously, changes in the soil environment are highly uncertain, and changes in moisture and conductivity will interact, further complicating the corrosion of the grounding grid. An increase in moisture can increase the erosive nature of the soil, while changes in conductivity can affect current distribution, accelerating localized corrosion. This multi-factor interwoven dynamic influence makes it difficult for traditional static detection methods to accurately predict long-term trends in grounding performance, and also makes it difficult to provide timely and effective early warnings for safety management. Therefore, how to build a dynamic grounding performance prediction model based on soil moisture and conductivity data, and identify the risk of increased grounding resistance in advance through real-time monitoring methods, has become a key issue in ensuring the safe operation of photovoltaic power stations. SUMMARY

[0003] The present application provides a photovoltaic system and power distribution network state collaborative monitoring method based on Internet of Things, mainly including:

[0004] The soil humidity change characteristics and the conductivity distribution state are acquired; the soil humidity change and the soil conductivity are denoised and standardized, to generate a normalized environmental parameter matrix; the normalized environmental parameter matrix and historical grounding grid corrosion data are combined to construct a time series-based analysis framework, to simulate the influence of the soil humidity change and the soil conductivity on the corrosion rate of the grounding grid, to take the corrosion depth and the corrosion area expansion speed as evaluation indexes, to obtain the dynamic change trend of the grounding grid corrosion; the mapping relationship between the soil environmental parameters and the grounding resistance value is identified according to the dynamic change trend of the grounding grid corrosion, and the real-time environmental parameters and the corrosion trend are combined to predict the resistance change; the predicted resistance change is compared with the current grounding resistance value, to trigger an abnormal early warning signal, to determine a high risk of electric shock area; the current distribution difference and the local corrosion degree data of the high risk of electric shock area are acquired, to perform risk level evaluation; the monitoring frequency of the sensor network is adjusted according to the risk level evaluation result, to acquire encrypted monitoring environmental parameters and resistance change data, to generate a risk distribution map; the data acquisition frequency of the sensor network is adjusted according to the risk distribution map, to acquire environmental and resistance correlation data, to establish a corrosion early warning mechanism and maintenance suggestion, to real-time feedback the grounding grid corrosion condition of the photovoltaic system.

[0005] Further, the soil humidity change characteristics and the conductivity distribution state are acquired, including:

[0006] The sensor deployment density is determined according to the terrain elevation difference and the soil type distribution of the photovoltaic power station, to acquire soil humidity data and soil conductivity data of different depth layers, to form a multi-dimensional data set containing time stamp, depth, humidity value and conductivity value; for the multi-dimensional data set, the soil humidity difference and the conductivity difference between adjacent collection time points are calculated, to obtain the humidity change rate and the conductivity change rate, and the measurement points exceeding the preset threshold are recorded as abnormal feature points; according to the spatial coordinates of the abnormal feature points, the spatial covariance function of the humidity value and the conductivity value is calculated, and the Kriging interpolation method is used to generate soil humidity spatial distribution data and conductivity spatial distribution data; the spatial gradient is calculated according to the soil humidity spatial distribution data and the conductivity spatial distribution data, the positions with gradient values exceeding the threshold are identified, the adjacent positions are connected to form a regional boundary, and the soil humidity change characteristics and the conductivity distribution state are determined.

[0007] Further, the soil humidity change and the soil conductivity are denoised and standardized to generate a normalized environmental parameter matrix, including:

[0008] The standard deviation of the soil humidity data and the conductivity data is calculated, data points with a standard deviation exceeding a threshold value are identified, a median filter method is used for processing to obtain denoised soil humidity sequences and conductivity sequences; based on the denoised soil humidity sequences and conductivity sequences, a linear interpolation method is used to complete missing data to generate a complete data set; the complete data set is normalized to generate normalized humidity data and normalized conductivity data; and a normalized environmental parameter matrix is constructed based on the normalized humidity data and the normalized conductivity data.

[0009] Further, the normalized environmental parameter matrix and historical grounding grid corrosion data are combined to construct a time series-based analysis framework to obtain a dynamic change trend of the grounding grid corrosion, including:

[0010] The soil humidity and conductivity time series data in the normalized environmental parameter matrix are time-aligned with corrosion depth records of the historical grounding grid corrosion data, and a corrosion depth increment sequence and a correlation coefficient are calculated; based on the correlation coefficient, a cross-correlation function is calculated to determine a delay length of soil humidity change and conductivity change to corrosion response; time shift processing is performed based on the delay length to calculate a predicted corrosion depth increment and a corrosion area expansion speed, and a corrosion depth time sequence and a corrosion area time sequence are generated.

[0011] Further, the mapping relationship between the soil environmental parameters and the grounding resistance value is identified based on the dynamic change trend of the grounding grid corrosion to predict resistance change, including:

[0012] Based on the dynamic change trend of the grounding grid corrosion, a radial corrosion ratio and a surface corrosion ratio are calculated to determine a reduction amount of the effective conductive cross-sectional area, and a numerical correspondence relationship between corrosion feature parameters and resistance increment is generated; based on the numerical correspondence relationship, a corrosion and resistance mapping function is constructed; based on the corrosion and resistance mapping function and real-time soil humidity and conductivity data, a corrosion depth prediction value and a corrosion area prediction value are calculated to generate a resistance increment prediction value and a resistance change rate.

[0013] Further, the predicted resistance change is compared with the current grounding resistance value to trigger an abnormal warning signal and determine a high risk area of electric shock, including:

[0014] Based on the predicted resistance change, a resistance deviation value is calculated to generate abnormal state data containing the deviation value and the corrosion rate, and a warning signal is triggered; based on the deviation value and the corrosion rate of the abnormal state data, a risk level is determined to generate a risk feature set; based on the risk feature set, measurement points with a spatial distance less than a threshold value are identified, an average value of the corrosion rate in a block is calculated, and a risk block distribution map is generated; based on the risk block distribution map, a comprehensive danger coefficient is calculated to determine a high risk area of electric shock.

[0015] Further, the current distribution difference and local corrosion degree data of the high electric shock risk area are acquired, and risk level assessment is performed, including:

[0016] By measuring the current value of the grounding conductor in the high electric shock risk area, the current distribution unevenness index is calculated;According to the current distribution unevenness index, the conductor residual thickness is detected, and the grounding conductor cross section loss rate is calculated;According to the grounding conductor cross section loss rate, the soil pH value and metal ion concentration are collected, the pH deviation and correlation coefficient are calculated, and the corrosion correlation intensity value is generated;According to the corrosion correlation intensity value and the normalized data, the comprehensive risk index is calculated, and the risk level is determined.

[0017] Further, the monitoring frequency of the sensor network is adjusted according to the risk level assessment result, and a risk distribution map is generated, including:

[0018] According to the risk level assessment result, the position where the current distribution unevenness exceeds the threshold value is screened, and the sensor sampling interval is adjusted;According to the adjusted sampling interval, the encrypted monitoring data sequence is collected, and the abnormal period where the change rate exceeds the threshold value is identified;According to the monitoring value of the abnormal period, the comprehensive risk index and the risk level distribution data are updated;According to the risk level distribution data, the measurement point coordinates are mapped to a two-dimensional plane, and a risk distribution map is generated by using a bilinear interpolation method.

[0019] Further, the data acquisition frequency of the sensor network is adjusted according to the risk distribution map, the environment and resistance correlation data are acquired, the corrosion early warning mechanism is established, and the maintenance suggestion is included:

[0020] According to the risk distribution map, high-risk measurement points are selected, and the shortest acquisition period is set;According to the high-frequency acquisition data, the correlation coefficient of soil humidity change and resistance increment is calculated, the time delay value is determined, and the corrosion stage is divided;According to the corrosion stage, the graded early warning trigger condition is set, and the information containing the early warning level and the maintenance operation suggestion is generated;According to the early warning information, real-time state data containing the corrosion stage and environmental risk factors are generated and sent to the monitoring terminal.

[0021] The technical scheme provided by the embodiment of the application can include the following beneficial effects:

[0022] The application discloses a photovoltaic system and power distribution network state cooperative monitoring method based on Internet of Things, which collects soil humidity and conductivity data through sensor network deployment in key areas, combines historical corrosion data to construct time series analysis framework, simulates the influence of environmental parameters on corrosion rate, and predicts the change of grounding resistance. When the resistance value exceeds the safety threshold, a warning signal is triggered and high-risk areas are identified. The application further analyzes the local environment and equipment state, generates a risk distribution map, adjusts the monitoring frequency accordingly, and strengthens the key point monitoring. Finally, an erosion warning mechanism and maintenance recommendations are established to provide real-time feedback on the grounding net corrosion and guide preventive maintenance. The application realizes intelligent monitoring and early warning of photovoltaic power station grounding net corrosion, improves system safety and maintenance efficiency. BRIEF DESCRIPTION OF DRAWINGS

[0023] Fig. 1 A flowchart of the photovoltaic system and power distribution network state cooperative monitoring method based on Internet of Things.

[0024] Fig. 2 A schematic diagram of the photovoltaic system and power distribution network state cooperative monitoring method based on Internet of Things. DETAILED DESCRIPTION

[0025] The technical solutions in the embodiments of the application will be described clearly and in detail below with reference to the accompanying drawings in the embodiments of the application. The described embodiments are only a part of the embodiments of the application.

[0026] As Figs. 1-2 , the photovoltaic system and power distribution network state cooperative monitoring method based on Internet of Things specifically can include:

[0027] Step S101, a multi-point acquisition device is deployed in key areas of a photovoltaic power station through a sensor network to obtain soil humidity change and soil conductivity regularly.

[0028] In a photovoltaic power station, the sensor deployment density is determined according to the terrain elevation difference and soil type distribution. The layered humidity data of different depths is obtained by the soil humidity sensor of each collection point, and the soil conductivity value of the corresponding depth is also obtained. The data is collected at a preset time interval to form a multidimensional data set containing time stamp, depth, humidity value and conductivity value of each measurement point. For the time series data of each measurement point in the multidimensional data set, the soil humidity difference between adjacent collection times is divided by the time interval to obtain the humidity change rate. The same method is used to calculate the conductivity change rate. If the humidity change rate of a measurement point exceeds the preset threshold and the conductivity change rate also exceeds the corresponding threshold, the spatial coordinates, humidity value, conductivity value and change rate data of the measurement point at that time are recorded as abnormal feature points. According to the spatial coordinates and corresponding humidity and conductivity values of the recorded abnormal feature points, the spatial covariance function of the humidity values and the spatial covariance function of the conductivity values between the measurement points are calculated. Based on the covariance function, the Kriging interpolation method is used to perform spatial interpolation on the humidity and conductivity of the whole region respectively to obtain the soil humidity spatial distribution data and the conductivity spatial distribution data covering the complete region of the photovoltaic power station. The spatial gradient of the soil humidity spatial distribution data and the conductivity spatial distribution data is calculated respectively to identify the positions with gradient values exceeding the threshold as parameter mutation points. The adjacent mutation points are connected to form a regional boundary. According to the humidity mean difference and the conductivity mean difference inside and outside the boundary, the soil humidity change characteristics and the soil conductivity distribution state of the key region of the photovoltaic power station are determined.

[0029] Specifically, in the soil monitoring of a photovoltaic power station, the determination of sensor deployment density needs to consider the terrain elevation difference and soil type distribution comprehensively.

[0030] Specifically, in a region with large terrain undulations, the soil moisture penetration and evaporation rates at different elevation positions are different, so the sensor deployment density needs to be increased.

[0031] For example, the soil moisture at the top of a slope is easy to flow away, while the soil at the bottom of the slope is easy to accumulate water, which directly affects the stability of the photovoltaic component foundation. Soil type also affects the deployment strategy. Sandy soil has strong permeability but poor water retention, while clay is the opposite, so monitoring points need to be increased at the junction of different soil types. The construction of a multidimensional data set is the basis for accurate monitoring.

[0032] In one possible implementation, each sensor collection point is configured with a multi-depth probe to obtain soil parameters of shallow, middle and deep layers respectively. The shallow layer data reflects the immediate influence of surface evaporation and rainfall infiltration, the middle layer data reflects the main storage state of soil moisture, and the deep layer data indicates the change trend of underground water level. The recording of time stamp ensures the time sequence continuity of the data, so that the subsequent change rate calculation has an accurate time reference.

[0033] It should be noted that the calculation of the humidity change rate and the conductivity change rate is realized by using a difference method. When the soil humidity changes sharply in a short time, it often indicates the occurrence of an abnormal situation, such as underground pipeline leakage or local water accumulation. The synchronous change of the conductivity further verifies the abnormality, because the soil conductivity is closely related to the water content and the concentration of dissolved salts. By setting a double threshold judgment mechanism, false judgments caused by single parameter fluctuations are avoided, and the reliability of abnormality detection is improved. The calculation of the spatial covariance function is the core step of the Kriging interpolation.

[0034] Exemplarily, when two measurement points are close to each other, the soil parameters of the two measurement points have a strong correlation, and the covariance value is large; as the distance increases, the correlation gradually weakens, and the covariance value tends to be stable. This spatial correlation feature enables the Kriging interpolation to reasonably estimate the parameter distribution of the unknown region according to the data of the known points. The interpolation process fully utilizes the spatial continuity feature of the soil parameters, and generates the full-area distribution data for the subsequent gradient analysis to provide a continuous spatial field. The calculation of the spatial gradient reveals the spatial variation intensity of the soil parameters.

[0035] In one embodiment, the gradient value of each position is obtained by calculating the ratio of the parameter difference value between adjacent grid points to the distance. The positions with gradient values exceeding a threshold value often correspond to the mutation interfaces of the soil properties, and the identification of these interfaces is of great significance for the zoning management of the photovoltaic power station. The region boundary formed by connecting adjacent mutation points divides the photovoltaic power station into different soil characteristic regions, and the soil humidity and conductivity in each region have relatively consistent distribution characteristics, which facilitates the development of targeted maintenance strategies.

[0036] In step S102, the collected soil humidity change and soil conductivity data are denoised and standardized to generate a normalized environmental parameter matrix.

[0037] For the collected soil humidity and conductivity raw data, the standard deviation of the data in the preset time window of each collection point is calculated, the data points with standard deviation exceeding the threshold are identified as potential noise points, the median filtering method is used to process the potential noise points, and the denoised soil humidity sequence and conductivity sequence are obtained. According to the denoised soil humidity sequence and conductivity sequence, the missing value position in the sequence is detected, if the data change amplitude of the time before and after the missing value is within the preset range, the linear interpolation method is used to complete the missing data, and the time-continuous complete humidity data set and complete conductivity data set are obtained. By calculating the maximum and minimum values in the complete humidity data set, each humidity data point is subtracted by the minimum value and divided by the difference between the maximum value and the minimum value, the normalized humidity data is obtained, and the same method is used to process each data point in the complete conductivity data set, the normalized conductivity data is obtained, and all data are mapped to the preset standard numerical interval. According to the time sequence order and spatial position information of the normalized humidity data and the normalized conductivity data, a humidity parameter matrix and a conductivity parameter matrix are constructed respectively, the rows of the two matrices correspond to the same time, and the columns correspond to the same spatial position sensor, and the normalized environmental parameter matrix containing double parameter information is formed by matrix splicing or multi-dimensional array combination.

[0038] Specifically, the noise identification of soil monitoring data depends on the accurate calculation of statistical characteristics.

[0039] In a possible implementation, the calculation of the standard deviation is based on the data distribution characteristics in the sliding time window. When the sensor is affected by electromagnetic interference or mechanical vibration, the collected data will deviate suddenly, and the standard deviation of these deviation points is obviously higher than the normal fluctuation range. The selection of the median filtering method is based on its good suppression ability to pulse noise, by selecting the middle value in the time window to replace the abnormal value, the overall trend of the data is retained, and the influence of instantaneous interference is eliminated.

[0040] Specifically, the generation of missing values is often caused by temporary failure of the sensor or interruption of data transmission. The linear interpolation method utilizes the continuity characteristics of soil parameters in a short time. When the missing data at a certain time is detected, the data difference between the adjacent time before and after is calculated, and the change amount is distributed according to the time ratio, so as to reconstruct the data at the missing time. This processing method is especially suitable for environmental parameters such as soil humidity and conductivity which have gradual change characteristics, because the physical characteristics of soil determine that the parameters of soil will not jump sharply in a very short time. The core of normalization processing is to eliminate the dimension difference of different physical quantities.

[0041] It should be noted that soil humidity is usually expressed in volume percentage, while conductivity is expressed in Siemens per meter, and the numerical range of the two is quite different. By subtracting the minimum value, the starting point of all data is unified; by dividing the range, the dynamic range of the data is compressed. This linear transformation preserves the relative size of the original data, so that subsequent data processing will not be biased due to dimensional differences.

[0042] In one embodiment, the construction of the environmental parameter matrix adopts a structured data organization method. The humidity parameter matrix and the conductivity parameter matrix store the corresponding physical quantity information respectively, and the two matrices have the same dimension structure. The row index corresponds to the sequential arrangement of the time stamp, ensuring the complete retention of the time sequence information; the column index is mapped to the sensor number of different spatial positions, reflecting the spatial distribution characteristics. Matrix splicing can adopt horizontal splicing to form an extended matrix, where the first half of the column stores humidity data and the second half of the column stores conductivity data. The implementation of multi-dimensional array provides another data organization idea. By constructing a three-dimensional array structure, the first dimension represents time, the second dimension represents spatial position, and the third dimension distinguishes between humidity and conductivity parameter types. The advantage of this organization method is that it can directly access the double-parameter information of a specific spatio-temporal position through indexing, which is convenient for subsequent correlation analysis and pattern recognition. The normalized environmental parameter matrix provides a standardized data basis for the soil state evaluation of the photovoltaic power station.

[0043] Step S103, combined with the normalized environmental parameter matrix and the historical grounding net corrosion data, a time series-based analysis framework is constructed to simulate the influence path of soil humidity change and soil conductivity on the corrosion rate of the grounding net, and the corrosion depth and the corrosion area expansion speed are used as key evaluation indicators to obtain the dynamic change trend of the grounding net corrosion.

[0044] According to the time sequence data of soil humidity and conductivity in the normalized environmental parameter matrix, and the corrosion depth record of the corresponding time period in the historical grounding net corrosion data, the time alignment is performed, the corrosion depth increment sequence is obtained through the corrosion depth difference of adjacent time points, the Pearson correlation coefficients of the humidity sequence and the corrosion depth increment sequence, and the Pearson correlation coefficients of the conductivity sequence and the corrosion depth increment sequence are calculated, to determine the correlation value of the environmental parameters and the corrosion process. For parameter sequences with correlation values exceeding a preset threshold, the cross-correlation function values under different time offsets are calculated through a moving time window,

[0045]

[0046] , R xy(τ) represents the cross-correlation function value of signal x and signal y at time offset τ, x(t) represents the amplitude of the first input signal at time t, y(t+τ) represents the amplitude of the second input signal at time t+τ, T represents the length of the integration time window, and τ represents the time offset parameter; determine the delay length of the soil humidity change to the corrosion depth response and the delay length of the conductivity change to the corrosion area response, and obtain a time sequence influence characteristic group containing the delay length and the correlation strength. Based on the delay length in the time sequence influence characteristic group, the environmental parameter sequence is subjected to time shift processing, so that the environmental parameters and the corrosion response are aligned in time, a linear regression equation of the corrosion depth increment with respect to the shifted humidity value and a linear regression equation of the corrosion area expansion speed with respect to the shifted conductivity value are established by using the least square method, and the quantitative influence path parameters are obtained.

[0047] ΔD = β0 + β1H * + ε

[0048] where ΔD represents the corrosion depth increment, H* represents the shifted humidity value, β_0 represents the regression intercept parameter, β1 represents the regression slope parameter, and ε represents a random error term;

[0049] v = a + b·σ'

[0050] where v represents the corrosion area expansion speed, a represents the intercept term of the regression straight line, b represents the regression coefficient, i.e., the slope, and σ' represents the shifted conductivity value; by substituting the environmental parameter value at the current time into the regression equation corresponding to the influence path parameter, the predicted corrosion depth increment and corrosion area expansion speed are calculated,

[0051]

[0052] This is the regression equation for predicting the corrosion area expansion speed, where V A represents the predicted corrosion area expansion speed, P i (t) represents the value of the i th influence path parameter at the current time t, γ i represents the regression coefficient corresponding to the i th influence path parameter, n represents the total number of influence path parameters, and δ0 represents the constant term of the regression equation;

[0053]

[0054] This is the regression equation for predicting the corrosion depth increment, where DeltaD represents the predicted corrosion depth increment, E k (t) represents the value of the k th environmental parameter at the current time t, β krepresents the regression coefficient corresponding to the kth environmental parameter, m represents the total number of environmental parameters, and a0 represents the intercept term of the regression equation; the corrosion depth time sequence is obtained by accumulating the corrosion depth increment at each time point starting from the initial corrosion state, and the corrosion area time sequence is obtained by accumulating the corrosion area expansion at each time point, thereby forming the dynamic change trend of the grounding grid corrosion.

[0055] Specifically, time alignment is a basic link for realizing accurate correlation analysis.

[0056] In a possible implementation, the data acquisition frequency in the environmental parameter matrix is usually once per hour, and the historical corrosion data can be monthly or quarterly detection results. By using a linear interpolation method, the low-frequency corrosion depth data is expanded to the same time resolution as the environmental parameters. The corrosion depth increment is calculated in a differential manner, that is, the corrosion depth at the current time point is subtracted from the corrosion depth at the previous time point. This increment reflects the corrosion development speed and can better reflect the immediate influence of environmental factors than the absolute depth value. The calculation of the Pearson correlation coefficient reveals the linear correlation degree between variables.

[0057] Specifically, when the soil humidity increases, water provides an electrolyte environment for electrochemical corrosion, promoting the progress of the corrosion reaction. The correlation coefficient has a numerical range from negative one to positive one, and a positive value indicates a positive correlation, that is, the higher the humidity, the more serious the corrosion. The correlation analysis of the conductivity is also important because the soil conductivity directly reflects the ion concentration and current conduction ability, and the corrosion current is more likely to form in a high-conductivity environment.

[0058] It should be noted that the calculation process of the cross-correlation function involves sliding matching of time sequences. The implementation method of the moving time window is to fix one sequence and shift the other sequence forward or backward step by step, and calculate the correlation coefficient of the two sequences after each shift. When the soil environment changes, the corrosion response does not appear immediately, but there is a certain delay. This delay is due to the electrochemical reaction kinetics characteristics of the corrosion process, and it takes time to accumulate to form a measurable change in corrosion depth. The time offset corresponding to the peak value of the cross-correlation function is the length of the delay. The process of establishing the regression equation based on the least squares method is based on the principle of minimizing the sum of squared errors.

[0059] In an embodiment, the humidity value after alignment is taken as the independent variable, and the corresponding corrosion depth increment is taken as the dependent variable, and the linear parameters that minimize the sum of squared perpendicular distances of all data points to the regression straight line are calculated. The obtained regression equation is in the form of corrosion depth increment equal to the slope multiplied by the humidity value plus the intercept, where the slope reflects the influence intensity of humidity on corrosion, and the intercept represents the basic corrosion rate. Cumulative calculation realizes the conversion from the corrosion rate to the total corrosion amount.

[0060] For example, the corrosion depth at a certain time is equal to the initial corrosion depth plus the cumulative sum of all corrosion depth increments from the initial time to the current time. The cumulative calculation of the corrosion area uses a similar method by accumulating the area expansion speed and the time interval at each time point. This cumulative process generates a time series that visually shows the evolution trajectory of the grounding grid corrosion state, forming a complete dynamic change trend, which provides a quantitative basis for the maintenance decision of the photovoltaic power station.

[0061] Step S104, according to the dynamic change trend of the grounding grid corrosion, the mapping relationship between the soil environment parameters and the grounding resistance value is identified, and the resistance change is predicted by combining the real-time environment parameters and the corrosion trend.

[0062] According to the corrosion depth time series and the corrosion area time series in the dynamic change trend of the grounding grid corrosion, the radial corrosion ratio is calculated by the ratio of the corrosion depth to the conductor radius, the surface corrosion ratio is calculated by the ratio of the corrosion area to the conductor surface area, the reduction of the effective conductive cross-sectional area is determined based on the circular conductor cross-sectional area calculation formula, and the corresponding grounding resistance increment value is calculated according to the physical law that the resistance is inversely proportional to the cross-sectional area, forming the numerical correspondence between the corrosion characteristic parameters and the resistance increment. The corrosion depth value and the corrosion area value in the numerical correspondence are used as input variables, and the grounding resistance increment is used as an output variable, and a function relationship from input to output is constructed by a polynomial fitting method,

[0063] ΔB=a0+a1d+a2A+a3d 2 +a4A 2 +a5dA, wherein ΔR represents the grounding resistance increment, d represents the corrosion depth value, A represents the corrosion area value, a_0 to a_5 represent the polynomial coefficients, the input-output relationship is established by determining the coefficient values by the least square method; if the fitting residual exceeds the preset threshold, increase the polynomial order to re-fit, and get the corrosion and resistance mapping function that meets the accuracy requirement. Based on the corrosion and resistance mapping function and the currently collected soil humidity and conductivity data, the linear regression equation of humidity and corrosion depth increment and the linear regression equation of conductivity and corrosion area expansion speed are used to calculate the corrosion depth prediction value and the corrosion area prediction value in the future period, and the prediction value is substituted into the corrosion and resistance mapping function to obtain the expected grounding resistance increment prediction value. By adding the initial resistance reference value measured when the grounding grid is installed to the grounding resistance increment prediction value, the absolute value of the grounding resistance at the predicted time is obtained, the difference between the resistance value at the current time and the resistance value at the predicted time is divided by the time interval to obtain the resistance change rate, and the resistance change result containing the current resistance value, the predicted resistance value and the change rate is output.

[0064] Specifically, the calculation of the cross-sectional area loss caused by corrosion needs to consider the geometric characteristics of the grounding conductor.

[0065] In one possible implementation, the grounding grid typically employs copper conductors or galvanized steel conductors with circular cross-sections. As corrosion progresses from the surface of the conductors to the interior, the effective radius of the conductors decreases. By measuring the obtained corrosion depth data, the remaining effective radius can be directly obtained by subtracting the corrosion depth from the original radius. According to the calculation principle of the circular cross-sectional area, the cross-sectional area is proportional to the square of the radius, so even a small corrosion depth will result in a significant loss of cross-sectional area.

[0066] It should be noted that the physical law that resistance is inversely proportional to cross-sectional area is derived from the resistivity characteristics of the material. When the cross-sectional area of the conductor decreases, the effective channel through which the current passes narrows, resulting in an increase in resistance. This relationship is manifested in engineering practice as follows: when the cross-sectional area is reduced by half, the resistance is doubled. The impact of corrosion area is reflected in the unevenness of the conductor surface, and local corrosion will form a current concentration effect, further increasing the grounding resistance. The selection of the polynomial fitting method is based on the nonlinear relationship between corrosion and resistance.

[0067] Specifically, the initial corrosion has a small impact on the resistance, but as the corrosion degree deepens, the resistance growth rate will accelerate. A low-order polynomial may not accurately describe this accelerating trend, so a suitable polynomial order needs to be determined through residual analysis. In the fitting process, the corrosion depth and the corrosion area are taken as two independent variables to construct a binary polynomial function, which can consider the combined effects of radial corrosion and surface corrosion.

[0068] In one embodiment, the prediction application of the regression equation embodies the forward-looking nature of time series analysis. By substituting the currently measured soil moisture and conductivity data into the established regression equation, the corrosion development in the future period of time can be calculated. This prediction takes into account the sustained effects of environmental factors, such as the high soil moisture during the rainy season, which will result in a consistently high corrosion rate. The predicted corrosion parameters are then converted into resistance increments through a mapping function, realizing the complete link from environmental monitoring to electrical parameter prediction. The determination of the initial resistance reference value is usually carried out immediately after the completion of the installation of the grounding grid, at which time the grounding conductor has not yet been affected by corrosion, and the measured value can truly reflect the design performance of the grounding grid. As the running time elapses, the actual resistance value will gradually deviate from the initial value. By adding the predicted resistance increment to the initial reference value, the absolute resistance value can be directly compared with the safety standard. The calculation of the resistance change rate provides trend information, and a rapidly increasing change rate indicates that monitoring needs to be strengthened or maintenance measures need to be taken, thereby realizing predictive maintenance of the grounding grid of the photovoltaic power station.

[0069] Step S105, compare the predicted resistance change with the current grounding resistance value in real time, if it is detected that the current grounding resistance value exceeds the safety threshold range, an abnormal early warning signal is triggered, and the potential high-risk area of electric shock is determined according to the resistance deviation degree and the corrosion rate.

[0070] The predicted resistance value in the predicted resistance change result is subtracted from the current grounding resistance value measured in real time to obtain a resistance deviation value. If the current grounding resistance value exceeds a preset upper limit of a safety threshold, abnormal state data including the deviation value, an over-limit time, a measurement position, and a corresponding corrosion rate are generated, and a corresponding early warning signal is triggered. Based on the resistance deviation value and the corrosion rate in the abnormal state data, a change amount of the resistance deviation in adjacent time periods is divided by a time interval to obtain a deviation change speed. The risk level is determined according to the deviation value and the deviation change speed, and a risk feature set including the risk level, the spatial position, and the corrosion rate is formed. The spatial position and the risk level information of each measurement point in the risk feature set are used to identify adjacent points based on a distance. Measurement points with a small spatial distance and similar risk levels are grouped into the same block. The average corrosion rate and the maximum resistance deviation value of all measurement points in each block are calculated to obtain a risk block distribution map. According to the ratio of the maximum resistance deviation value of each block in the risk block distribution map to the safety threshold, and the ratio of the average corrosion rate in the block to the reference corrosion rate, a comprehensive risk coefficient of each block is calculated. Blocks with a comprehensive risk coefficient exceeding a preset risk threshold are determined as potential high-risk areas of electric shock.

[0071] Specifically, the calculation of the resistance deviation value and the comparison with the threshold value are key links of the electric shock risk assessment.

[0072] In a possible implementation, the grounding resistance safety standard of the photovoltaic power station generally requires not to exceed 4 ohms. When the measured value reaches 6 ohms, the deviation value is 2 ohms, indicating that the grounding performance has decreased significantly. This deviation not only reflects the current safety hazard, but more importantly, it, in combination with the corrosion rate, can predict the future deterioration trend. The complete record of the abnormal state data contains the space-time information and the corrosion parameters, and provides comprehensive data support for subsequent risk classification.

[0073] It should be noted that the calculation of the deviation change speed reveals the dynamic characteristics of the grounding system deterioration. When the resistance deviation of a measurement point increases from 0.5 ohms to 1.5 ohms in a month, the change speed is 1 ohm per month. This rapid deterioration often indicates that the corrosion has entered an accelerated period. The determination of the risk level adopts a two-factor determination mechanism: the static deviation value reflects the current risk level, and the dynamic change speed predicts the development trend. The area with a high deviation value and a high change speed is assigned the highest risk level and needs to be given priority.

[0074] Specifically, if both neighboring ground grid nodes in a photovoltaic power station exhibit high-risk characteristics, it is likely that the regional corrosion is caused by poor local soil environment or abnormal groundwater level. By setting a reasonable distance threshold, for example, high-risk points within a 50-meter range are classified into the same block, the aggregation effect of corrosion can be identified. The calculation of the average value of the corrosion rate in the block eliminates the influence of individual abnormal points, and the maximum resistance deviation value ensures the conservatism of the safety assessment.

[0075] In an embodiment, the calculation of the comprehensive risk coefficient embodies the risk assessment idea of multi-factor coupling. The ratio of the resistance deviation value to the safety threshold quantifies the current exceeding degree, the larger the ratio, the worse the grounding performance. The ratio of the corrosion rate to the reference rate reflects the abnormality of corrosion, under normal circumstances, the annual corrosion rate is about 0.1 mm, when the actual rate reaches 0.3 mm, the ratio is 3, indicating that the corrosion is seriously abnormal. The multiplication of the two ratios obtains the comprehensive risk coefficient, which considers both the current state and the development speed. The determination of the high electric shock risk area directly serves the safety management of the photovoltaic power station. These areas usually have obvious spatial aggregation characteristics, and may be concentrated in low-lying, high-salinity soil or dense underground pipeline locations. By comparing the comprehensive risk coefficient with the preset risk threshold, the blocks exceeding the threshold are marked as high-risk areas. This zoning management method enables precise allocation of maintenance resources, prioritizes the corrosion problem of the ground grid in high-risk areas, and effectively prevents electric shock accidents.

[0076] Step S106, obtain the current distribution difference and local corrosion degree data of the high electric shock risk area, including the loss rate of the grounding conductor cross section, the change of soil acidity and alkalinity, and the concentration of metal ions, perform correlation analysis and risk level evaluation, and form a local environment and equipment state report.

[0077] For the grounding conductor in the area with high risk of electric shock, the unevenness of current distribution is obtained by measuring the current values at different positions and calculating the standard deviation of the current difference between adjacent measuring points. Meanwhile, the ultrasonic thickness gauge is used to detect the remaining thickness of the conductor. By subtracting the remaining thickness from the original thickness and dividing by the original thickness, the cross-section loss rate value and its spatial distribution of the grounding conductor are calculated. According to the unevenness of current distribution and the spatial distribution of the cross-section loss rate value, soil samples are collected at the overlapping positions where the unevenness is high and the cross-section loss rate exceeds the threshold. The pH value of the soil is measured by the acid-base tester, and the concentration of copper ions or iron ions in the soil is detected by ion chromatography, forming a comprehensive measurement dataset containing position coordinates, current unevenness, cross-section loss rate, pH value, and metal ion concentration. Based on the comprehensive measurement dataset, the Pearson correlation coefficient of the unevenness of current distribution and the cross-section loss rate is calculated, the absolute value of the difference between the pH value and the neutral value 7 is taken as the pH deviation, and the correlation coefficient of the pH deviation and the metal ion concentration is calculated. If both correlation coefficients exceed the preset threshold, the product of the two correlation coefficients is taken as the corrosion correlation strength value. By multiplying the corrosion correlation strength value, the normalized value of the unevenness of current distribution, the cross-section loss rate, the normalized value of the pH deviation, and the metal ion concentration exceeding multiple by the preset weight coefficients and summing them up, the comprehensive risk index is obtained. According to the comparison of the comprehensive risk index and the grading threshold, the risk level is determined, and the local environment and equipment state report containing the environmental parameters of each measuring point, the corrosion state of the equipment, and the risk level is output.

[0078] Specifically, the standard deviation calculation of the unevenness of current distribution reveals the abnormality degree of the current path in the grounding grid.

[0079] In one possible implementation, under normal circumstances, the current of each node of the grounding grid should be distributed according to the design proportion. When some conductors are severely corroded, the resistance increases, causing the current to redistribute. By installing current transformers at the grid nodes, the branch currents can be monitored in real time. The current difference between adjacent measuring points reflects the local current concentration phenomenon, and the standard deviation of these differences quantifies the unevenness of the overall distribution. The larger the standard deviation, the more uneven the current distribution, and some conductors may be subjected to excessive current and accelerated aging.

[0080] It should be noted that the ultrasonic thickness measurement technology has unique advantages in detecting metal corrosion. When ultrasonic waves propagate in metal, they will be reflected when encountering corrosion pits or cracks. By measuring the round-trip time of ultrasonic waves, the remaining thickness can be accurately calculated. The calculation of the cross-section loss rate directly reflects the corrosion degree of the conductor. When the loss rate reaches 30%, the current-carrying capacity of the conductor will decrease significantly. This quantitative evaluation provides a reliable data basis for subsequent risk analysis. The selection of the collection position of the soil environmental parameters reflects the problem-oriented detection strategy.

[0081] Specifically, the areas with high current non-uniformity tend to have more severe local corrosion, and the soil environmental parameters at these locations can best reflect the corrosion mechanism. The pH meter measures the hydrogen ion activity of the soil solution through a glass electrode. The pH value of normal soil is between 6.5 and 7.5, and deviation from this range will accelerate metal corrosion. Ion chromatography can accurately separate and quantitatively analyze metal ions in the soil. An increase in copper or iron ion concentration directly proves the occurrence of metal dissolution.

[0082] In one embodiment, the calculation of the correlation coefficient uses the Pearson method, which can quantify the degree of linear correlation between two variables. The high correlation between current non-uniformity and cross-sectional loss rate indicates that the cross-sectional reduction caused by corrosion is the main reason for current redistribution. The calculation of the pH deviation uses the absolute value form because both acidic and alkaline environments can accelerate corrosion. When the pH value is 5, the deviation is 2; when the pH value is 9, the deviation is also 2, and the corrosion risk in both cases is equivalent. The product of the two sets of correlation coefficients is used as the correlation strength value, which comprehensively reflects the coupling effect of electrical parameters and chemical environment. The weighted calculation of the comprehensive risk index realizes the quantitative evaluation of multiple factors. The normalization of each parameter eliminates the dimensional differences, allowing different types of indicators to be mathematically operated. The setting of the weight coefficient reflects the contribution of each factor to the risk of electric shock. Generally, the current non-uniformity and cross-sectional loss rate have higher weights because they directly affect the grounding performance. The metal ion concentration exceeding the standard multiple indicates the trend of continuous corrosion development. By comparing with the grading threshold, the risk can be divided into low, medium and high levels, and the state report formed provides comprehensive information support for maintenance decisions.

[0083] Step S107, according to the local environment and equipment state report, identify the location with large current distribution difference, automatically adjust the monitoring frequency of the sensor in the high-risk area, get the encrypted monitoring environmental parameters and resistance change data, and generate the corresponding risk distribution map according to the risk level evaluation result.

[0084] According to the current distribution unevenness value and risk level of each measuring point in the local environment and device state report, the position coordinates with unevenness exceeding the preset threshold are screened out. For the sensors at these positions, according to the corresponding acceleration factors of two, four and eight for low, medium and high risk levels respectively, the new sampling interval is calculated by dividing the original sampling interval by the corresponding acceleration factor, and the control instruction containing the new sampling interval is sent to the sensor. Based on the new sampling interval, the sensor collects soil moisture, conductivity and grounding resistance data according to the updated time frequency, forms an encrypted monitoring data sequence, identifies the abnormal period with the change rate exceeding the threshold by calculating the change rate of data of a plurality of consecutive sampling points, and obtains a time series data set containing normal and abnormal segment markers. The latest monitoring values marked as abnormal segments in the time series data set are used to recalculate the comprehensive risk index of each measuring point. If the comprehensive risk index of a measuring point rises sharply, the risk level of the measuring point and the measuring points within the preset range from the measuring point are synchronously promoted by one level to form a dynamically updated risk level distribution data. By mapping the spatial coordinates of each measuring point in the risk level distribution data to the pixel position of the two-dimensional plane, the corresponding color value is selected according to the risk level value, wherein low risk corresponds to green, medium risk corresponds to yellow, and high risk corresponds to red. The color filling is performed on the blank area between the measuring points by using the bilinear interpolation method to generate a continuous and gradual risk distribution map.

[0085] Specifically, the dynamic adjustment of the monitoring frequency embodies the concept of resource optimization configuration.

[0086] In a possible implementation, the sensor may collect data once an hour in the normal state, the sampling interval is changed to 30 minutes when it is identified as a low-risk area, the sampling interval is 15 minutes in a medium-risk area, and the sampling interval is shortened to 7.5 minutes in a high-risk area. This differentiated monitoring strategy not only ensures close attention to high-risk areas, but also avoids data redundancy and increased energy consumption caused by global high-frequency monitoring. The setting of the acceleration factor is based on the corresponding relationship between the risk level and the monitoring demand. The higher the risk is, the higher the time resolution required is.

[0087] It should be noted that the implementation of the sensor control instruction depends on modern Internet of Things technology. Each sensor is equipped with a programmable microcontroller that can receive and analyze instructions from the central monitoring system. The instruction contains a new sampling interval parameter, and the timer inside the sensor resets the interrupt period according to this parameter. This remote dynamic adjustment mechanism enables the entire monitoring network to flexibly respond to changes in risk in different areas. The calculation of the data change rate provides a quantitative basis for anomaly detection.

[0088] Specifically, when the soil humidity rises from 20% to 35% in three consecutive sampling points, the change rate reaches 5% per hour, which is significantly higher than the normal slow change. This rapid change may indicate a sudden rise in groundwater level or a pipeline leak. The sharp change in conductivity is also of warning significance, jumping from 2 to 5 millisiemens per centimeter, indicating a significant change in ion concentration in the soil, accelerating the electrochemical corrosion process.

[0089] In one embodiment, the linkage promotion mechanism of risk level takes into account the diffusion effect of corrosion. When the comprehensive risk index of a certain sampling point rises sharply, the sampling points within a 50-meter range around it are considered as potential impact areas. This spatial correlation is based on the continuity of the soil environment, and local adverse conditions tend to spread to the surrounding area. The simultaneous promotion of risk level ensures preventive monitoring of potential danger areas and avoids the problem of focusing only on single points while ignoring regional risks. The bilinear interpolation method plays a key role in generating a continuous risk distribution map. The discrete distribution of sampling points cannot directly form a continuous visual effect, and the risk values of the areas between the sampling points need to be estimated by interpolation algorithms. Bilinear interpolation uses the risk values of the four nearest neighbor sampling points, and according to the distance of the point to be interpolated to these four points, it performs a weighted average. The closer the sampling point, the greater the weight, and the risk distribution map generated by this method presents a smooth color transition. The gradient from green to yellow to red intuitively shows the spatial distribution of risk from low to high, enabling maintenance personnel to quickly identify areas that need to be focused on, providing visual support for the development of targeted maintenance plans.

[0090] Step S108, adjust the data acquisition frequency of the sensor network according to the risk distribution map, strengthen the monitoring of key points, obtain the environment and resistance correlation data, establish the corrosion early warning mechanism and maintenance suggestions through the correlation data, and real-time feedback the corrosion situation of the photovoltaic system grounding net and guide the preventive maintenance work.

[0091] According to the coordinates of the high-risk area marked in the risk distribution map, the measuring points with risk values exceeding the preset high-risk threshold are selected as key monitoring points, the shortest collection period is set for the sensors of the key points, and the collection frequency of the low-risk area is reduced to half of the baseline period. High-frequency environmental data and resistance measurement values of the key area are obtained through differentiated monitoring density configuration. Based on the high-frequency environmental data and resistance measurement values, the correlation coefficients of soil humidity change and resistance increment at different time offsets are calculated through a moving time window, the time delay value corresponding to the maximum correlation coefficient is determined, and the time interval between the occurrence time of the conductivity peak value and the time when the corrosion rate accelerates is counted. According to the corrosion rate, the corrosion process is divided into three stages of initial stage, development stage and acceleration stage, and an associated data set containing environmental parameter threshold, resistance change rate and corrosion stage identifier is formed. Through the environmental parameter threshold, resistance change rate and corrosion stage identifier in the associated data set, the hierarchical early warning trigger condition is set. If the current environmental parameter exceeds the threshold of the corresponding stage and the resistance change rate continues to rise, early warning information containing the warning level, location information and predicted corrosion progress time is generated. According to the corrosion stage, a preset maintenance operation suggestion is matched. In the initial stage, it is suggested to strengthen monitoring, in the development stage, it is suggested to take local protection treatment, and in the acceleration stage, it is suggested to replace the damaged conductor. Real-time status reports are generated using early warning information and maintenance operation suggestions. The reports contain the current corrosion stage, environmental risk factor value, warning level and specific maintenance operation of each monitoring point. Through the timing push mechanism, the status report is sent to the monitoring terminal, and the corrosion condition of the photovoltaic system grounding grid is fed back in real time to guide the implementation of preventive maintenance work.

[0092] Specifically, the implementation of differentiated monitoring density configuration is based on the risk grading management concept.

[0093] In a possible implementation, the setting of the high-risk threshold generally considers the safety standard of grounding resistance and the critical value of corrosion rate. When the comprehensive risk index of a certain area reaches more than 70% of the full score, the area is marked as a high-risk area. The sensor collection period of the key monitoring point may be shortened from the regular 60 minutes to 5 minutes, achieving a 12-fold increase in data density. This high-frequency collection can capture the rapid changes of environmental parameters and the subtle fluctuations of resistance values, providing data support for accurate early warning.

[0094] It should be noted that the moving time window technique plays a key role in time delay correlation analysis. The influence of soil environment change on the corrosion of grounding grid does not appear immediately, but there is a certain hysteresis effect. By setting different lengths of time window, from 1 hour to 24 hours, the correlation coefficient of the humidity change sequence and the resistance increment sequence in each window is calculated. When the time window moves to a certain position, the correlation coefficient reaches a peak, and the time offset at this time is the delay time of the influence of environmental factors on corrosion. Accurate grasp of this delay characteristic enables the early warning system to respond in advance. The division of corrosion stages is based on the change characteristics of the corrosion rate.

[0095] Specifically, the corrosion rate in the initial stage remains below 0.05 mm per year, and the metal surface is only slightly oxidized; the corrosion rate in the development stage rises to 0.1 to 0.2 mm per year, and obvious corrosion pits appear locally; in the acceleration stage, the corrosion rate exceeds 0.3 mm per year and shows an accelerating upward trend. The threshold values of environmental parameters corresponding to different stages are also different. In the initial stage, attention should be paid when the soil humidity exceeds 30%, while in the acceleration stage, even if the humidity is 25%, it may maintain a high corrosion rate.

[0096] In one embodiment, the matching of maintenance operation recommendations follows the principle of gradual intervention. The strengthened monitoring in the initial stage includes increasing the data acquisition frequency and increasing the number of inspections, which is low in cost but can timely discover abnormal trends. The local protection treatment in the development stage involves coating the corrosion serious areas with anticorrosive coating or installing sacrificial anode, which can effectively delay the corrosion process. The suggestion to replace the damaged conductor in the acceleration stage is based on safety considerations. When the corrosion causes the cross-sectional loss of the conductor to exceed 40%, its mechanical strength and electrical conductivity have been seriously reduced, and there is a risk of fracture if it continues to be used. The generation and push mechanism of real-time status reports ensures the timeliness of the information. The report uses a structured format, and each monitoring point is displayed in groups according to the risk level, with high-risk points marked in a conspicuous way. The display of environmental risk factor values includes current value, historical trend and predicted value, enabling maintenance personnel to fully understand the corrosion development trend. Timed push can adjust the frequency according to the risk level, with high-risk areas being pushed every hour and low-risk areas being pushed daily, which ensures the timely delivery of important information and avoids information overload.

[0097] The above is only a preferred embodiment of the present application, and does not limit the patent scope of the present application, and any equivalent structure or equivalent flow transformation using the content of the specification and drawings, or direct or indirect application in other related technical fields, are also included in the patent protection scope of the present application.

Claims

1. A method for collaborative monitoring of the status of a photovoltaic system and a power distribution network based on the Internet of Things, characterized in that, The method includes: The process involves: acquiring soil moisture variation characteristics and conductivity distribution; denoising and standardizing the soil moisture variation and conductivity data to generate a standardized environmental parameter matrix; constructing a time-series-based analysis framework based on the standardized environmental parameter matrix and historical grounding grid corrosion data to simulate the impact of soil moisture variation and conductivity on the grounding grid corrosion rate, using corrosion depth and corrosion area expansion rate as evaluation indicators to obtain the dynamic trend of grounding grid corrosion; identifying the mapping relationship between soil environmental parameters and grounding resistance values ​​based on the dynamic trend of grounding grid corrosion, and predicting resistance changes based on real-time environmental parameters and the corrosion trend; comparing the predicted resistance changes with the current grounding resistance value to trigger an abnormal early warning signal and identify high-risk areas for electric shock; acquiring current distribution difference and local corrosion degree data in the high-risk areas for risk level assessment; adjusting the monitoring frequency of the sensor network based on the risk level assessment results to acquire encrypted monitoring environmental parameters and resistance change data, generating a risk distribution map; adjusting the data acquisition frequency of the sensor network based on the risk distribution map to acquire environmental and resistance correlation data, establishing a corrosion early warning mechanism and maintenance suggestions, and providing real-time feedback on the corrosion status of the photovoltaic system grounding grid.

2. The method for collaborative monitoring of photovoltaic systems and distribution network status based on the Internet of Things according to claim 1, characterized in that, The acquisition of soil moisture variation characteristics and electrical conductivity distribution includes: The sensor deployment density is determined based on the topographic elevation differences and soil type distribution of the photovoltaic power station. Soil moisture and electrical conductivity data at different depths are acquired to form a multidimensional dataset containing timestamps, depths, moisture values, and electrical conductivity values. For this multidimensional dataset, the difference in soil moisture and electrical conductivity between adjacent acquisition times is calculated to obtain the rate of change in moisture and electrical conductivity. Measurement points exceeding preset thresholds are recorded as abnormal feature points. Based on the spatial coordinates of the abnormal feature points, the spatial covariance function of moisture and electrical conductivity values ​​is calculated, and Kriging interpolation is used to generate spatial distribution data of soil moisture and electrical conductivity. Based on the spatial distribution data of soil moisture and electrical conductivity, a spatial gradient is calculated, and locations where the gradient value exceeds the threshold are identified. Adjacent locations are connected to form regional boundaries, thus determining the characteristics of soil moisture change and the distribution state of electrical conductivity.

3. The method for collaborative monitoring of photovoltaic systems and distribution network status based on the Internet of Things according to claim 1, characterized in that, The process of denoising and standardizing the soil moisture changes and soil electrical conductivity to generate a standardized environmental parameter matrix includes: By calculating the standard deviation of soil moisture and electrical conductivity data, data points with standard deviations exceeding a threshold are identified and processed using median filtering to obtain denoised soil moisture and electrical conductivity sequences. Based on the denoised soil moisture and electrical conductivity sequences, missing data are filled in using linear interpolation to generate a complete dataset. The complete dataset is then normalized to generate normalized moisture and electrical conductivity data. Based on the normalized moisture and electrical conductivity data, a standardized environmental parameter matrix is ​​constructed.

4. The method for collaborative monitoring of photovoltaic systems and distribution network status based on the Internet of Things according to claim 1, characterized in that, By combining the standardized environmental parameter matrix and historical grounding grid corrosion data, a time-series-based analysis framework is constructed to obtain the dynamic trend of grounding grid corrosion, including: Based on the soil moisture and conductivity time-series data in the normalized environmental parameter matrix, the corrosion depth records of historical grounding grid corrosion data are time-aligned to calculate the corrosion depth increment sequence and correlation coefficient. Based on the correlation coefficient, the cross-correlation function is calculated to determine the delay time from soil moisture and conductivity changes to the corrosion response. Based on the delay time, time shifting is performed to calculate the predicted corrosion depth increment and corrosion area expansion rate, generating corrosion depth time series and corrosion area time series.

5. The method for collaborative monitoring of photovoltaic systems and distribution network status based on the Internet of Things according to claim 1, characterized in that, The step of identifying the mapping relationship between soil environmental parameters and grounding resistance values ​​based on the dynamic change trend of grounding grid corrosion, and predicting resistance changes, includes: Based on the dynamic trend of grounding grid corrosion, the radial corrosion ratio and surface corrosion ratio are calculated to determine the reduction in effective conductive cross-sectional area, and a numerical correspondence between corrosion characteristic parameters and resistance increment is generated. Based on the numerical correspondence, a corrosion-resistance mapping function is constructed. Based on the corrosion-resistance mapping function and real-time soil moisture and conductivity data, the predicted corrosion depth and corrosion area are calculated, and the predicted resistance increment and resistance change rate are generated.

6. The method for collaborative monitoring of photovoltaic systems and distribution network status based on the Internet of Things according to claim 1, characterized in that, The step of comparing the predicted resistance change with the current grounding resistance value to trigger an abnormal warning signal and identify high-risk areas for electric shock includes: Based on the predicted resistance change, the resistance deviation value is calculated, and abnormal state data containing the deviation value and corrosion rate is generated to trigger an early warning signal. Based on the deviation value and corrosion rate of the abnormal state data, the risk level is determined, and a risk feature set is generated. Based on the risk feature set, measuring points with a spatial distance less than a threshold are identified, the average corrosion rate within the block is calculated, and a risk block distribution map is generated. Based on the risk block distribution map, a comprehensive hazard coefficient is calculated to determine high-risk areas for electric shock.

7. The method for collaborative monitoring of photovoltaic systems and distribution network status based on the Internet of Things according to claim 1, characterized in that, The process of acquiring data on current distribution differences and localized corrosion levels in the high-risk areas for electric shock, and conducting risk level assessments, includes: By measuring the current value of the grounding conductor in areas with high risk of electric shock, the current distribution non-uniformity index is calculated; based on the current distribution non-uniformity index, the remaining thickness of the conductor is detected, and the cross-sectional loss rate of the grounding conductor is calculated; based on the cross-sectional loss rate of the grounding conductor, soil pH and metal ion concentration are collected, pH deviation and correlation coefficient are calculated, and a corrosion correlation strength value is generated; based on the corrosion correlation strength value and normalized data, a comprehensive risk index is calculated, and the risk level is determined.

8. The method for collaborative monitoring of photovoltaic system and distribution network status based on the Internet of Things according to claim 1, characterized in that, The step of adjusting the monitoring frequency of the sensor network based on the risk level assessment results and generating a risk distribution map includes: Based on the risk level assessment results, locations where the current distribution non-uniformity exceeds the threshold are screened, and the sensor sampling interval is adjusted; based on the adjusted sampling interval, encrypted monitoring data sequences are collected, and abnormal time periods with change rates exceeding the threshold are identified; based on the monitoring values ​​of the abnormal time periods, the comprehensive risk index and risk level distribution data are updated; based on the risk level distribution data, the coordinates of the measuring points are mapped to a two-dimensional plane, and a risk distribution map is generated using a bilinear interpolation method.

9. The method for collaborative monitoring of photovoltaic system and distribution network status based on the Internet of Things according to claim 1, characterized in that, The step of adjusting the data acquisition frequency of the sensor network according to the risk distribution map, obtaining environmental and resistance correlation data, and establishing a corrosion early warning mechanism and maintenance recommendations includes: Based on the risk distribution map, high-risk measurement points are selected, and the shortest acquisition cycle is set. Based on the high-frequency acquisition data, the correlation coefficient between soil moisture change and resistance increment is calculated, the time delay value is determined, and corrosion stages are divided. Based on the corrosion stages, graded early warning trigger conditions are set, and information including early warning level and maintenance operation suggestions is generated. Based on the early warning information, real-time status data including corrosion stages and environmental risk factors is generated and sent to the monitoring terminal.

Images