A photovoltaic power generation and desert sand prevention cooperative three-dimensional monitoring method and system

Abstract

The application provides a photovoltaic power generation and desert sand prevention and control synergistic three-dimensional monitoring method and system, belongs to the photovoltaic sand control monitoring technical field, and the method comprises the following steps: determining a three-dimensional monitoring strategy based on wind and sand damage levels and photovoltaic sand control facility layout and executing the three-dimensional monitoring strategy to obtain synergistic monitoring data; obtaining the sand and dust accumulation amount on the surface of a photovoltaic panel and calculating a sand and dust accumulation rate; if the sand and dust accumulation rate is greater than or equal to a dust accumulation threshold value and the photovoltaic panel power generation efficiency attenuation rate is greater than or equal to a first efficiency attenuation threshold value, all the synergistic monitoring data in the current monitoring period is taken as a second synergistic monitoring data set to obtain a target monitoring data set; otherwise, the synergistic monitoring data is taken as the target monitoring data set; a coupling prediction model is used to obtain a predicted wind and sand migration trend and a photovoltaic efficiency attenuation trend; synergistic risk assessment is carried out to obtain a synergistic risk level; when the synergistic risk level is greater than or equal to a risk threshold value, a synergistic regulation and control strategy is matched, regulation and control instructions are generated and executed.

Description

Technical Field

[0001] This application relates to the field of intelligent monitoring technology for photovoltaic desertification control, and in particular to a three-dimensional monitoring method and system for the synergistic effect of photovoltaic power generation and desertification control. Background Technology

[0002] Along railway lines in deserts and sandy areas, photovoltaic (PV) desertification control facilities are often constructed to achieve comprehensive wind and sand management through PV power generation and vegetation-based sand fixation. Currently, such scenarios mostly rely on single ground-based monitoring points, resulting in limited monitoring range and fixed deployment methods. This fails to allow for differentiated monitoring schemes based on regional wind and sand damage levels and PV layout. Furthermore, existing monitoring systems rely on single data sources and lack multi-source data fusion methods from the sky and ground, failing to comprehensively collect multi-dimensional information on meteorological wind and sand, PV operation, ecological vegetation, and railway safety. Continuous dust accumulation causes PV efficiency degradation, and existing methods cannot accurately identify abnormal dust accumulation conditions. Raw monitoring data is susceptible to deviations due to wind and sand disturbances, sensor drift, and communication interference, leading to insufficient data reliability. Moreover, current methods only achieve data collection and cannot predict trends based on the coupling patterns of wind and sand and PV. Risk assessment is limited in scope and cannot comprehensively assess the synergistic risks of wind and sand, PV, and ecology. Moreover, when risks exceed limits, manual intervention is required, resulting in delayed control and low efficiency. This fails to meet the actual needs of integrated intelligent management and control of railway protection, PV operation and maintenance, and ecological governance in windy and sandy areas.

[0003] Therefore, there is an urgent need for a three-dimensional monitoring method and system that integrates photovoltaic power generation with desert sand control. Summary of the Invention

[0004] To address the aforementioned technical issues, this application provides a three-dimensional monitoring method and system for the coordinated use of photovoltaic power generation and desert sand control.

[0005] A first aspect of this application provides a three-dimensional monitoring method for the coordinated use of photovoltaic power generation and desert sand control, comprising: Based on the wind and sand erosion levels and the layout of photovoltaic desertification control facilities in the target sections along the railway line, a corresponding three-dimensional monitoring strategy was determined. The aforementioned three-dimensional monitoring strategy is implemented to acquire collaborative monitoring data through the constructed integrated sky-ground monitoring network; the collaborative monitoring data includes meteorological and sandstorm parameters, photovoltaic system operating parameters, ecological vegetation parameters, and railway facility safety parameters. The cumulative amount of dust on the surface of the photovoltaic panel at multiple time points during the monitoring period is obtained, and the dust accumulation rate is calculated using the time series of the cumulative amount and the average wind speed during the monitoring period. If the dust accumulation rate is greater than or equal to a preset dust accumulation threshold, and the photovoltaic panel power generation efficiency decay rate is greater than or equal to a first efficiency decay threshold, then all collaborative monitoring data of the current monitoring cycle will be used as the second collaborative monitoring dataset, and the second collaborative monitoring dataset will be evaluated and corrected to obtain the target monitoring dataset. Otherwise, the collaborative monitoring data will be directly used as the target monitoring dataset; Based on the target monitoring dataset, the wind-sand-photovoltaic coupled prediction model is used to obtain the predicted wind-sand migration trend and photovoltaic efficiency degradation trend. Based on the ecological vegetation parameters in the target monitoring dataset, the predicted wind and sand transport trend, and the photovoltaic efficiency degradation trend, a collaborative risk assessment is conducted to obtain the collaborative risk level. If the collaborative risk level is greater than or equal to the preset risk threshold, a preset collaborative control strategy is matched, a collaborative control instruction is generated and executed.

[0006] A second aspect of this application provides a three-dimensional monitoring system that coordinates photovoltaic power generation and desert sand control, comprising: The strategy determination module is used to determine the corresponding three-dimensional monitoring strategy based on the wind and sand erosion level and the layout of photovoltaic desertification control facilities in the target section along the railway line. The execution strategy module is used to execute the three-dimensional monitoring strategy and acquire collaborative monitoring data through the constructed integrated sky-ground monitoring network; the collaborative monitoring data includes meteorological and sandstorm parameters, photovoltaic system operating parameters, ecological vegetation parameters, and railway facility safety parameters. The dust accumulation module is used to obtain the amount of dust accumulation on the surface of the photovoltaic panel at multiple time points during the monitoring period, and to calculate the dust accumulation rate using the time series of the accumulation amount and the average wind speed during the monitoring period. The target data module is used to take all the collaborative monitoring data of the current monitoring cycle as the second collaborative monitoring dataset if the dust accumulation rate is greater than or equal to a preset dust accumulation threshold and the photovoltaic power generation efficiency decay rate is greater than or equal to a first efficiency decay threshold, and to perform quality assessment and correction on the second collaborative monitoring dataset to obtain the target monitoring dataset. Otherwise, the collaborative monitoring data will be directly used as the target monitoring dataset; The trend prediction module is used to obtain the predicted wind and sand movement trend and photovoltaic efficiency degradation trend based on the target monitoring dataset and using the wind-sand-photovoltaic coupled prediction model. The collaborative assessment module is used to conduct a collaborative risk assessment based on the ecological vegetation parameters in the target monitoring dataset, the predicted wind and sand transport trend, and the photovoltaic efficiency degradation trend, and obtain a collaborative risk level. The instruction execution module is used to match a preset collaborative control strategy, generate a collaborative control instruction, and execute it when the collaborative risk level is greater than or equal to a preset risk threshold.

[0007] A third aspect of this application provides an electronic device, including a memory, a processor, and a computer program stored in the memory and running on the processor, wherein the processor executes the computer program to implement the steps of the above-described three-dimensional monitoring method for the coordinated use of photovoltaic power generation and desert sand control.

[0008] In a fourth aspect of this application, a computer-readable storage medium is provided, the computer-readable storage medium storing a computer program, which, when executed by a processor, implements the steps of the above-described three-dimensional monitoring method for the coordinated use of photovoltaic power generation and desert sand control.

[0009] The beneficial effects of the three-dimensional monitoring method and system for the coordinated operation of photovoltaic power generation and desert sand control provided in this application are as follows: This application formulates a dedicated three-dimensional monitoring strategy based on the wind and sand erosion level along the railway line and the layout of photovoltaic sand control, and realizes multi-dimensional synchronous data collection through an integrated sky-ground network. By using a dual-threshold judgment data processing method based on the dust accumulation rate and power generation efficiency decay rate, targeted data quality control and correction are carried out, effectively improving the accuracy of monitoring data. The application predicts the changing trend through a wind-sand-photovoltaic coupling model, completes a coordinated risk assessment in conjunction with ecological parameters, and matches and executes control strategies when risks exceed the limits. This achieves multi-objective coordinated management and control of photovoltaic power generation, desert sand control, and railway protection, improves the automation and intelligence level of photovoltaic sand control projects along desert railway lines, and can proactively avoid multiple risks such as wind and sand, dust accumulation, and ecological degradation, ensuring the efficient power generation of photovoltaic power and the safety of railway operation. Attached Figure Description

[0010] Figure 1 A schematic flowchart illustrating a three-dimensional monitoring method for the synergistic effect of photovoltaic power generation and desert sand control provided in an embodiment of this application; Figure 2 This is a structural block diagram of a three-dimensional monitoring system for the coordinated use of photovoltaic power generation and desert sand control, provided in one embodiment of this application. Figure 3 This is a schematic block diagram of an electronic device provided in an embodiment of this application. Detailed Implementation

[0011] In the following description, specific details such as particular system architectures and techniques are set forth for illustrative purposes and not for limitation, in order to provide a thorough understanding of the embodiments of this application. However, those skilled in the art will understand that this application may also be implemented in other embodiments without these specific details. In other instances, detailed descriptions of well-known systems, apparatuses, circuits, and methods have been omitted so as not to obscure the description of this application with unnecessary detail.

[0012] To make the purpose, technical solution, and advantages of this application clearer, the following will be described in conjunction with the appendix. Figure 1 -Appendix Figure 3 The following is an explanation using specific examples.

[0013] Please refer to Figure 1 , Figure 1 This is a flowchart illustrating a three-dimensional monitoring method for the coordinated use of photovoltaic power generation and desert sand control, provided in an embodiment of this application. The method includes: S101: Based on the wind and sand erosion level and the layout of photovoltaic desertification control facilities in the target section along the railway, determine the corresponding three-dimensional monitoring strategy.

[0014] In this embodiment, the target section along the railway line is a continuous control area within a desert / sandstorm region, adjacent to the railway line and equipped with photovoltaic sand control facilities. This is the designated operational area for monitoring and control, and can be segmented according to mileage, administrative boundaries, and sandstorm characteristics, distinguishing it from other ordinary sections of the entire railway line. The sandstorm damage level is a risk classification result comprehensively assessed based on regional historical sandstorm data, geographical environment, and sandstorm incidents. It is divided into three levels: severe damage area, moderate damage area, and minor damage area, used to characterize the severity of sandstorm erosion and sand accumulation damage to the section, and serves as the basis for setting the monitoring intensity and scope.

[0015] In this embodiment, the layout of photovoltaic desertification control facilities refers to the overall deployment of photovoltaic arrays that have the dual functions of photovoltaic power generation and desert sand prevention and fixation within the region, including a complete set of layout parameters such as the parameters of the photovoltaic panels themselves, the array arrangement, and their relative position to the railway.

[0016] In this embodiment, the three-dimensional monitoring strategy is a comprehensive monitoring plan for the target area, including a complete set of rules such as the delineation of the monitoring area, the configuration of air-space-ground monitoring units, the location / density of equipment deployment, the monitoring cycle, the data acquisition frequency, and the multi-unit collaborative working logic. It serves as the execution basis for guiding the operation of on-site monitoring equipment.

[0017] S102: Implement a three-dimensional monitoring strategy to acquire collaborative monitoring data through a constructed integrated sky-ground monitoring network; collaborative monitoring data includes meteorological and wind and sand parameters, photovoltaic system operating parameters, ecological vegetation parameters, and railway facility safety parameters.

[0018] In this embodiment, the integrated air-space-ground monitoring network is a pre-built three-layered, three-dimensional collaborative monitoring hardware system, including space-based, air-based, and ground-based components. The space-based component is satellite remote sensing for large-scale, macroscopic monitoring of an area; the air-based component is UAV aerial surveying for mesoscale, detailed inspections and imaging of sand accumulation distribution; and the ground-based component is a ground-based Internet of Things (IoT) sensor array for fixed-point, high-frequency, continuous, and high-precision monitoring. These three components are synchronized in time and space, complementing each other to form a fully covered, three-dimensional sensing network.

[0019] In this embodiment, the collaborative monitoring data is multi-source fusion data that is spatiotemporally aligned and synchronously collected by multiple devices. The collaborative monitoring data includes meteorological and wind and sand parameters, photovoltaic system operating parameters, ecological vegetation parameters, and railway facility safety parameters. These four types of parameters are coupled with each other and jointly characterize the operating status of the wind and sand-photovoltaic-ecological-railway coupled system, rather than independent single data.

[0020] Specifically, meteorological wind and sand parameters characterize the intensity of regional wind and sand activity and the meteorological environment: wind speed, wind direction, dust concentration, air temperature and humidity, instantaneous sand transport, and near-surface wind and sand flow characteristics. Photovoltaic system operating parameters characterize the working status and dust accumulation characteristics of photovoltaic sand control components: dust accumulation on the photovoltaic panel surface, real-time power generation, power generation efficiency, component temperature, string power generation, and efficiency degradation degree. Ecological vegetation parameters characterize the region's sand-fixing ecological barrier capacity: vegetation coverage, soil moisture content, vegetation growth status, surface desertification degree, and vegetation spatial distribution. Railway facility safety parameters characterize the safety of railway subgrade and lines under wind and sand threats: thickness of sand accumulation on railway subgrade, amount of sand accumulation on slopes, height of sand accumulation around the track, and the integrity of protective facilities.

[0021] S103: Obtain the cumulative amount of dust on the surface of the photovoltaic panel at multiple time points during the monitoring period, and calculate the dust accumulation rate using the time series of the cumulative amount and the average wind speed during the monitoring period.

[0022] In this embodiment, the monitoring period is a set, continuous, and fixed data statistical calculation cycle, providing a unified time benchmark for wind and sand dust accumulation observation, wind speed statistics, and rate calculation, ensuring that the calculation results are comparable and reusable. Multiple time points are sampling moments evenly and discretely distributed within the monitoring period, rather than single-point instantaneous sampling, used to comprehensively capture the continuous change process of dust accumulation.

[0023] In this embodiment, the cumulative dust accumulation on the photovoltaic panel surface refers to the mass of dust adhering to and accumulating on the surface of the photovoltaic panel module, which is direct physical observation data representing the degree of dust pollution. The time series of the cumulative amount is a dynamic data sequence formed by arranging the cumulative dust amounts collected at multiple time points in chronological order, which can characterize the patterns of dust growth, stagnation, and acceleration.

[0024] In this embodiment, the average wind speed during the monitoring period is the statistical average of all sampled wind speeds throughout the entire monitoring cycle. This is used to normalize the wind field conditions and eliminate the interference of instantaneous strong winds and gust fluctuations on dust accumulation judgment. The dust accumulation rate characterizes the rate at which dust accumulates on photovoltaic panels per unit time and is used to determine whether dust accumulation rapidly exceeds the standard and whether the data correction process is triggered.

[0025] S104: If the dust accumulation rate is greater than or equal to the preset dust accumulation threshold and the photovoltaic power generation efficiency decay rate is greater than or equal to the first efficiency decay threshold, then all the collaborative monitoring data of the current monitoring cycle will be used as the second collaborative monitoring dataset, and the second collaborative monitoring dataset will be quality evaluated and corrected to obtain the target monitoring dataset. Otherwise, the collaborative monitoring data will be used directly as the target monitoring dataset.

[0026] In this embodiment, the preset dust accumulation threshold is a calibrated critical value for dust accumulation growth, used to determine whether rapid and abnormal dust accumulation has occurred, distinguishing between normal settling and dust accumulation caused by wind and sand disasters. The photovoltaic panel power generation efficiency degradation rate is the percentage decrease in the current photovoltaic power generation efficiency compared to the clean baseline efficiency, directly representing the degree to which dust accumulation inhibits power generation performance.

[0027] The preset dust accumulation threshold is obtained through on-site monitoring and experimental calibration. Specifically, multiple typical photovoltaic panel measuring points in the target area are selected, and the dust accumulation rate under different wind speed conditions during the sandstorm season and non-sandstorm season is continuously collected. The upper limit of stable dust accumulation rate under normal conditions is statistically analyzed. Based on the characteristic that the dust accumulation rate of photovoltaic panels will quickly lead to the decline of power generation efficiency and the increase of regional sand accumulation risk after reaching this value, and with reference to the photovoltaic cleaning operation start-up standard and railway sand prevention and control indicators, abnormal extreme values ​​caused by instantaneous gusts are eliminated. After multiple rounds of sample statistics and on-site verification, the final preset dust accumulation threshold is determined.

[0028] In this embodiment, the first efficiency degradation threshold is a preset power generation efficiency degradation threshold, used to determine whether dust accumulation has significantly affected photovoltaic power generation performance. The second collaborative monitoring dataset is the original collaborative monitoring dataset to be inspected and corrected when both thresholds are triggered simultaneously. It contains environmental disturbance errors and cannot be used directly.

[0029] The first efficiency degradation threshold was determined through on-site tests and data statistics. Multiple photovoltaic panels of the same specifications were selected to conduct dust accumulation comparison monitoring tests. The changes in power generation efficiency under different dust accumulation amounts and rates were recorded. The critical degradation ratio that could not meet the normal power generation indicators of the power station after efficiency degradation and was highly correlated with abnormal wind and sand conditions was screened out. Through photovoltaic operation and maintenance cleaning trigger rules and railway line comprehensive protection and control standards, instantaneous efficiency fluctuations caused by short-term environmental disturbances were eliminated. After verification by samples from multiple time periods and multiple regions, the first efficiency degradation threshold was obtained.

[0030] In this embodiment, quality assessment and correction involves verifying the completeness, consistency, and physical correlation of multi-source monitoring data, identifying defects such as wind and sand disturbance, sensor drift, and transmission errors, and correcting data deviations. The target monitoring dataset is a standard dataset that, after screening and correction, is ultimately used for prediction, risk assessment, and control decision-making.

[0031] S105: Based on the target monitoring dataset, the wind-sand-photovoltaic coupled prediction model is used to obtain the predicted wind-sand migration trend and photovoltaic efficiency degradation trend.

[0032] In this embodiment, the wind-sand-photovoltaic coupled prediction model is a coupled model constructed by integrating the wind-sand dynamics mechanism and the characteristics of photovoltaic power generation. It can quantify the correlation between wind-sand movement and photovoltaic dust accumulation and power generation efficiency, and can realize the linkage prediction of the two types of trends.

[0033] Specifically, the wind-sand-photovoltaic coupled prediction model is divided into four layers according to the data flow sequence: data preprocessing layer, feature fusion layer, dual-branch prediction layer, coupling constraint layer, and result output layer. The data preprocessing layer receives the original target monitoring dataset and completes data cleaning, outlier removal, and normalization. The feature fusion layer integrates four types of related features: meteorological wind and sand, photovoltaic operation, ecological environment, and spatial layout, to construct a comprehensive feature vector. The dual-branch prediction layer sets up wind-sand transport prediction branch and photovoltaic efficiency degradation prediction branch in parallel, and completes single-dimensional trend extrapolation respectively. The coupling constraint layer performs bidirectional linkage correction on the results of the two branches based on the physical correlation between wind-sand accumulation and photovoltaic performance. The result output layer organizes the standardized prediction data and outputs the trend sequence. Each layer flows sequentially, and the data progresses unidirectionally, forming a complete prediction processing chain.

[0034] The wind-sand photovoltaic coupled prediction model is built using a dual-branch coupled architecture. Based on a time-series prediction network, two sub-models are constructed: a wind-sand transport sub-model and a photovoltaic efficiency degradation sub-model. These two sub-models share the front-end preprocessing and feature fusion modules. A coupling correlation module is added between the two-branch prediction structures. This module explores the field mechanisms by which wind-sand, sand transport, and dust accumulation affect the photovoltaic module's power generation efficiency. Through cross-branch data interaction channels and correlation constraint rules, it achieves the mutual constraint and synchronous prediction of wind-sand evolution and photovoltaic performance changes.

[0035] The input feature dimension of the wind-sand-solar photovoltaic coupling prediction model is determined based on the total number of monitoring indicators. The time-series input step size is set to 15 monitoring periods, and the prediction step size is set to the next 24 monitoring periods. The model learning rate is set to 0.001, the batch sample size is configured to 32, and the regularization coefficient is set to 0.001 to prevent overfitting. The loss function is the mean absolute error, and the convergence threshold is set to 10. -4 For the wind and sand model, the key features to be configured are wind speed, dust concentration, and sand transport volume. For the photovoltaic sub-model, the focus is on binding the weights of parameters such as dust accumulation rate, ambient temperature and humidity, and initial efficiency of the modules. The coupling module sets linkage correction coefficients according to different wind and sand damage levels. All parameters are determined after multiple sets of simulation tests and field verifications under various working conditions during and outside the wind and sand seasons.

[0036] The training and online application phases of the wind-sand-photovoltaic coupled prediction model maintain consistent input. The input data is a target monitoring dataset that has undergone quality assessment and correction, specifically including multi-dimensional time-series data composed of meteorological wind-sand parameters, photovoltaic system operating parameters, ecological vegetation parameters, and railway facility safety parameters. The model output consists of two parts: first, the wind-sand transport prediction branch outputs the future time period's wind speed profile, sand transport volume, sand accumulation thickness, and other parameter change sequences and wind-sand transport trends; second, the photovoltaic efficiency degradation prediction branch outputs the photovoltaic power generation efficiency, power generation, and other parameter change sequences and photovoltaic efficiency degradation trends. The two types of output results are interrelated and jointly support subsequent collaborative risk assessment work.

[0037] In this embodiment, the wind-blown sand transport trend is a prediction of the characteristics of wind-blown sand activity, such as the direction of sand flow, the amount of sand transported, the extent of sand accumulation, and changes in wind speed, in the future period, representing the dynamic evolution of sand diffusion and deposition. The photovoltaic efficiency degradation trend is a prediction of the continuous changes in the power generation efficiency of photovoltaic panels in the future period, reflecting the decline and pace of power generation performance under the influence of continuous sand accumulation and wind-blown sand environment.

[0038] S106: Based on the ecological vegetation parameters, predicted wind and sand transport trends, and photovoltaic efficiency degradation trends in the target monitoring dataset, a collaborative risk assessment is conducted to obtain the collaborative risk level.

[0039] In this embodiment, collaborative risk assessment is an evaluation process that comprehensively assesses three related risks: wind and sand damage, photovoltaic operation, and ecological degradation, which differs from single-dimensional risk analysis. The collaborative risk level is a risk hierarchy divided after integrating multiple risk indicators, used to intuitively characterize the degree of threat to the overall railway-photovoltaic-ecological system.

[0040] S107: If the collaborative risk level is greater than or equal to the preset risk threshold, match the preset collaborative control strategy, generate collaborative control instructions and execute them.

[0041] In this embodiment, the preset risk threshold is a calibrated risk threshold used to distinguish between a safe and controllable state and a risky state that requires intervention. It is the threshold for initiating intelligent regulation.

[0042] The preset risk threshold was determined by combining historical data statistics with field tests. The records of risk events in the target section over the years were summarized, and the correlation between changes in the comprehensive risk index and problems such as excessive sand accumulation on the roadbed, a sharp drop in photovoltaic efficiency, and large-scale vegetation withering was analyzed. The critical index value of the risk that is about to cause harm was extracted. At the same time, comparative verification was carried out in areas with severe, moderate and slight wind and sand, taking into account the bottom line requirements of the three control targets of railway protection, photovoltaic power generation and ecological sand fixation. The interference caused by occasional data fluctuations was eliminated. After repeated field tests and calibrations, the final preset risk threshold was obtained.

[0043] In this embodiment, the preset collaborative control strategy is a strategy knowledge base that configures specific control strategies, combined control strategies, and execution priority rules for the three main types of risks: wind and sand, photovoltaics, and ecology. The collaborative control instructions are standardized control instructions automatically generated based on risk type and risk level, which can be recognized and executed by on-site execution agencies. These include operational instructions such as sweeping, angle adjustment, sand barrier deployment, ecological irrigation, and vegetation replanting. Execution involves issuing the control instructions to the corresponding on-site execution equipment to complete the entire process of physical operations and risk management.

[0044] As can be seen from the above, this application formulates a dedicated three-dimensional monitoring strategy based on the wind and sand erosion levels along railway lines and the layout of photovoltaic desertification control, achieving multi-dimensional synchronous data collection through an integrated sky-ground network. By using a dual-threshold data processing method based on dust accumulation rate and power generation efficiency decay rate, targeted data quality control and correction are carried out, effectively improving the accuracy of monitoring data. A wind-sand-photovoltaic coupling model is used to predict changing trends, and combined with ecological parameters, a collaborative risk assessment is completed. Once the risk exceeds the standard, a matching control strategy is implemented and executed. This achieves multi-objective collaborative management and control of photovoltaic power generation, desert sand control, and railway protection, improving the automation and intelligence level of photovoltaic desertification control projects along desert railway lines. It can proactively avoid multiple risks such as wind and sand, dust accumulation, and ecological degradation, ensuring efficient photovoltaic power generation and the safety of railway operation.

[0045] In one embodiment of this application, the method for determining the level of wind and sand damage includes: Acquire historical wind and sand disaster data and geographical environment data for the target area. Historical wind and sand disaster data includes historical wind speed observation sequences, historical sandstorm frequency, historical sand accumulation thickness, and historical sand damage accident records. Geographical environment data includes topographic relief, surface vegetation coverage, soil desertification degree, and the angle between the prevailing wind direction and the railway line. Based on historical wind speed observation sequences, the annual average duration of sand-raising wind speed and the probability of extreme wind speed occurrence in the target area are calculated to obtain the first sandstorm intensity sub-index. Based on historical sandstorm frequency and historical sand accumulation thickness, the annual average sand transport flux and sand accumulation rate of the target section are calculated to obtain the second wind and sand intensity sub-index. Based on the topographic relief and the angle between the prevailing wind direction and the railway line, the wind and sand convergence effect index of the target section is calculated to obtain the first wind and sand environment sub-index. Based on the surface vegetation coverage and soil desertification degree, the surface wind erosion resistance index of the target area is calculated to obtain the second wind and sand environment sub-index. Based on historical records of sandstorm incidents, the frequency and severity of sandstorm incidents in the target area were statistically analyzed to obtain sub-indicators of wind and sand hazards. The first wind and sand intensity sub-index, the second wind and sand intensity sub-index, the first wind and sand environment sub-index, the second wind and sand environment sub-index, and the wind and sand hazard sub-index are weighted and integrated to calculate the comprehensive wind and sand damage index. Based on the preset threshold range of the comprehensive wind and sand damage index, the target area is divided into severely damaged areas, moderately damaged areas, or slightly damaged areas.

[0046] In this embodiment, historical wind and sand disaster data is a long-term accumulated statistical data set of wind and sand disasters in the target area, used to characterize the level of wind and sand activity in the region, including historical wind speed observation sequences, historical sandstorm frequency, historical sand accumulation thickness, and historical sand damage accident records. Geographic environmental data are the underlying topographic and ecological parameters that determine the local wind and sand convergence, diffusion, and sand fixation capabilities, including topographic relief, surface vegetation cover, soil desertification degree, and the angle between the prevailing wind direction and the railway line.

[0047] In this embodiment, the historical wind speed observation sequence is a multi-year continuous hourly wind speed time series data, used to statistically analyze the patterns of sandstorms and the probability of extreme wind conditions. The annual average sandstorm wind speed duration is the cumulative duration within a year when the wind speed continuously exceeds the critical sandstorm wind speed, representing the activity level of sandstorm initiation. The probability of extreme wind speed occurrence is the annual frequency of dangerous wind speeds such as strong winds and sandstorms, characterizing the risk of severe sandstorms. The first sandstorm intensity sub-index is obtained from wind speed time series statistics and is a sandstorm intensity scoring index characterizing the strength of wind dynamics.

[0048] In this embodiment, the annual average sand transport flux and sand accumulation rate are obtained from the frequency of sandstorms and the historical sand accumulation thickness, representing the sand transport and deposition capacity. The second wind and sand intensity sub-index is a quantitative scoring index representing the intensity of sand transport and deposition.

[0049] In this embodiment, the wind and sand convergence effect index is calculated based on topographic relief, wind direction, and the angle between the railway and the terrain, characterizing the effect of topography and railway line layout on the convergence, acceleration, and deposition of wind and sand. The first wind and sand environment sub-index characterizes the degree of aggravation of wind and sand accumulation caused by local topography and railway line location.

[0050] In this embodiment, the surface wind erosion resistance index is calculated based on vegetation cover and soil desertification level, characterizing the surface's ability to fix sand and resist wind and sand erosion. The second wind and sand environment sub-index is an evaluation index characterizing the ecological surface sand fixation and protection capacity.

[0051] In this embodiment, the frequency and severity of sandstorm incidents are based on historical sandstorm records and represent the actual degree of damage caused by wind and sand to railways and photovoltaic facilities. The wind and sand hazard sub-indicators are evaluation indicators characterizing the severity of actual sandstorm consequences in a given section.

[0052] In this embodiment, the comprehensive wind and sand damage index is a total quantitative score that integrates wind dynamics, sand transport intensity, topographic convergence, surface erosion resistance, and historical disasters, serving as the basis for determining the sand damage level of a section. The preset grading threshold range is a pre-defined boundary of the index range, used to distinguish between severe, moderate, and minor wind and sand damage levels.

[0053] Specifically, a weighted fusion of the first, second, first, and second wind and sand intensity sub-indicators, the second wind and sand environment sub-indicators, and the wind and sand hazard sub-indicator was performed. The weights of each indicator were determined using the analytic hierarchy process (AHP) combined with industry expert experience. First, a pairwise comparison judgment matrix was constructed using the five sub-indicators as evaluation objects. The matrix was assigned values ​​based on the impact of wind and sand intensity, environmental conditions, and historical hazards on regional wind and sand damage. Then, the judgment matrix underwent a consistency check to eliminate logical biases. Next, the eigenvectors of the matrix were solved using the eigenvalue method and normalized to obtain the initial weight coefficients for each indicator. The weights were fine-tuned using industry standards in railway wind and sand protection and photovoltaic desertification control, along with years of on-site operation and maintenance experience. Finally, verification was performed using multi-section measured samples to ensure that the weight allocation accurately reflects the wind and sand characteristics of different regions, thus finalizing the determination of the weight coefficients.

[0054] The preset grading threshold range is determined by comprehensively considering long-term wind and sand observation data of the target section, the occurrence pattern of sand damage, and on-site operation and maintenance experience. First, the distribution range of the comprehensive wind and sand damage index corresponding to the full sample is statistically analyzed. Based on the correspondence between different index values ​​and the intensity of wind and sand activity, the frequency of sand damage, and the damage to facilities, the index ranges corresponding to three types of damage—severe, moderate, and minor—are divided. Then, cross-validation is carried out for sections with different terrain and vegetation conditions to eliminate interference from abnormal samples. After verification by field measurements over multiple seasons and years, the preset grading threshold range is finally determined.

[0055] As can be seen from the above, this embodiment selects indicators from multiple dimensions such as wind and sand intensity, environmental conditions, and degree of harm, and integrates data on historical wind and sand disasters, geographical environment, and sand damage incidents. Through weighted fusion of multiple sub-indicators, a comprehensive wind and sand damage index is obtained and classified into levels. This approach takes into account factors such as the natural environment, wind and sand activity, and actual disaster impacts, avoiding the one-sidedness of single-indicator assessments and making the wind and sand damage level determination results more consistent with the actual situation on site. It accurately distinguishes the severity of sand damage in different sections, providing a reliable grading basis for subsequent differentiated monitoring, protection, and control work. It also lays a data foundation for the zoned management of photovoltaic sand control areas along railway lines, improving the scientific nature of regional wind and sand risk assessment.

[0056] In one embodiment of this application, a corresponding three-dimensional monitoring strategy is determined based on the wind and sand erosion level and the layout of photovoltaic desertification control facilities in the target section along the railway line, including: Obtain the wind and sand erosion level of the target area; Obtain the layout parameters of the photovoltaic desertification control facilities, including the tilt angle of the photovoltaic panels, row spacing, height above the ground, installation density, and distance between the photovoltaic array and the centerline of the railway line; Based on the level of wind and sand damage and layout parameters, the sand prevention monitoring area between the photovoltaic power station and the railway was determined; Based on the sand control monitoring area, the collaborative configuration scheme of satellite remote sensing monitoring units, UAV aerial survey units and ground IoT monitoring units, the deployment density and location of each monitoring sub-unit, and the differentiated monitoring cycle and data acquisition frequency are determined.

[0057] In this embodiment, the wind and sand damage level is divided into severely damaged areas, moderately damaged areas, and slightly damaged areas based on the risk level of the section according to the comprehensive wind and sand damage index. This indicates the intensity of wind and sand disasters in the region and serves as the basis for setting the monitoring intensity. The layout parameters of photovoltaic desertification control facilities are a series of parameters describing the overall layout and spatial location of the photovoltaic array. They are used to analyze the impact of the photovoltaic structure on the wind and sand flow field, including the following sub-items: Photovoltaic panel tilt angle: the angle between the photovoltaic panel surface and the horizontal plane, affecting the wind and sand circulation and sand accumulation distribution; row spacing: the horizontal distance between two adjacent rows of photovoltaic panels; height above ground: the vertical distance between the bottom of the photovoltaic panel and the ground surface, determining the near-ground wind and sand circulation range; installed capacity density: the photovoltaic installed capacity per unit area, reflecting the density of the photovoltaic array; distance between the photovoltaic array and the centerline of the railway line: the horizontal distance between the photovoltaic field area and the main railway line, defining the protective buffer space.

[0058] In this embodiment, the sand control monitoring area is a key control area located between the photovoltaic power station and the railway, affected by both wind and sand transport and photovoltaic bypass, and is the scope of three-dimensional monitoring. The satellite remote sensing monitoring unit is a space-based large-scale monitoring device responsible for regional area monitoring, acquiring macroscopic information on vegetation, sand dunes, and wind and sand distribution. The UAV aerial survey unit is an airborne inspection device undertaking mesoscale fine-grained monitoring, collecting data on sand accumulation distribution, photovoltaic and railway local conditions. The ground-based IoT monitoring unit is a ground-based fixed-point sensing device, achieving close-range, high-frequency continuous monitoring, collecting detailed data on meteorology, sand accumulation, and equipment operation. The collaborative configuration scheme is the division of labor, linkage, and data interaction rules for the three types of monitoring units, clarifying the monitoring focus and cooperation logic of each unit. Deployment density and location refer to the installation spacing and specific locations of the monitoring equipment; high-risk areas are densely deployed, while low-risk areas are sparsely deployed. The differentiated monitoring cycle and data collection frequency are the inspection intervals and data sampling frequencies set based on regional risk and seasonal characteristics, distinguishing between wind and sand seasons and key / ordinary areas to implement different standards.

[0059] As can be seen from the above, this embodiment delineates sandstorm prevention monitoring areas based on wind and sand erosion levels and photovoltaic facility layout parameters, and then strategically configures three types of monitoring units: satellite, drones, and ground-based IoT. The deployment density, location, monitoring cycle, and data acquisition frequency of equipment are determined according to regional risk differences, achieving a rational allocation of monitoring resources. High-risk areas are enhanced with high-frequency, high-density monitoring, while low-risk areas employ large-scale, lightweight monitoring, ensuring monitoring accuracy in key areas while reducing equipment energy consumption and data redundancy. This fully leverages the advantages of different monitoring devices, constructing a clearly defined collaborative monitoring system, reducing operation and maintenance costs while improving overall monitoring coverage and operational efficiency.

[0060] In one embodiment of this application, a sand prevention monitoring area between a photovoltaic power station and a railway is determined based on the level of wind and sand damage and layout parameters, including: Obtain the height of the photovoltaic panel above the ground; Obtain the angle between the prevailing wind direction and the railway alignment; Based on the level of wind and sand damage, the basic monitoring width is determined as follows: when the level of wind and sand damage is a severe damage area, the basic monitoring width is taken as the first preset width value; when the level of wind and sand damage is a moderate damage area, the basic monitoring width is taken as the second preset width value; when the level of wind and sand damage is a slight damage area, the basic monitoring width is taken as the third preset width value; wherein the first preset width value is greater than the second preset width value, and the second preset width value is greater than the third preset width value. Based on the height of the photovoltaic panel above the ground and the angle between the prevailing wind direction and the railway line, the distance of wind and sand bypass influence is calculated. The total monitoring width is obtained by adding the basic monitoring width to the distance affected by wind and sand flow. Based on the centerline of the railway line, the total monitoring width is extended upwind of the prevailing wind direction, and the preset downwind width is extended downwind of the prevailing wind direction to determine the sand prevention monitoring area. The photovoltaic array is located within the sand control monitoring area, which covers the area between the rows of photovoltaic panels and the buffer zone between the photovoltaic array and the railway.

[0061] In this embodiment, the height of the photovoltaic panel above the ground is the vertical height of the bottom of the photovoltaic panel from the ground surface, which directly determines the range of influence of near-surface windblown sand lifting, circulation, and deposition. The angle between the prevailing wind direction and the railway line is the horizontal angle between the region's perennial prevailing wind direction and the railway alignment, which determines the strength of the oblique intrusion and lateral deposition of windblown sand.

[0062] In this embodiment, the basic monitoring width is a baseline monitoring width determined solely by the level of wind and sand intrusion, representing the baseline of the influence range of a purely natural wind and sand field. The first preset width value, the second preset width value, and the third preset width value are gradient baseline widths corresponding to severe, moderate, and slight wind and sand intrusion areas, decreasing progressively to achieve a larger baseline monitoring range for higher risks.

[0063] The first preset width is the benchmark monitoring width for severely wind-blown sand infestation sections, determined based on extreme wind-blown sand transport intensity, large-scale wind-blown sand diffusion radius, and the span of high-risk sand hazard impact on railways. By statistically analyzing the historical maximum wind-blown sand diffusion range and the width of large-area sand accumulation on the roadbed in severely affected sections, combined with the boundary of large-scale sand disturbance impact from photovoltaic arrays, and eliminating short-term, occasional wind-blown sand disturbance ranges, the first preset width value is obtained by taking the maximum impact envelope width of stable wind-blown sand over many years. The specific range of the first preset width value is 90m-110m, with 100m being the preferred value in this embodiment.

[0064] The second preset width is the benchmark monitoring width for sections affected by moderate wind and sandstorms. It is determined comprehensively based on the annual average wind and sand diffusion range, the width of conventional sand accumulation coverage, and the scale of local wind and sand convergence in this section. According to the stable wind and sand transport width under moderate wind and sand conditions and the range of local sand disturbance caused by photovoltaic power, after excluding extreme wind and sand events, the stable coverage width affected by normal wind and sand events throughout the year is taken to calibrate the second preset width value. This value is less than the first preset width, and the specific range of the second preset width value is 50m-70m. In this embodiment, a value of 60m is preferred.

[0065] The third preset width is the baseline monitoring width for sections with minor wind and sand damage, and is designed for sections with low wind and sand activity, weak sand transport, and weak accumulation. It is determined based on the normal wind and sand diffusion scale, natural settlement range, and extremely low sand hazard impact radius of the minor wind and sand section. It only covers areas with minor near-field wind and sand activity, and takes the smallest value to suit the monitoring needs of low-risk sections. The specific value of the third preset width ranges from 20m to 40m, and in this embodiment, a value of 30m is preferred.

[0066] In this embodiment, the wind and sand diversion influence distance is the additional wind and sand disturbance range caused by the obstruction, lifting, and deflection of wind and sand flow after the photovoltaic array is erected; it is the area affected by additional sand damage brought about by the photovoltaic structure. The total monitoring width is the range of the natural wind and sand baseline and the disturbance range of the photovoltaic structure, which is the complete wind and sand influence width in the upwind direction.

[0067] In this embodiment, the preset downwind width is a fixed, reserved downwind sand settling buffer width, used to monitor the tail sand accumulation area after wind and sand pass through the photovoltaic array. The railway line centerline is the railway's horizontal reference axis, serving as the reference line for delineating the monitoring area. Upwind expansion and downwind expansion respectively expand the monitoring range along the direction of wind and sand inflow and reserve the settlement monitoring range along the direction of wind and sand outflow. The preset downwind width is determined based on the wind and sand flow settlement law; after being blocked, slowed, and disturbed by the photovoltaic array, wind and sand will form a sand particle settling accumulation area downwind. By statistically analyzing the downwind sand accumulation and settlement extension distance of the photovoltaic array under different wind and sand conditions, and combining this with the railway subgrade safety protection buffer width requirements, the stable downwind settlement coverage width under normal wind and sand conditions is taken to obtain the preset downwind width.

[0068] In this embodiment, the area between the rows of photovoltaic panels is the gap between the rows of photovoltaic arrays, which is the core area where wind and sand accumulation is most severe and sand accumulation hotspots are most frequent. The photovoltaic and railway buffer zone is the blank protective belt between the outer edge of the photovoltaic array and the railway subgrade, which is the inevitable channel for wind and sand to invade the railway.

[0069] As can be seen from the above, this embodiment sets a gradient-based basic monitoring width based on the level of wind and sand intrusion, then calculates the wind and sand flow impact distance based on the photovoltaic panel's ground clearance and wind direction angle, and finally determines the total monitoring width by superimposing these values. The monitoring area is delineated by distinguishing upwind and downwind directions using the railway centerline as a reference, ensuring complete coverage of the photovoltaic array, the inter-row area, and the road-ground buffer zone. The influence of the photovoltaic facility's shape and wind and sand flow field characteristics on the monitoring range is fully considered, avoiding issues of missed monitoring areas or insufficient coverage. The core areas of high-incidence wind and sand transport and accumulation are accurately identified, ensuring that the monitoring range matches the on-site wind and sand activity patterns, thus guaranteeing the effectiveness and completeness of the three-dimensional monitoring work.

[0070] In one embodiment of this application, based on the sand control monitoring area, a collaborative configuration scheme for satellite remote sensing monitoring units, UAV aerial survey units, and ground-based IoT monitoring units is determined, including the deployment density and location of each monitoring sub-unit, and differentiated monitoring cycles and data acquisition frequencies, including: The sand control monitoring area is divided into key monitoring sub-areas and general monitoring sub-areas; the key monitoring sub-areas include the area between photovoltaic panels, the section prone to sand accumulation, and the railway main body area; the general monitoring sub-areas are the areas within the sand control monitoring area other than the key monitoring sub-areas. For key monitoring sub-areas, ground-based IoT monitoring units are configured for continuous fixed-point monitoring, and UAV aerial survey units are configured for periodic inspections. For general monitoring sub-areas, satellite remote sensing monitoring units are configured for area coverage monitoring, and sparsely deployed ground-based Internet of Things (IoT) monitoring units are also provided. The deployment density of each monitoring subunit in the ground-based IoT monitoring unit is as follows: In key monitoring sub-areas, the deployment spacing of wind and sand monitoring subunits is the first deployment distance; photovoltaic system monitoring subunits are deployed in units of photovoltaic panel strings; and ecological and soil monitoring subunits are deployed between rows of photovoltaic panels at the second deployment distance. In general monitoring sub-areas, the deployment spacing of each monitoring subunit is a multiple of the first preset multiple of the key monitoring sub-area; the first deployment distance is less than the second deployment distance. The differentiated monitoring cycle and data collection frequency are as follows: During the sandstorm season, the data collection frequency for key monitoring sub-areas is on the minute level, and the data collection frequency for general monitoring sub-areas is on the hour level; during the non-sandstorm season, the data collection frequency for key monitoring sub-areas is on the hour level, and the data collection frequency for general monitoring sub-areas is on the day level.

[0071] In this embodiment, the key monitoring sub-area is the control area with the highest risk of sand accumulation and the most direct impact on railways and photovoltaics within the sand prevention monitoring area. It includes three types of zones: the area between photovoltaic panel rows, the sand accumulation-prone sections, and the railway base area. Among them, the area between photovoltaic panel rows is the area where the photovoltaic deceleration and sand accumulation are most significant; the sand accumulation-prone sections are areas with low-lying terrain, wind and sand convergence, and high annual sand accumulation; the railway base area is the core protection area directly related to railway traffic safety.

[0072] In this embodiment, the general monitoring sub-area is the peripheral transitional area outside the key monitoring sub-area within the sand control monitoring area. This area has relatively mild wind and sand activity, a low probability of sand accumulation, and a weak risk level, making it suitable for large-scale, lightweight monitoring. The satellite remote sensing monitoring unit is responsible for large-scale, area-based, macro-scale monitoring, used to cover vegetation, sand dunes, and overall wind and sand environment changes in general areas.

[0073] In this embodiment, the UAV aerial survey unit is responsible for mesoscale, refined, and periodic inspections, focusing on imaging monitoring of sand accumulation distribution, photovoltaic panel dust accumulation, and roadbed sand damage in key areas. The ground-based IoT monitoring unit is a high-precision, high-frequency fixed-point sensing device that collects data on wind speed, dust, sand accumulation thickness, photovoltaic operating conditions, and soil ecology, serving as the core high-precision data source.

[0074] In this embodiment, the first deployment distance is the minimum numerical value and maximum density of the wind and sand monitoring sensors within the key monitoring sub-area. The second deployment distance is the greater than the first deployment distance, which is the minimum spacing between the ecological and soil monitoring devices within the key monitoring sub-area.

[0075] Specifically, the initial deployment distance was determined based on field observations and experiments. Specifically, wind and sand monitoring points with different spacings were sequentially set up in key areas such as photovoltaic rows and railway subgrades. The ability of data at different spacings to capture local wind and sand abrupt changes and short-term sand accumulation processes was compared and analyzed. Given the characteristics of intense wind and sand activity and large spatial differences in parameters in key areas, the optimal spacing was selected that comprehensively represents the regional wind and sand state while controlling the scale of equipment deployment. This eliminated resource waste caused by overly dense monitoring points and data gaps caused by overly sparse monitoring points. After full-cycle verification during both wind and sand seasons, the final initial deployment distance was obtained.

[0076] The second deployment distance was determined through field comparative experiments. Specifically, ecological and soil monitoring points with different spacings were sequentially set up in the area between the rows of photovoltaic panels to analyze the ability of sampling results at different spacings to reconstruct the ecological state of the entire area. Considering that the spatial variation of ecological parameters is gradual, it is not necessary to deploy points with the high density required for wind and sand monitoring. On the basis of avoiding sampling blind spots and ensuring data validity, the deployment spacing was reasonably increased, and then verified through full-time operation data during both wind and sand seasons to obtain the second deployment distance.

[0077] In this embodiment, the photovoltaic monitoring units are deployed according to the photovoltaic string positions, accurately corresponding to the power generation and sand accumulation status of each photovoltaic array. The first preset multiplier is the equipment sparse amplification factor of the general monitoring sub-area relative to the key area, which is to achieve reduced equipment deployment in low-risk areas.

[0078] Specifically, the first preset magnification factor was determined through a zoned comparative experiment. Specifically, referring to the deployment standards for key monitoring sub-areas, different magnification factor spacings were set in general monitoring sub-areas to verify whether the sampling data at different magnification factors could accurately represent the overall state of the area. Taking advantage of the low risk level and small parameter spatial gradient of general areas, the spacing between monitoring points was appropriately increased to avoid situations where excessive spacing caused monitoring gaps or excessive magnification resulted in resource waste. After full-cycle data verification during both sandstorm seasons and non-sandstorm seasons, the first preset magnification factor was finally obtained.

[0079] In this embodiment, the differentiated monitoring cycle is the equipment inspection and data update cycle differentiated according to seasonal risk and regional risk level. High-frequency sampling during the sandstorm season is used when the risk is high due to active sandstorms, with minute-level sampling in key areas and hour-level sampling in general areas. Low-frequency sampling during the non-sandstorm season is used when the environment is stable and sandstorm activity is weak, with hour-level sampling in key areas and daily sampling in general areas, reducing energy consumption and data redundancy.

[0080] As can be seen from the above, this embodiment divides the monitoring area into two sub-areas: key and general. Monitoring equipment and deployment density are configured differently to match the risk levels of different areas. Key areas achieve continuous monitoring through terrestrial IoT and periodic inspections by drones, ensuring high-frequency and accurate sampling. General areas achieve areal monitoring through satellite, combined with sparse ground equipment to reduce investment. Simultaneously, differentiated sampling frequencies are set according to the sandstorm season and non-sandstorm season to adapt to sandstorm activity patterns. This strengthens monitoring of high-risk areas such as sand accumulation and roadbeds while reasonably controlling the overall operation and maintenance load, effectively balancing monitoring accuracy, coverage, and operating costs, and improving the adaptability and economy of the comprehensive monitoring system.

[0081] In one embodiment of this application, the second collaborative monitoring dataset is quality-assessed and corrected to obtain the target monitoring dataset, including: Data integrity verification, time series consistency analysis, and multi-parameter physical correlation constraint verification were performed on the second collaborative monitoring dataset to obtain quality assessment results. Establish a mapping relationship library between monitoring data error types and collaborative correction strategies. The monitoring data error types include instantaneous disturbance error of wind and sand, cumulative attenuation error of dust accumulation, sensor temperature drift error, communication transmission packet loss error, and multi-source spatiotemporal mismatch error. Based on the error type and degree identified in the quality assessment results, the corresponding candidate collaborative correction strategies are matched from the mapping relationship library; If the matched candidate collaborative correction strategy is a comparison correction strategy based on a digital twin model, then the second collaborative monitoring dataset is corrected using a preset multi-physics coupled digital twin model to obtain the target monitoring dataset. If the matched candidate collaborative correction strategy is an error compensation correction strategy based on historical data statistics, then the second collaborative monitoring dataset is corrected using a preset multivariate error compensation algorithm to obtain the target monitoring dataset.

[0082] In this embodiment, the second collaborative monitoring dataset is the original monitoring dataset marked as distorted by wind and sand disturbance when both the dust accumulation rate and power generation efficiency decay thresholds exceed the limits, requiring quality control correction. Data integrity verification is a quality inspection process that checks whether the dataset has missing measurements, null values, broken frames, or partial data loss, judging the completeness of the data samples. Time series consistency analysis verifies the sampling interval, timestamp continuity, and smoothness of data time series changes, identifying time series jumps, misalignments, and time series disorders. Multi-parameter physical correlation constraint verification uses the objective physical coupling law between wind-sand, photovoltaic, ecological, and railway parameters to verify the authenticity of the data. For example, an increase in wind speed should lead to a synchronous change in dust accumulation rate, and an increase in dust accumulation should lead to a decrease in power generation efficiency, used to identify abnormal data that violates physical mechanisms. The quality assessment result is a structured quality inspection report including the location of data defects, error type, and error severity, which serves as the basis for matching correction strategies.

[0083] In this embodiment, the error type and collaborative correction strategy mapping database is a pre-built one-to-one correspondence knowledge base of error and correction strategy. The instantaneous disturbance error caused by sandstorms is the instantaneous change and abnormal jump error in monitoring data caused by sudden gusts and short-term strong sandstorms in the desert. The cumulative attenuation error caused by dust accumulation is the slow variable accumulation deviation brought about by the attenuation of photovoltaic efficiency and the gradual shift of sandstorm parameters during the continuous accumulation of sand. The sensor temperature drift error is the systematic error caused by the zero-point offset and measurement accuracy drift of the sensing device due to the diurnal fluctuation of ambient temperature. The communication transmission packet loss error is the data packet loss, discontinuity, and partial missing error caused by wireless transmission interference. The multi-source spatiotemporal mismatch error is the data spatiotemporal misalignment and mismatch error caused by the inconsistent sampling time and spatial location of multi-source devices (sky, ground, and air). The candidate collaborative correction strategy is a dedicated data correction scheme automatically retrieved and matched according to the error type, divided into two categories: digital twin model comparison correction and historical statistical error compensation correction.

[0084] In this embodiment, the multiphysics coupled digital twin model is a high-precision simulation digital model that replicates the on-site wind and sand flow field, photovoltaic dust accumulation field, and temperature field. It can output standard true value data for comparison and correction of abrupt changes and nonlinear complex errors.

[0085] Specifically, the multi-physics coupled digital twin model is divided into five layers based on the data flow sequence. The first layer is the parameter access and preprocessing layer, which receives meteorological boundary conditions and initial state parameters and completes data cleaning, anomaly identification, and format standardization. The second layer is the single-physics independent simulation layer, which runs parallel computations on five sub-units: wind field, dust transport field, temperature field, photovoltaic operation field, and soil ecological field. The third layer is the field coupling and interaction layer, which completes data communication and linkage constraints through the interaction mechanism between various physical fields and corrects the simulation results of single fields. The fourth layer is the theoretical data generation layer, which integrates the simulation results of various fields and outputs a standardized theoretical monitoring data set. The fifth layer is the data comparison and correction layer, which compares the theoretical values ​​with the measured data to calculate the deviation, completes the correction of the original dataset, and outputs the target monitoring dataset. Each layer flows sequentially and logically in a closed loop.

[0086] The multiphysics coupled digital twin model is built using an overall architecture of independent single-field simulation and cross-field linkage coupling. First, for five research objects—wind and sand, temperature, photovoltaics, soil, and ecology—independent single-physics simulation sub-models are built, each following the fluid mechanics, heat conduction, and equipment operation laws of its corresponding field. Coupling and interaction interfaces and linkage rules are set between the single-physics sub-models, combining the impact mechanisms of wind and sand erosion, dust accumulation, and temperature and humidity changes on photovoltaic equipment and surface ecology to achieve dynamic data transfer and result mutual correction between fields. The overall model is deeply integrated with on-site photovoltaic desertification control and railway wind and sand protection scenarios, replicating the real spatial layout, equipment morphology, and natural environmental characteristics of the monitoring area, forming a digital mirror system capable of real-time mapping of on-site conditions.

[0087] The spatial parameters of the multiphysics coupled digital twin model were set with a grid size of 20m based on the monitoring area, and the time simulation step was set to 1 minute, consistent with the field data sampling period. The wind field was configured with turbulence intensity, critical wind speed for sand lifting, and sand particle jumping coefficient; the dust transport field was configured with sand particle settling coefficient and sand transport flux conversion coefficient; the temperature field was configured with air heat conduction coefficient and surface heat transfer coefficient; and the ecological field was configured with soil moisture permeability coefficient and vegetation soil and water conservation influence coefficient. Simultaneously, fixed structural parameters such as photovoltaic panel height above ground, array spacing, and panel tilt angle were bound, and the influence weight coefficients of each physical field were fine-tuned in each region based on three levels of wind and sand intrusion: severe, moderate, and minor. All parameters were repeatedly calibrated through field measurements and simulation benchmarking experiments under different seasons and wind and sand intensities, and finalized after the simulation deviation was controlled within the allowable range.

[0088] The inputs to the multiphysics coupled digital twin model during its operation are divided into two categories: first, meteorological boundary conditions, including wind speed, wind direction, dust concentration, ambient temperature, and air humidity; and second, initial state parameters, including initial dust content on the photovoltaic panel surface, initial soil moisture content, and vegetation cover. The model output is a theoretical monitoring data set, specifically including theoretical wind speed profiles, theoretical sediment transport, theoretical photovoltaic panel power generation efficiency, and theoretical soil moisture content. This output is used to calculate the difference between the actual monitoring values ​​in the second collaborative monitoring dataset and the actual monitoring values, obtaining a simulation deviation vector to support subsequent data correction processes.

[0089] In this embodiment, the comparison and correction strategy based on the digital twin model is a high-precision correction method that uses the true value of the digital twin simulation as a benchmark to correct abnormal measured data on site, and is adapted to complex, sudden, and nonlinear wind and sand disturbance errors.

[0090] In this embodiment, the multivariate error compensation algorithm is a statistical compensation model trained based on historical big data, which can calculate system deviation, temperature drift deviation, and cumulative deviation and complete the compensation.

[0091] Specifically, the multivariate error compensation algorithm adopts a multiple linear regression model, and is structured into three progressive layers: a multi-source feature input layer, a linear correlation fitting layer, and a dynamic error compensation output layer. The feature input layer receives four types of error influencing factors: dust accumulation rate, photovoltaic temperature deviation, equipment operating time, and data transmission delay, and performs unified normalization processing. The linear correlation fitting layer constructs a multidimensional linear mapping relationship between multiple factors and monitoring errors, and solves for the regression weights and bias parameters corresponding to each variable. The error compensation output layer dynamically calculates the overall system error based on real-time operating conditions, achieving targeted compensation and correction for sensor temperature drift, dust accumulation attenuation, and communication steady-state deviation.

[0092] The parameters of the multivariate error compensation algorithm are configured using a desert photovoltaic sandstorm monitoring scenario. The model input features are four-dimensional operating condition variables, including sand accumulation rate, photovoltaic panel temperature difference, continuous equipment operating time, and data transmission latency. Feature preprocessing employs Min-Max range normalization, and the extreme values ​​of each feature are obtained and fixed from historical samples. The model uses a least-squares fitting method to configure the basic solution parameters, with the fitting convergence threshold set to 10. -4 The regularization coefficient was set to 0.002 to suppress overfitting and improve the model's generalization ability. The dataset was divided into training, validation, and test sets in a 7:2:1 ratio. Convergence determination was based on both the loss function and the validation set error fluctuation. The regression coefficients of each feature and the model bias parameters were determined through iterative optimization based on historical samples from different dust accumulation levels during the sandstorm season and non-sandstorm season. All parameters were solidified after multi-scenario simulation benchmarking and on-site measurement verification.

[0093] The multivariate error compensation algorithm employs supervised learning for training. First, the entire year-round, full-cycle second collaborative monitoring dataset is cleaned and screened, removing abrupt and abnormal data. Samples are organized by single monitoring cycle, with each sample bound to four types of feature variables and corresponding error labels. The training, validation, and test sets are divided in a 7:2:1 ratio. After uniformly performing Min-Max normalization on the four types of input features, multi-dimensional operating condition features are used as model input, and the actual equipment calibration error values ​​are used as supervision labels. Iterative fitting is performed using the least squares method. Training uses mean squared error as the loss function. When the loss function value falls below 10 after 10 consecutive iterations... -4 Furthermore, if the validation set error fluctuation does not exceed ±0.0005, the model is considered converged, training is terminated, and parameters are fixed. In the online application phase, the model input consists of the dust accumulation rate, photovoltaic temperature difference, equipment operating time, and transmission delay parameters for the current monitoring period. After normalization preprocessing, these parameters are fed into the model for computation. The output is the expected system error value corresponding to each monitoring indicator. Based on the output error value, the original monitoring data is compensated and corrected point by point, ultimately obtaining high-precision target monitoring data that eliminates steady-state offset.

[0094] In this embodiment, the error compensation and correction strategy based on historical data statistics is a correction method that compensates for deviations by utilizing long-term sample statistical patterns, adapting to systematic, gradual, and regular steady-state errors. The target monitoring dataset is the final usable dataset that has undergone error identification and targeted correction, eliminating various interferences, conforming to physical laws, and achieving high precision.

[0095] As can be seen from the above, this embodiment establishes a mapping library between error types and correction strategies for various common field monitoring errors, such as wind and sand disturbance, sensor errors, transmission packet loss, and spatiotemporal mismatch. Based on data quality inspection results, it intelligently matches two types of solutions: digital twin comparison correction and historical statistical error compensation, addressing data anomalies of different causes in a categorized manner. First, a multi-dimensional quality assessment of data integrity, temporal sequence, and physical correlation is completed to accurately locate data defects, and then corresponding strategies are applied for correction. This reduces error interference from multi-source monitoring data in complex desert environments, improves the reliability of collaborative monitoring datasets, and provides high-quality basic data support for subsequent trend prediction and risk assessment.

[0096] In one embodiment of this application, if the matched candidate collaborative correction strategy is a comparison correction strategy based on a digital twin model, then a preset multiphysics coupled digital twin model is used to correct the second collaborative monitoring dataset to obtain the target monitoring dataset, including: Error assessment is performed on the meteorological boundary conditions and initial state parameters in the second collaborative monitoring dataset. If the error exceeds the preset confidence threshold, the error compensation and correction strategy based on historical data statistics is switched to the next step. Otherwise, the meteorological boundary conditions and initial state parameters are input into the preset multi-physics coupled digital twin model. The meteorological boundary conditions include wind speed, wind direction, dust concentration, temperature and humidity, and the initial state parameters include the initial dust content on the photovoltaic panel surface, the initial soil moisture content and the initial vegetation coverage. Run the preset digital twin model to simulate the theoretical monitoring data set corresponding to the monitoring period. The theoretical monitoring data set includes theoretical wind speed profile, theoretical sediment transport rate, theoretical photovoltaic power generation efficiency, and theoretical soil moisture content. The difference between the actual monitored values ​​in the second collaborative monitoring dataset and the corresponding theoretical values ​​in the theoretical monitoring dataset is calculated to obtain the simulation deviation vector; If the magnitude of the simulated deviation vector is less than the preset deviation tolerance threshold, then the second collaborative monitoring dataset will be directly used as the target monitoring dataset. If the magnitude of the simulated deviation vector is greater than or equal to the preset deviation tolerance threshold, the Kalman filter data assimilation algorithm is used to fuse the actual monitoring value with the theoretical value to obtain the assimilated and corrected monitoring data, which is used as the target monitoring dataset.

[0097] In this embodiment, meteorological boundary conditions are external environmental parameters that drive the operation of the digital twin model, including wind speed, wind direction, dust concentration, temperature, and humidity, which determine the external field environment simulated by the model. Initial state parameters are the basic system state parameters at the start of the model operation, including the initial dust content on the photovoltaic panel surface, the initial soil moisture content, and the initial vegetation cover, which characterize the initial operating conditions of the monitoring area.

[0098] In this embodiment, the preset credibility threshold is a critical value used to determine whether meteorological boundaries and initial state parameters are reliable, and is used to filter out severely distorted input parameters. Specifically, the preset credibility threshold is determined by comprehensively considering the normal measurement error of the sensor and the short-term fluctuation range of meteorological parameters, statistically analyzing the reasonable error range under normal operating conditions, and using parameter distortion that would cause the digital twin simulation results to deviate significantly from the actual operating conditions as the judgment boundary. The preset credibility threshold is finally determined through repeated verification through multiple simulation comparison tests.

[0099] In this embodiment, the theoretical monitoring data set is the standard reference data output by the digital twin model simulation, including theoretical wind speed profile, theoretical sand transport rate, theoretical photovoltaic panel power generation efficiency, and theoretical soil moisture content. Specifically, the theoretical wind speed profile is the wind speed distribution pattern at different heights obtained from the model simulation; the theoretical sand transport rate is the total amount of sand and dust transported per unit time obtained from the simulation; the theoretical photovoltaic panel power generation efficiency is the theoretical power generation capacity of the photovoltaic module obtained from the simulation under standard operating conditions; and the theoretical soil moisture content is the simulated standard value of regional soil moisture content.

[0100] In this embodiment, the actual monitored value is the raw measured data collected by field equipment in the second collaborative monitoring dataset. The simulated deviation vector is a vector composed of the differences between multiple sets of actual values ​​and their corresponding theoretical values, comprehensively representing the degree to which the overall data deviates from the theoretical true value. The magnitude of the deviation vector is a scalar value obtained by performing a modulo operation on the deviation vector, serving as a quantitative evaluation index of the overall deviation magnitude. The preset deviation tolerance threshold is a critical value used to determine whether data needs to be fused and corrected; a magnitude less than this value indicates that the data deviation is acceptable.

[0101] The preset deviation tolerance threshold is set by combining the accuracy index of the monitoring system and the allowable error range of wind and sand disturbance on site. It is determined by statistically analyzing the normal deviation between the measured data under normal working conditions and the theoretical data of the digital twin, distinguishing between normal small fluctuations and abnormal large deviations, taking into account both data usage requirements and algorithm correction capabilities. The preset deviation tolerance threshold is obtained after joint verification by on-site measurement and simulation.

[0102] In this embodiment, the Kalman filter data assimilation algorithm is a fusion algorithm that balances the authenticity of measured data with the rationality of theoretical model data, and can correct data bias and suppress random errors. The assimilated and corrected monitoring data is high-precision data obtained after Kalman filter fusion calculation, which is the final target monitoring dataset.

[0103] Specifically, the Kalman filter data assimilation algorithm is divided into three functional layers: state prediction layer, observation update layer, and state output layer. The state prediction layer constructs a priori state estimates using theoretical values ​​output from a multi-physics coupled digital twin model, and simultaneously incorporates the model's inherent error characteristics to complete preliminary deduction. The observation update layer accesses actual field monitoring values, combines two types of noise characteristics to complete filter gain calculation, bias correction, and posterior state update, achieving fusion calibration of theoretical and measured data. The state output layer outputs the optimal state estimation results and completes data format adaptation. Each layer is executed sequentially according to time, with data flowing unidirectionally, forming a complete data assimilation processing chain.

[0104] The Kalman filter data assimilation algorithm requires no iterative training. It uses historical simulation data and measured calibration data to complete parameter calibration and performance verification. By comparing the data fusion errors under different noise matrix parameters, it optimizes and finalizes the matrix parameters suitable for the specific scenario. In actual operation, the algorithm input consists of two parts: first, the theoretical monitoring values ​​output by the digital twin model serve as prior predicted state variables; second, the actual monitoring values ​​from the second collaborative monitoring dataset serve as observed values. After completing state prediction, gain calculation, and data assimilation, the algorithm outputs the optimal state estimate after fusion correction. This result, after format reconstruction, becomes the final assimilated and corrected monitoring data, used in the subsequent risk assessment stage.

[0105] As can be seen from the above, this embodiment avoids model failure by first verifying the reliability of the input parameters of the digital twin model and automatically switching the correction scheme when the parameters are abnormal. It utilizes the theoretical data output from the multiphysics coupled digital twin model, compares it with the measured data to obtain a deviation vector, and executes different processing logics for different scenarios. When the deviation is small, the original data is directly used; when the deviation exceeds the standard, data assimilation correction is initiated, fully leveraging the advantages of the physical laws of the simulation model. This effectively corrects the deviations in measured data such as wind and sand, temperature and humidity, and photovoltaic parameters. By using simulation scenarios, it compensates for the shortcomings of field monitoring, which is susceptible to instantaneous environmental interference, further enhancing the rationality and accuracy of data correction.

[0106] In one embodiment of this application, if the magnitude of the simulated deviation vector is greater than or equal to a preset deviation tolerance threshold, a Kalman filter data assimilation algorithm is used to fuse the actual monitored values ​​with the theoretical values ​​to obtain assimilated and corrected monitoring data, including: Based on the model error statistical characteristics of the preset digital twin model, the process noise covariance matrix of Kalman filtering is determined; Based on the measurement error statistical characteristics of the second collaborative monitoring dataset, the measurement noise covariance matrix of the Kalman filter is determined. The theoretical monitoring values ​​in the theoretical monitoring data set are used as predicted observations, and the actual monitoring values ​​are used as observations. The Kalman filter update equation is executed to calculate the optimal state estimate. The optimal state estimate is reconstructed into assimilated and corrected monitoring data in the same format as the second collaborative monitoring dataset, resulting in assimilated and corrected monitoring data.

[0107] In this embodiment, the statistical characteristics of model error are the systematic deviation, fluctuation range, and error distribution patterns between the long-term simulation output of the multiphysics coupled digital twin model and the true value, used to characterize the simulation uncertainty of the model itself. The process noise covariance matrix (Q matrix) is a core parameter of the Kalman filter, used to describe the degree of error fluctuation in the digital twin model simulation process, and represents the reliability of the model's prediction results.

[0108] In this embodiment, the statistical characteristics of measurement errors are the distribution patterns of measurement deviations, random jitter, and sensor drift in long-term data collected by ground-based, UAV-based, and satellite-based multi-source monitoring equipment, characterizing the uncertainty of the field-measured data. The measurement noise covariance matrix (R matrix) is a core parameter of the Kalman filter, used to describe the noise level of the field sensor measured data and represent the reliability of the monitoring observations.

[0109] In this embodiment, the theoretical monitoring value (predicted observation value) is the simulated true value sequence output by the digital twin model, serving as the prior predicted state quantity for the filtering algorithm. The actual monitoring value (observation value) is the real monitoring data collected by the field equipment, serving as the real-time observation correction quantity for the filtering algorithm. The Kalman filter update equation is a complete set of iterative calculation equations including state prediction, gain calculation, state correction, and covariance update, which is the core mathematical logic for realizing the fusion of model data and measured data. The optimal state estimate is the optimal data state quantity that best fits the actual working conditions, obtained after filtering out random errors and offsetting system deviations from the model simulation law and the actual field measured conditions.

[0110] In this embodiment, the assimilated and corrected monitoring data is generated by reconstructing the optimal state estimate according to the original data time sequence, fields, and format, resulting in standardized high-quality monitoring data with the same structure as the original dataset and greater accuracy.

[0111] As can be seen from the above, this embodiment constructs Kalman filter noise matrices based on digital twin model errors and field measurement errors, and sets filter parameters to closely match actual working conditions. Using theoretical model values ​​as predicted values ​​and actual field measurements as observed values, the optimal state estimate is calculated iteratively through Kalman filtering, thus achieving the fusion and assimilation of the two types of data. This leverages the advantages of Kalman filtering in suppressing random noise and optimizing state estimation, accurately correcting the deviation between measured data and theoretical simulation data. The final output is formatted corrected data, effectively eliminating data fluctuations caused by random disturbances in the desert environment, improving the stability of monitoring data, and ensuring the accuracy of subsequent analysis results.

[0112] In one embodiment of this application, if the matched candidate collaborative correction strategy is an error compensation correction strategy based on historical data statistics, then a preset multivariate error compensation algorithm is used to correct the second collaborative monitoring dataset to obtain the target monitoring dataset, including: Based on the current dust accumulation rate, photovoltaic panel temperature difference and sandstorm season information corresponding to the second collaborative monitoring dataset, a historical error sample set matching the working conditions is retrieved from the historical error database. Analyze the statistical distribution characteristics of each error value in the historical error sample set, and establish an error probability distribution model under the current operating conditions; Using the error probability distribution model, the expected error estimate for each monitoring parameter in the second collaborative monitoring dataset is calculated; A multivariate error compensation algorithm is used to subtract the expected error estimate from the actual monitoring values ​​in the second collaborative monitoring dataset to obtain the corrected monitoring data after error compensation. The spatial continuity of the corrected monitoring data after error compensation is optimized using spatial kriging interpolation to obtain a spatially continuous target monitoring dataset.

[0113] In this embodiment, the current dust accumulation rate is the rate of dust accumulation and growth on the photovoltaic panel surface during the current monitoring period, serving as a characteristic variable for matching the intensity of wind and sand conditions. The photovoltaic panel temperature difference is the deviation between the actual monitored temperature of the photovoltaic panel and the standard reference temperature, used to characterize the degree of influence of temperature drift error. Wind and sand seasonal information is a time-series operating condition label that distinguishes between active wind and sand seasons and non-stable wind and sand seasons, used to match seasonal error patterns. The historical error database is a long-term accumulated multi-condition, multi-parameter error sample library, storing sensor deviations, dust accumulation attenuation errors, and system offset data under different wind and sand intensities, temperature conditions, and seasonal operating conditions. The historical error sample set is a collection of historical error data highly similar to the current dust rate, temperature deviation, and seasonal operating conditions, serving as the basis for current error statistical modeling.

[0114] In this embodiment, the error statistical distribution characteristics are the mathematical and statistical regularities of the mean, variance, distribution range, and central tendency of historical error samples, representing the inherent fluctuation characteristics of the error under this operating condition. The error probability distribution model is a probabilistic statistical model fitted based on samples from similar operating conditions, used to quantitatively characterize the error distribution patterns of each monitoring parameter under the current operating condition. The expected error estimate is the optimal theoretical error compensation amount obtained by solving the probability distribution model, representing the system offset of the data under the current operating condition.

[0115] Specifically, the error probability distribution model adopts a normal distribution (Gaussian distribution) model, divided into four layers: operating condition matching layer, statistical feature analysis layer, probability modeling layer, and error solving layer. The operating condition matching layer retrieves corresponding historical error samples based on the dust accumulation rate, photovoltaic panel temperature difference, and wind and sand season information; the statistical feature analysis layer extracts the distribution patterns of the sample's mean, variance, quantiles, etc.; the probability modeling layer fits a standard probability distribution form suitable for the current operating conditions; and the error solving layer calculates the expected error estimate through the distribution model. Each layer is connected in the order of business logic to realize the entire process from sample selection to error quantification.

[0116] The training process of the error probability distribution model is based on the modeling and calibration of historical monitoring error samples throughout the entire cycle. First, the historical error data is cleaned and grouped. Then, probability distribution functions are fitted one by one according to different operating conditions. The fitting effect is continuously verified and the parameters are optimized until the goodness-of-fit requirement is met, at which point the model is finalized. The input to the error probability distribution model is the characteristic parameters of the current operating condition and the retrieved historical error sample set. The output is the expected error estimate corresponding to each monitoring parameter. This estimate is used as a compensation amount and fed into a multivariate error compensation algorithm to complete data correction, supporting subsequent spatial interpolation optimization.

[0117] In this embodiment, the corrected monitoring data after error compensation is the preliminary corrected data after deducting system errors and eliminating steady-state offsets. The timing accuracy is improved, but the spatial continuity still needs to be optimized.

[0118] In this embodiment, the spatial kriging interpolation method is an optimal interpolation algorithm based on spatial autocorrelation. It can use the spatial correlation of surrounding measurement points to repair spatial discreteness defects, smooth spatial abrupt changes, and fill spatial gaps. The spatially continuous target monitoring dataset is the final standardized dataset that is accurate in time, spatially continuous, and has a reasonable field distribution after error compensation and spatial smoothing optimization.

[0119] As can be seen from the above, this embodiment retrieves and matches historical error samples based on on-site working conditions, constructs an error probability distribution model, solves for the expected error, and completes error compensation using a multivariate algorithm. For the discrepancy after compensation, spatial kriging interpolation is used to optimize spatial continuity, solving the data discontinuity problem caused by uneven distribution of field measuring points. Error correction is achieved through historical big data, eliminating the need for complex simulation models and adapting to application scenarios without digital twin support. It can quickly complete batch correction of multi-source monitoring data, effectively improving the discreteness and distortion problems of monitoring data in desert areas and broadening the applicability of data correction schemes.

[0120] In one embodiment of this application, a collaborative risk assessment is performed based on ecological vegetation parameters, predicted wind erosion trends, and photovoltaic efficiency degradation trends in the target monitoring dataset to obtain a collaborative risk level, including: Based on the predicted sand transport trend, the risk indicators of sandstorm damage are assessed. The risk indicators of sandstorm damage include the predicted sand accumulation thickness, the predicted sand transport volume, and the predicted probability of wind speed exceeding the limit. Based on the predicted photovoltaic efficiency degradation trend, the photovoltaic system operation risk indicators are assessed. The photovoltaic system operation risk indicators include the predicted power generation efficiency degradation rate, the predicted power generation gap, and the predicted equipment failure probability. Based on the ecological vegetation parameters in the target monitoring dataset, ecological degradation risk indicators are assessed. These indicators include the trend of vegetation coverage decline, the degree of soil moisture deficit, and the predicted value of vegetation survival rate. Based on the pre-defined analytic hierarchy process (AHP), the weights of wind and sand damage risk indicators, photovoltaic system operation risk indicators, and ecological degradation risk indicators are determined. After normalizing each risk indicator, a weighted sum is performed to obtain a comprehensive risk index. Based on the pre-defined threshold range of the comprehensive risk index, the comprehensive risk level is determined, and the dominant risk factors are identified by ranking each risk indicator according to its contribution to the comprehensive risk level.

[0121] In this embodiment, the wind and sand attack risk index is a set of parameters representing the future degree of wind and sand hazard, used to assess the magnitude of the threat posed by wind and sand to a region and its facilities. The wind and sand attack risk index includes predicted sand accumulation thickness, predicted sand transport volume, and predicted wind speed exceeding probability. The predicted sand accumulation thickness is a predicted value of the future regional sand accumulation depth derived from wind and sand transport trends. The predicted sand transport volume is the estimated total amount of sand and dust transported per unit time, representing the intensity of sand and dust transport. The predicted wind speed exceeding probability is the probability that wind speeds will exceed safe limits in the future.

[0122] In this embodiment, the photovoltaic system operation risk index is a risk parameter characterizing the operating status and power generation efficiency of photovoltaic equipment, reflecting the operational hazards caused by wind and sand accumulation. The predicted power generation efficiency degradation rate is an estimate of the percentage decrease in the power generation efficiency of the photovoltaic modules. The predicted power generation gap is the predicted value of the missing power generation compared to the rated power generation. The predicted equipment failure probability is a prediction of the likelihood of photovoltaic equipment failure based on efficiency degradation and the windy and sandy environment.

[0123] In this embodiment, the ecological degradation risk index is an evaluation parameter obtained through ecological vegetation parameters, representing the degree of deterioration of the regional ecological sand-fixing system. The vegetation cover decline trend is the gradual decrease in vegetation cover area over time. The soil moisture deficit level is the order of magnitude of the current soil moisture content compared to a reasonable value. The predicted vegetation survival rate is a forecast of the subsequent vegetation survival rate, reflecting the stability of the ecosystem.

[0124] In this embodiment, the Analytic Hierarchy Process (AHP) is a commonly used decision-making method for multi-indicator comprehensive evaluation. It constructs a judgment matrix through pairwise comparisons to scientifically allocate the weights of each risk indicator. The indicator weight is a coefficient representing the importance of each risk indicator in the comprehensive evaluation. Normalization maps indicator data with different dimensions and numerical ranges to the same numerical interval, eliminating the influence of dimensional differences. The collaborative risk comprehensive index is a comprehensive evaluation value obtained by weighted summation of all indicators, representing the overall risk level of the system across the entire domain.

[0125] In this embodiment, the preset threshold range is a pre-defined index range used to classify different levels of collaborative risk. The method for determining the preset threshold range of the collaborative risk comprehensive index is to aggregate comprehensive index samples under different operating conditions, combine the occurrence nodes of adverse events such as sandstorms, photovoltaic shutdowns, and ecological degradation to define the interval boundaries, and verify and correct through multi-regional and multi-seasonal field measurements to finally determine the preset threshold range used for classification. In this embodiment, the contribution of an indicator is the percentage of a single risk indicator's impact on the composite index, reflecting the magnitude of that indicator's role in the overall risk. The dominant risk factor is the risk category or specific indicator that contributes the most and plays a decisive role in the overall risk.

[0126] As can be seen from the above, this embodiment comprehensively covers various hidden dangers in the photovoltaic desertification control collaborative system by extracting multiple detailed risk indicators from three dimensions: wind and sand erosion, photovoltaic operation, and ecological degradation. The analytic hierarchy process (AHP) is used to scientifically allocate the weights of each indicator, and a comprehensive risk index is obtained through normalization and weighted calculation, accurately classifying risk levels and identifying dominant risk factors. This breaks through the limitations of single-dimensional risk assessment, achieving comprehensive risk assessment of multiple coupled systems and clearly distinguishing the impact proportion of various risks. This makes the risk assessment results more comprehensive and based on more sufficient evidence, providing clear guidance for subsequent precise matching of control strategies and identification of key governance areas.

[0127] In one embodiment of this application, if the collaborative risk level is greater than or equal to a preset risk threshold, a preset collaborative control strategy is matched, a collaborative control instruction is generated and executed, including: Based on the dominant risk factors, the corresponding target collaborative strategy is matched from the pre-set collaborative control strategy library; The pre-set collaborative control strategy library includes: First strategy, when the dominant risk factor is the risk of wind and sand damage, generating automated sand barrier deployment instructions and / or photovoltaic panel tilt angle adjustment instructions; Second strategy, when the dominant risk factor is the risk of photovoltaic efficiency degradation, generating photovoltaic panel intelligent cleaning instructions and / or energy storage system charging and discharging scheduling instructions; Third strategy, when the dominant risk factor is the risk of ecological degradation, generating intelligent irrigation instructions and / or vegetation replanting instructions; Fourth strategy, when the collaborative risk level is at the severe risk level and multiple risk factors coexist, based on the contribution ranking of the dominant risk factors and the pre-set strategy conflict resolution rules, generating combined instructions of the above first to third strategies, and determining the execution priority, time window and execution intensity of each basic control instruction; The generated coordinated control commands are sent to the actuator array, which includes one or more of the following: intelligent irrigation actuator, photovoltaic panel intelligent cleaning robot, automated sand barrier deployment device, and grid-connected energy storage controller.

[0128] In this embodiment, the first strategy (sandstorm control strategy) is a specific governance strategy targeting the risks of sandstorm accumulation and excessive sand transport, including two sand control methods: automated sand barrier deployment and photovoltaic panel tilt adjustment. The second strategy (photovoltaic efficiency degradation control strategy) is an operation and maintenance optimization strategy targeting photovoltaic dust accumulation, efficiency degradation, and insufficient power generation, including intelligent cleaning and energy storage charging and discharging scheduling. The third strategy (ecological degradation control strategy) is an ecological restoration strategy targeting vegetation decline, soil water shortage, and decreased sand-fixing capacity, including intelligent irrigation and vegetation replanting. The fourth strategy (multi-risk composite linkage strategy) is a combined control scheme for scenarios with severe risk levels and multiple coexisting risks, addressing the problems of incomplete governance by a single strategy and mutual interference between multiple strategies executed simultaneously.

[0129] In this embodiment, the strategy conflict resolution rules are preset priority determination, timing staggering, and intensity matching rules, used to avoid problems such as interference from multiple devices operating simultaneously, resource contention, and functional cancellation. Execution priority refers to the sequential execution order of multiple control commands, ensuring that core risks are addressed first. The time window is the optimal execution period for each control action, avoiding unfavorable operating conditions such as strong winds, strong sunlight, and high temperatures. Execution intensity refers to the equipment's operational effort, including quantitative parameters such as sweeping frequency, irrigation water volume, sand barrier deployment density, and tilt angle adjustment range.

[0130] In this embodiment, the actuator array is a cluster of intelligent devices on site that can receive control commands and automatically complete governance operations, including irrigation actuators, cleaning robots, automated sand barrier devices, and energy storage controllers.

[0131] As can be seen from the above, this embodiment classifies and matches specific control strategies based on dominant risk factors. For a single risk, a corresponding specific instruction is activated. When multiple risks coexist, combined instructions are generated by combining contribution and conflict rules, clearly defining execution priorities, timing, and intensity. Control instructions can be issued to various types of on-site execution equipment, automating operations such as sand barrier deployment, photovoltaic cleaning, and ecological irrigation. Layered control logic is established for different risk scenarios, avoiding the blind stacking of control measures and improving the targeting of risk management. Seamless integration of risk assessment and intelligent control is achieved, shortening emergency response time and comprehensively mitigating collaborative risks such as wind erosion, dust accumulation, and ecological degradation.

[0132] In one embodiment of this application, after using all collaborative monitoring data of the current monitoring period as the second collaborative monitoring dataset, the method further includes: Acquire the time period and spatial location information corresponding to the portion of monitoring data where the photovoltaic panel power generation efficiency degradation rate is greater than or equal to the first efficiency degradation threshold; Based on time period and spatial location information, a dust hotspot distribution map is generated. The dust hotspot distribution map is used to identify specific photovoltaic panels or panel groups in the photovoltaic array that have severe dust accumulation and require priority cleaning. The dust hotspot distribution map is used as additional information in the second collaborative monitoring dataset for error tracing and correction strategy selection in the quality assessment and correction steps.

[0133] In this embodiment, the power generation efficiency degradation rate is the percentage decrease in the real-time power generation efficiency of the photovoltaic module relative to the clean baseline efficiency, directly representing the degree of dust accumulation on the panel surface. The time period information is the specific time interval within which the photovoltaic efficiency exceeds the standard degradation, used to determine the occurrence and duration of dust accumulation. The spatial location information is the coordinate position of the photovoltaic panel, panel group, or array area to which the abnormal degradation data belongs, used to accurately locate the spatial distribution of dust accumulation.

[0134] In this embodiment, the dust hotspot distribution map is a spatial distribution map drawn based on spatiotemporal anomaly data, used to visually identify photovoltaic panel / panel areas with severe dust accumulation, excessive attenuation, and requiring priority maintenance. The supplementary information in the second collaborative monitoring dataset consists of spatial labels, hotspot locations, and auxiliary information on abnormal time periods stored synchronously with the original monitoring data. This does not alter the original data structure and is used to assist in data quality control and correction. Error tracing distinguishes between genuine performance degradation caused by dust accumulation and false anomalies caused by sensor and transmission errors through the distribution of dust hotspots. The preferred correction strategy intelligently selects an appropriate correction method based on the hotspot distribution characteristics, prioritizing digital twin correction for dust hotspot areas and conventional statistical compensation correction for stable areas.

[0135] As can be seen from the above, this embodiment, after identifying high dust accumulation and low efficiency data, extracts the corresponding spatiotemporal information to generate a dust hotspot distribution map, intuitively displaying the severely dusty areas of the photovoltaic array. Incorporating the hotspot map as supplementary information into the dataset can, on the one hand, assist in tracing data errors and quickly locate the high-incidence locations and time periods of data anomalies; on the other hand, it can prioritize photovoltaic panel cleaning operations, giving priority to addressing dust hotspot areas. This achieves a combination of data anomaly analysis and on-site operation and maintenance guidance, improving the efficiency and targeting of data correction while proactively identifying potential photovoltaic module failure points, thus addressing both data governance and equipment operation and maintenance needs.

[0136] In one embodiment of this application, if the collaborative risk level is greater than or equal to a preset risk threshold, after matching a preset collaborative control strategy, generating a collaborative control instruction, and executing it, the method further includes: Acquire effect monitoring data within a preset monitoring time window after executing the coordinated control command; Based on the effect monitoring data, the control effect indicators are calculated, including the wind and sand damage reduction rate, the photovoltaic power generation efficiency recovery rate, and the ecological vegetation improvement rate. The control effect indicators are compared with the preset expected effect indicators to calculate the control achievement rate; If the control achievement rate is less than the preset achievement rate threshold, the risk level judgment and coordinated control strategy matching steps will be re-executed to generate an upgraded coordinated control instruction. If the achievement rate of adjustments is consistently lower than the preset achievement rate threshold, a manual intervention alarm will be generated to prompt maintenance personnel to conduct on-site inspections.

[0137] In this embodiment, the preset monitoring time window is a fixed observation period reserved after the control action is completed. It is used to observe whether the wind and sand, photovoltaic, and ecological conditions have been effectively restored, and serves as a unified timing benchmark for effect evaluation. The effect monitoring data consists of wind and sand parameters, photovoltaic operation parameters, and ecological vegetation parameters collected within the monitoring time window after the control operation is completed. These data are used to quantify the effectiveness of the control measures.

[0138] In this embodiment, the wind and sand damage reduction rate is the percentage decrease in wind and sand transport intensity, sand accumulation thickness, and the risk of exceeding wind speed limits before and after regulation, representing the effectiveness of wind and sand control. The photovoltaic power generation efficiency recovery rate is the percentage increase in photovoltaic power generation efficiency after regulation compared to before regulation, representing the effectiveness of photovoltaic operation and maintenance. The ecological vegetation improvement rate is the percentage increase in soil moisture content, vegetation coverage, and vegetation survival rate after regulation, representing the effectiveness of ecological restoration. The regulation effect indicators are a comprehensive evaluation index system used to quantify the effectiveness of coordinated regulation and control in three dimensions.

[0139] In this embodiment, the preset expected effect indicator is a preset standard governance compliance value, which serves as a benchmark threshold for judging whether the regulation is qualified. The regulation achievement rate is the degree of agreement between the actual regulation effect and the expected effect, used to quantify the quality of the regulation task completion. The preset achievement rate threshold is a critical threshold for determining whether the regulation has met the standard and whether it needs to be iterated and optimized again.

[0140] The pre-set expected effect indicators were determined comprehensively based on the regional wind and sand erosion level, photovoltaic desertification control facility operation and maintenance standards, and ecological restoration and management requirements. First, wind and sand erosion areas were divided into three categories: severe, moderate, and minor. For each category, the normal improvement levels of wind and sand control, photovoltaic operation and maintenance, and ecological restoration in historical effective control cases were statistically analyzed to establish benchmark target values ​​for wind and sand erosion reduction rate, photovoltaic power generation efficiency recovery rate, and ecological vegetation improvement rate. Then, the benchmark values ​​were calibrated and corrected in conjunction with railway safety protection standards, photovoltaic power station operation indicators, and regional ecological governance acceptance standards. Differentiated indicators were set for normal operating conditions and extreme wind and sand conditions. After multiple rounds of on-site trial operation to verify the rationality of the indicators, they were finalized and used as a unified reference for evaluating the control effect.

[0141] The preset achievement rate threshold was determined by combining historical operating condition statistics with field tests. Historical control effect data corresponding to different risk levels and control strategies were compiled, and sample values ​​of control achievement rate under each scenario were calculated. Based on the criteria of wind and sand damage, photovoltaic efficiency, and ecological restoration to a safe operating range, a minimum qualified achievement standard was defined, and the fault tolerance space of iterative control was determined. After full-cycle operation verification and parameter correction during wind and sand seasons and non-wind and sand seasons, the preset achievement rate threshold was obtained.

[0142] In this embodiment, the upgraded collaborative control command is an optimized control scheme that increases the intensity of operations, increases the frequency of operations, and adds strategy combination dimensions based on the original control strategy. Manual intervention alarms automatically generate maintenance prompts to alert personnel for on-site troubleshooting and repair of persistent and complex risks that cannot be resolved by fully automated control.

[0143] As can be seen from the above, this embodiment collects effect data through a dedicated monitoring window after the control command is executed, quantifies and calculates indicators such as wind and sand reduction rate and efficiency recovery rate, and determines the control achievement rate. For cases where control fails to meet standards, the control strategy is iteratively optimized, and manual alarms are triggered after multiple failures. A closed-loop management process of monitoring, evaluation, control, review, and optimization is constructed, continuously verifying and improving the control effect. Shortcomings in automated control are identified in a timely manner, and manual intervention avoids repeated risk rebounds, continuously optimizing the system's control capabilities and improving overall safety and stability.

[0144] Corresponding to the three-dimensional monitoring method for the coordinated use of photovoltaic power generation and desert sand control in the above embodiment, Figure 2 This is a structural block diagram of a three-dimensional monitoring system for the coordinated use of photovoltaic power generation and desert sand control, provided in one embodiment of this application. For ease of explanation, only the parts relevant to the embodiment of this application are shown. References Figure 2 The three-dimensional monitoring system 20, which integrates photovoltaic power generation and desert sand control, includes: a strategy determination module 21, a strategy execution module 22, a sand and dust accumulation module 23, a target data module 24, a trend prediction module 25, a collaborative evaluation module 26, and an instruction execution module 27.

[0145] Among them, the strategy determination module 21 is used to determine the corresponding three-dimensional monitoring strategy based on the wind and sand erosion level and the layout of photovoltaic desertification control facilities in the target section along the railway. The execution strategy module 22 is used to execute the three-dimensional monitoring strategy and acquire collaborative monitoring data through the constructed integrated sky-ground monitoring network. The collaborative monitoring data includes meteorological and wind and sand parameters, photovoltaic system operation parameters, ecological vegetation parameters, and railway facility safety parameters. The dust accumulation module 23 is used to obtain the amount of dust accumulation on the surface of the photovoltaic panel at multiple time points during the monitoring period, and to calculate the dust accumulation rate using the time series of the accumulation amount and the average wind speed during the monitoring period. The target data module 24 is used to take all the collaborative monitoring data of the current monitoring cycle as the second collaborative monitoring dataset if the dust accumulation rate is greater than or equal to the preset dust accumulation threshold and the photovoltaic power generation efficiency decay rate is greater than or equal to the first efficiency decay threshold, and to perform quality assessment and correction on the second collaborative monitoring dataset to obtain the target monitoring dataset. Otherwise, the collaborative monitoring data will be directly used as the target monitoring dataset; The trend prediction module 25 is used to obtain the predicted wind and sand movement trend and photovoltaic efficiency degradation trend based on the target monitoring dataset and the wind-sand-photovoltaic coupled prediction model. The collaborative assessment module 26 is used to conduct collaborative risk assessment based on the ecological vegetation parameters, predicted wind and sand transport trends and photovoltaic efficiency degradation trends in the target monitoring dataset, and obtain the collaborative risk level. The instruction execution module 27 is used to match the preset collaborative control strategy, generate collaborative control instructions, and execute them when the collaborative risk level is greater than or equal to the preset risk threshold.

[0146] See Figure 3 , Figure 3 This is a schematic block diagram of an electronic device provided according to an embodiment of this application. Figure 3 The electronic device 300 in this embodiment may include one or more processors 301, one or more input devices 302, one or more output devices 303, and one or more memories 304. The processors 301, input devices 302, output devices 303, and memories 304 communicate with each other via a communication bus 305. The memories 304 store computer programs, including program instructions. The processors 301 execute the program instructions stored in the memories 304. Specifically, the processors 301 are configured to invoke the program instructions to perform the functions of the modules in the aforementioned device embodiments, for example... Figure 2 The functions of the strategy determination module 21, strategy execution module 22, dust accumulation module 23, target data module 24, trend prediction module 25, collaborative evaluation module 26, and instruction execution module 27 are shown.

[0147] It should be understood that, in the embodiments of this application, the processor 301 may be a central processing unit (CPU), or it may be other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. The general-purpose processor may be a microprocessor or any conventional processor.

[0148] Input device 302 may include a touchpad, a fingerprint sensor (for collecting the user's fingerprint information and fingerprint orientation information), a microphone, etc., and output device 303 may include a display (LCD, etc.), a speaker, etc.

[0149] The memory 304 may include read-only memory and random access memory, and provides instructions and data to the processor 301. A portion of the memory 304 may also include non-volatile random access memory. For example, the memory 304 may also store device type information.

[0150] In specific implementations, the processor 301, input device 302, and output device 303 described in the embodiments of this application can execute the implementation methods described in any embodiment of the three-dimensional monitoring method for the coordinated photovoltaic power generation and desert sand prevention provided in the embodiments of this application, or they can execute the implementation methods of the electronic devices described in the embodiments of this application, which will not be repeated here.

[0151] In another embodiment of this application, a computer-readable storage medium is provided. This computer-readable storage medium stores a computer program, which includes program instructions. When executed by a processor, the program instructions implement all or part of the processes in the methods described above. Alternatively, the computer program can instruct related hardware to complete the process. The computer program can be stored in a computer-readable storage medium, and when executed by a processor, it can implement the steps of the various method embodiments described above. The computer program includes computer program code, which can be in the form of source code, object code, executable files, or certain intermediate forms. The computer-readable medium can include any entity or device capable of carrying computer program code, a recording medium, a USB flash drive, a portable hard drive, a magnetic disk, an optical disk, a computer memory, a read-only memory (ROM), a random access memory (RAM), an electrical carrier signal, a telecommunication signal, and a software distribution medium, etc.

[0152] The computer-readable storage medium can be an internal storage unit of the electronic device in any of the foregoing embodiments, such as a hard disk or memory of the electronic device. The computer-readable storage medium can also be an external storage device of the electronic device, such as a plug-in hard disk, smart media card (SMC), secure digital card (SD), flash card, etc., equipped on the electronic device. Furthermore, the computer-readable storage medium can include both internal and external storage units of the electronic device. The computer-readable storage medium is used to store computer programs and other programs and data required by the electronic device. The computer-readable storage medium can also be used to temporarily store data that has been output or will be output.

[0153] Those skilled in the art will recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, computer software, or a combination of both. To clearly illustrate the interchangeability of hardware and software, the components and steps of the various examples have been generally described in terms of functionality in the foregoing description. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementations should not be considered beyond the scope of this application.

[0154] Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the specific working process of the electronic devices and units described above can be referred to the corresponding process in the foregoing method embodiments, and will not be repeated here.

[0155] In the several embodiments provided in this application, it should be understood that the disclosed electronic devices and methods can be implemented in other ways. For example, the device embodiments described above are merely illustrative; for instance, the division of units is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. In addition, the mutual coupling or direct coupling or communication connection shown or discussed may be indirect coupling or communication connection through some interfaces or units, or it may be an electrical, mechanical, or other form of connection.

[0156] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of the embodiments of this application, depending on actual needs.

[0157] Furthermore, the functional units in the various embodiments of this application can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or as a software functional unit.

[0158] The above are merely specific embodiments of this application, but the scope of protection of this application is not limited thereto. Any person skilled in the art can easily conceive of various equivalent modifications or substitutions within the technical scope disclosed in this application, and these modifications or substitutions should all be covered within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

Claims

1. A three-dimensional monitoring method for the synergistic effect of photovoltaic power generation and desert sand control, characterized in that, include: Based on the wind and sand erosion levels and the layout of photovoltaic desertification control facilities in the target sections along the railway line, a corresponding three-dimensional monitoring strategy was determined. The aforementioned three-dimensional monitoring strategy is implemented to acquire collaborative monitoring data through the constructed integrated sky-ground monitoring network; the collaborative monitoring data includes meteorological and sandstorm parameters, photovoltaic system operating parameters, ecological vegetation parameters, and railway facility safety parameters. The cumulative amount of dust on the surface of the photovoltaic panel at multiple time points during the monitoring period is obtained, and the dust accumulation rate is calculated using the time series of the cumulative amount and the average wind speed during the monitoring period. If the dust accumulation rate is greater than or equal to a preset dust accumulation threshold, and the photovoltaic panel power generation efficiency decay rate is greater than or equal to a first efficiency decay threshold, then all collaborative monitoring data of the current monitoring cycle will be used as the second collaborative monitoring dataset, and the second collaborative monitoring dataset will be evaluated and corrected to obtain the target monitoring dataset. Otherwise, the collaborative monitoring data will be directly used as the target monitoring dataset; Based on the target monitoring dataset, the wind-sand-photovoltaic coupled prediction model is used to obtain the predicted wind-sand migration trend and photovoltaic efficiency degradation trend. Based on the ecological vegetation parameters in the target monitoring dataset, the predicted wind and sand transport trend, and the photovoltaic efficiency degradation trend, a collaborative risk assessment is conducted to obtain the collaborative risk level. If the collaborative risk level is greater than or equal to the preset risk threshold, a preset collaborative control strategy is matched, a collaborative control instruction is generated and executed.

2. The three-dimensional monitoring method for the synergistic effect of photovoltaic power generation and desert sand control according to claim 1, characterized in that, The method for determining the level of wind and sand damage includes: The historical wind and sand disaster data and geographical environment data of the target section are obtained. The historical wind and sand disaster data includes historical wind speed observation sequences, historical sandstorm frequency, historical sand accumulation thickness and historical sand damage accident records. The geographical environment data includes topographic relief, surface vegetation coverage, soil desertification degree and the angle between the prevailing wind direction and the railway line. Based on the historical wind speed observation sequence, the annual average duration of sand-raising wind speed and the probability of extreme wind speed occurrence in the target section are calculated to obtain the first sandstorm intensity sub-index. Based on the historical frequency of sandstorms and the historical thickness of sand accumulation, the annual average sand transport flux and sand accumulation rate of the target section are calculated to obtain the second wind and sand intensity sub-index. Based on the topographic relief and the angle between the prevailing wind direction and the railway line, the wind and sand convergence effect index of the target section is calculated to obtain the first wind and sand environment sub-index. Based on the surface vegetation coverage and the degree of soil desertification, the surface wind erosion resistance index of the target section is calculated to obtain the second wind and sand environment sub-index. Based on the historical sandstorm incident records, the frequency and severity of sandstorm incidents in the target area are statistically analyzed to obtain wind and sand hazard sub-indicators. The first wind and sand intensity sub-index, the second wind and sand intensity sub-index, the first wind and sand environment sub-index, the second wind and sand environment sub-index, and the wind and sand hazard sub-index are weighted and fused to calculate the comprehensive wind and sand damage index. Based on the preset grading threshold range of the comprehensive wind and sand damage index, the target area is divided into a severely damaged area, a moderately damaged area, or a slightly damaged area.

3. The three-dimensional monitoring method for the coordinated use of photovoltaic power generation and desert sand control according to claim 2, characterized in that, Based on the wind and sand erosion levels and the layout of photovoltaic desertification control facilities along the target sections of the railway line, a corresponding three-dimensional monitoring strategy is determined, including: Obtain the wind and sand erosion level of the target area; Obtain the layout parameters of the photovoltaic desertification control facility, including the photovoltaic panel tilt angle, row spacing, ground clearance, installation density, and distance between the photovoltaic array and the centerline of the railway line; Based on the wind and sand erosion level and the layout parameters, the sand prevention monitoring area between the photovoltaic power station and the railway is determined; Based on the aforementioned sand control monitoring area, the collaborative configuration scheme of satellite remote sensing monitoring units, UAV aerial survey units, and ground IoT monitoring units, as well as the deployment density and location of each monitoring sub-unit and the differentiated monitoring cycle and data acquisition frequency, are determined.

4. The three-dimensional monitoring method for the synergistic effect of photovoltaic power generation and desert sand control according to claim 3, characterized in that, The determination of the sandstorm prevention monitoring area between the photovoltaic power station and the railway based on the wind and sand erosion level and the layout parameters includes: Obtain the height of the photovoltaic panel above the ground; Obtain the angle between the prevailing wind direction and the railway alignment; Based on the wind and sand damage level, the basic monitoring width is determined as follows: when the wind and sand damage level is a severe damage area, the basic monitoring width is a first preset width value; when the wind and sand damage level is a moderate damage area, the basic monitoring width is a second preset width value; when the wind and sand damage level is a slight damage area, the basic monitoring width is a third preset width value; wherein the first preset width value is greater than the second preset width value, and the second preset width value is greater than the third preset width value. Based on the height of the photovoltaic panel above the ground and the angle between the prevailing wind direction and the railway line, the distance of wind and sand flow around the affected area is calculated. The total monitoring width is obtained by adding the basic monitoring width to the wind and sand obstruction distance; Using the centerline of the railway line as a reference, the total monitoring width is extended upwind of the prevailing wind direction, and the preset downwind width is extended downwind of the prevailing wind direction to determine the sand prevention monitoring area; The photovoltaic array is located within the sand control monitoring area, which covers the area between the rows of photovoltaic panels and the buffer zone between the photovoltaic array and the railway.

5. A three-dimensional monitoring method for the synergistic effect of photovoltaic power generation and desert sand control according to claim 3, characterized in that, Based on the aforementioned sand control monitoring area, the coordinated configuration scheme of satellite remote sensing monitoring units, UAV aerial survey units, and ground-based IoT monitoring units is determined, along with the deployment density and location of each monitoring sub-unit, and the differentiated monitoring cycle and data acquisition frequency, including: The sand control monitoring area is divided into key monitoring sub-areas and general monitoring sub-areas; wherein, the key monitoring sub-areas include the area between photovoltaic panels, sand-prone sections, and the railway main body area; the general monitoring sub-areas are the areas within the sand control monitoring area other than the key monitoring sub-areas; For the key monitoring sub-areas, a ground-based Internet of Things (IoT) monitoring unit is configured for continuous fixed-point monitoring, and an unmanned aerial vehicle (UAV) aerial survey unit is configured for periodic inspections. For the general monitoring sub-area, satellite remote sensing monitoring units are configured for area coverage monitoring, and sparsely deployed ground-based Internet of Things monitoring units are also provided. The deployment density of each monitoring subunit in the ground-based IoT monitoring unit is as follows: In the key monitoring sub-area, the deployment spacing of the wind and sand monitoring subunit is a first deployment distance; the photovoltaic system monitoring subunit is deployed in units of photovoltaic panel strings; and the ecological and soil monitoring subunit is deployed between rows of photovoltaic panels at a second deployment distance. In the general monitoring sub-area, the deployment spacing of each monitoring subunit is a first preset multiple of the key monitoring sub-area; the first deployment distance is less than the second deployment distance. The differentiated monitoring cycle and data collection frequency are as follows: during the sandstorm season, the data collection frequency for key monitoring sub-areas is on the minute level, and the data collection frequency for general monitoring sub-areas is on the hour level; during the non-sandstorm season, the data collection frequency for key monitoring sub-areas is on the hour level, and the data collection frequency for general monitoring sub-areas is on the day level.

6. The three-dimensional monitoring method for the synergistic effect of photovoltaic power generation and desert sand control according to claim 1, characterized in that, The process of quality assessment and correction of the second collaborative monitoring dataset to obtain the target monitoring dataset includes: The second collaborative monitoring dataset is subjected to data integrity verification, time series consistency analysis, and multi-parameter physical correlation constraint verification to obtain quality assessment results. Establish a mapping relationship library between monitoring data error types and collaborative correction strategies. The monitoring data error types include instantaneous wind and sand disturbance error, cumulative dust attenuation error, sensor temperature drift error, communication transmission packet loss error, and multi-source spatiotemporal mismatch error. Based on the error type and error degree identified in the quality assessment results, the corresponding candidate collaborative correction strategy is matched from the mapping relationship library; If the matched candidate collaborative correction strategy is a comparison correction strategy based on a digital twin model, then the second collaborative monitoring dataset is corrected using a preset multi-physics coupled digital twin model to obtain the target monitoring dataset. If the matched candidate collaborative correction strategy is an error compensation correction strategy based on historical data statistics, then the second collaborative monitoring dataset is corrected using a preset multivariate error compensation algorithm to obtain the target monitoring dataset.

7. A three-dimensional monitoring method for the synergistic effect of photovoltaic power generation and desert sand control according to claim 6, characterized in that, If the matched candidate collaborative correction strategy is a comparison correction strategy based on a digital twin model, then a preset multiphysics coupled digital twin model is used to correct the second collaborative monitoring dataset to obtain the target monitoring dataset, including: Error assessment is performed on the meteorological boundary conditions and initial state parameters in the second collaborative monitoring dataset. If the error exceeds a preset confidence threshold, the error compensation and correction strategy based on historical data statistics is switched to the next step. Otherwise, the meteorological boundary conditions and initial state parameters are input into the preset multiphysics coupled digital twin model. The meteorological boundary conditions include wind speed, wind direction, dust concentration, temperature, and humidity. The initial state parameters include the initial dust content on the photovoltaic panel surface, the initial soil moisture content, and the initial vegetation coverage. The preset digital twin model is run to simulate and obtain a theoretical monitoring data set corresponding to the monitoring period. The theoretical monitoring data set includes theoretical wind speed profile, theoretical sediment transport rate, theoretical photovoltaic power generation efficiency, and theoretical soil moisture content. Calculate the difference between the actual monitored values ​​in the second collaborative monitoring dataset and the corresponding theoretical values ​​in the theoretical monitoring dataset to obtain the simulated deviation vector; If the magnitude of the simulated deviation vector is less than the preset deviation tolerance threshold, then the second collaborative monitoring dataset is directly used as the target monitoring dataset. If the magnitude of the simulated deviation vector is greater than or equal to the preset deviation tolerance threshold, then the Kalman filter data assimilation algorithm is used to fuse the actual monitoring value with the theoretical value to obtain the assimilated and corrected monitoring data, which is used as the target monitoring dataset.

8. A three-dimensional monitoring method for the synergistic effect of photovoltaic power generation and desert sand control according to claim 7, characterized in that, If the magnitude of the simulated deviation vector is greater than or equal to a preset deviation tolerance threshold, then a Kalman filter data assimilation algorithm is used to fuse the actual monitored value with the theoretical value to obtain assimilated and corrected monitoring data, including: Based on the model error statistical characteristics of the preset digital twin model, the process noise covariance matrix of the Kalman filter is determined; Based on the measurement error statistical characteristics of the second collaborative monitoring dataset, the measurement noise covariance matrix of the Kalman filter is determined; Using the theoretical monitoring values ​​in the theoretical monitoring data set as predicted observations and the actual monitoring values ​​as observations, the Kalman filter update equation is executed to calculate the optimal state estimate. The optimal state estimate is reconstructed into assimilated and corrected monitoring data in the same format as the second collaborative monitoring dataset to obtain the assimilated and corrected monitoring data.

9. A three-dimensional monitoring method for the synergistic effect of photovoltaic power generation and desert sand control according to claim 6, characterized in that, If the matched candidate collaborative correction strategy is an error compensation correction strategy based on historical data statistics, then a preset multivariate error compensation algorithm is used to correct the second collaborative monitoring dataset to obtain the target monitoring dataset, including: Based on the current dust accumulation rate, photovoltaic panel temperature difference and sandstorm season information corresponding to the second collaborative monitoring dataset, a historical error sample set matching the working conditions is retrieved from the historical error database. Analyze the statistical distribution characteristics of each error value in the historical error sample set, and establish an error probability distribution model under the current operating conditions; Using the error probability distribution model, calculate the expected error estimate for each monitoring parameter in the second collaborative monitoring dataset; A multivariate error compensation algorithm is used to subtract the expected error estimate from the actual monitoring values ​​in the second collaborative monitoring dataset to obtain the error-compensated corrected monitoring data. The spatial continuity of the corrected monitoring data after error compensation is optimized using spatial kriging interpolation to obtain a spatially continuous target monitoring dataset.

10. A three-dimensional monitoring system for the coordinated use of photovoltaic power generation and desert sand control, characterized in that, include: The strategy determination module is used to determine the corresponding three-dimensional monitoring strategy based on the wind and sand erosion level and the layout of photovoltaic desertification control facilities in the target section along the railway line. The execution strategy module is used to execute the three-dimensional monitoring strategy and acquire collaborative monitoring data through the constructed integrated sky-ground monitoring network; the collaborative monitoring data includes meteorological and wind and sand parameters, photovoltaic system operation parameters, ecological vegetation parameters, and railway facility safety parameters. The dust accumulation module is used to obtain the amount of dust accumulation on the surface of the photovoltaic panel at multiple time points during the monitoring period, and to calculate the dust accumulation rate using the time series of the accumulation amount and the average wind speed during the monitoring period. The target data module is used to take all the collaborative monitoring data of the current monitoring cycle as the second collaborative monitoring dataset if the dust accumulation rate is greater than or equal to a preset dust accumulation threshold and the photovoltaic power generation efficiency decay rate is greater than or equal to a first efficiency decay threshold, and to perform quality assessment and correction on the second collaborative monitoring dataset to obtain the target monitoring dataset. Otherwise, the collaborative monitoring data will be directly used as the target monitoring dataset; The trend prediction module is used to obtain the predicted wind and sand movement trend and photovoltaic efficiency degradation trend based on the target monitoring dataset and using the wind-sand-photovoltaic coupled prediction model. The collaborative assessment module is used to conduct a collaborative risk assessment based on the ecological vegetation parameters in the target monitoring dataset, the predicted wind and sand transport trend, and the photovoltaic efficiency degradation trend, and obtain a collaborative risk level. The instruction execution module is used to match a preset collaborative control strategy, generate a collaborative control instruction, and execute it when the collaborative risk level is greater than or equal to a preset risk threshold.

Images