A land wind power monitoring method and system based on remote sensing and GIS

Abstract

The application discloses a land wind power monitoring method and system based on remote sensing and GIS, and belongs to the technical field of wind power monitoring and geographic information processing. The application obtains multi-type image data of a wind farm through satellite remote sensing technology, unmanned aerial vehicle remote sensing technology and the like, the image data is subjected to geometric correction, spectral correction, radiation correction and data integration to generate a geographic information model, and environmental risks such as vegetation, hydrology and soil erosion are monitored; wind turbine sensor data is collected, spatiotemporal double-dimension adaptive weighted data fusion analysis is adopted in combination with remote sensing data, and the health state of a unit is evaluated; a risk quantization model of the environment and equipment is constructed, the wind turbine environmental exposure index, the environmental and equipment time correlation coefficient and the dynamic coupling risk index are calculated, and risk assessment is realized; the application displays the operation result of the wind farm in real time through a GIS platform, provides operation and maintenance decision support for the operation of the wind farm, improves the wind power monitoring precision and efficiency, and can be popularized to other energy and infrastructure monitoring fields.

Description

Technical Field

[0001] This invention relates to the interdisciplinary technical field of onshore wind power monitoring, geographic information processing, and data fusion analysis, specifically to an onshore wind power monitoring method and system based on remote sensing and GIS. Background Technology

[0002] With the global energy structure transformation and increased environmental awareness, wind energy, as a clean and renewable energy source, has been widely promoted and applied. Onshore wind farms have achieved large-scale development due to their advantages such as low construction costs and convenient operation and maintenance. However, as wind farms operate for longer periods and expand in scale, environmental changes in the wind farm area, such as vegetation degradation, soil erosion, hydrological anomalies, and wind turbine equipment failures, have gradually become more prominent, placing higher demands on the continuous and healthy operation of wind farms and the protection of the surrounding ecological environment.

[0003] Currently, traditional onshore wind power monitoring methods mainly rely on ground-based sensors to collect data and regular on-site inspections. This approach has significant drawbacks, such as the limited monitoring range of ground-based sensors, making it difficult to achieve environmental monitoring of the entire wind farm area; high cost and low efficiency of manual on-site inspections; and insufficient real-time and comprehensiveness of monitoring data. Existing technologies often use remote sensing or GIS (Geographic Information System) technologies for localized monitoring, lacking the integration of multiple technologies and failing to clarify the temporal sequence and logical relationships of each monitoring step, leading to chaotic implementation of technical solutions. While some monitoring methods attempt to integrate sensor and remote sensing data, the weights for data integration are often statically set, unable to adaptively adjust according to the spatial topography and real-time meteorological conditions of the wind farm. The parallel logic of time series analysis and spatial analysis is not clearly defined, resulting in low accuracy in assessing the health status of the units. Existing technologies fail to achieve coupled analysis of wind farm environmental changes and equipment operating status, and do not construct coupled risk assessment models according to progressive logic. They can only independently monitor environmental risks and equipment failures, unable to pinpoint the root causes of equipment failures caused by environmental risks, making it difficult to achieve accurate predictive maintenance.

[0004] Remote sensing technology can acquire large-scale, high-resolution surface images, while GIS technology can efficiently process and analyze spatial data. The combination of these two technologies provides a new solution for wind power monitoring. Therefore, there is an urgent need to develop a comprehensive onshore wind power monitoring method that clearly defines the temporal sequence and logical relationships of each step, integrates remote sensing, GIS, and advanced data fusion algorithms, and achieves full-area, real-time, and accurate monitoring of the wind farm environment and equipment. Simultaneously, it should complete the coupled risk assessment of the environment and equipment according to a progressive logic, providing scientific decision support for wind farm operation and maintenance, and improving the level of wind farm operation and management. Summary of the Invention

[0005] This invention addresses the shortcomings of existing onshore wind power monitoring technologies by providing an onshore wind power monitoring method and system based on remote sensing and GIS. This method enables comprehensive and real-time monitoring of wind farm environmental changes and wind turbine equipment status, solves the problems of fixed data fusion weights and the inability to achieve environment-equipment coupling risk assessment, improves the accuracy and efficiency of wind power monitoring, and provides scientific decision support for wind farm operation and maintenance and ecological protection.

[0006] To achieve the above objectives, the technical solution of this invention is as follows: A method for monitoring onshore wind power based on remote sensing and GIS includes the following steps: S1. Acquire multispectral and high-resolution images of the wind farm area using remote sensing technology, including at least one of satellite remote sensing, UAV remote sensing, and aerial remote sensing. The images include at least one of visible light images, infrared images, and radar images. Multispectral images can provide multi-dimensional information on surface vegetation, water bodies, and soil, while high-resolution images can accurately display surface details. The combination of multiple remote sensing technologies and image types enables comprehensive acquisition of image data for the entire wind farm area, laying the foundation for all subsequent monitoring and analysis steps.

[0007] S2. After acquiring the imagery in step S1, GIS technology is used to preprocess the imagery to generate a geographic information system model containing topographic, land use, and vegetation cover information. The preprocessing is performed in the order of geometric correction, spectral correction, radiometric correction, and data integration. Each step serves as the basis for the next, ensuring the accuracy and standardization of the imagery data. Specifically, this includes: 201. Geometric correction: Based on the high-resolution digital elevation model (DEM) and sensor observation data, the remote sensing image is aligned with the geographic coordinate system to eliminate geometric distortion caused by terrain undulation and sensor angle changes, providing a geometrically error-free base image for subsequent spectral correction. 202. Spectral correction: After geometric correction, atmospheric correction models are used to eliminate spectral distortion in the image caused by changes in atmospheric conditions such as wind speed and humidity, ensuring the accuracy of the image spectral data and providing spectrally accurate images for subsequent radiometric correction. 203. Radiometric correction: After completing spectral correction, the sensor response characteristics adjustment algorithm is combined to correct the non-uniformity of sensor response and the changes in radiation intensity in the image, so as to ensure the consistency of radiation brightness in the image and provide a radiometrically unified image for subsequent data integration. 204. Data Integration: After completing geometric, spectral, and radiometric corrections, the corrected image data, digital elevation model (DEM), and geographic coordinate data are integrated into the GIS system to generate a comprehensive geographic information system model. This enables the normalized management of multi-source spatial data and provides complete GIS model support for the environmental analysis in step S3.

[0008] S3. After generating the geographic information system model in step S2, conduct parallel analyses on the model that are independent of temporal sequence and can be performed simultaneously. Identify and monitor three types of environmental changes and potential risks of the wind farm: vegetation changes, hydrological changes, and soil erosion. These three types of analyses can be carried out simultaneously to improve environmental monitoring efficiency. Specific analysis methods include: Vegetation change monitoring utilizes multispectral imagery to calculate the Normalized Difference Vegetation Index (NDVI) or Enhanced Vegetation Index (EVI), compares index values ​​at different times, identifies trends in vegetation cover change and key areas of vegetation reduction, and analyzes potential causes. Hydrological change monitoring combines high-resolution remote sensing images and digital elevation models (DEMs) to calculate the water body index NDWI to extract the water body range, analyze the changes in water area and river flow direction in the wind farm area, and assess water quality indicators such as suspended solids concentration and transparency. Soil erosion monitoring, based on the Universal Soil Loss Equation (USLE) model, combined with high-resolution imagery and DEM data, quantifies soil erosion intensity, locates high-risk areas of soil erosion, and predicts the potential threat of soil erosion to wind farm infrastructure. The Universal Soil Loss Equation (USLE) model is as follows: = , For the region At any moment t The rate of soil erosion, The rainfall erosivity factor was calculated dynamically on a monthly basis using real-time rainfall monitoring data from regional meteorological stations, combined with rainfall intensity and duration, and over time. Updated in real time; The soil erodibility factor is assigned a value based on soil sampling and testing results in the wind farm area, and a fixed value is used for areas with the same soil type. The slope length and slope factor are calculated using slope and slope length formulas by extracting topographic data based on the DEM digital elevation model. The vegetation cover factor is obtained by inverting the NDVI index calculated from multispectral images, and it changes dynamically with the vegetation status. The soil and water conservation measures factors are statically assigned based on the existing soil and water conservation engineering types such as slope protection and vegetation soil stabilization in the wind farm. The calculation or assignment methods for each factor in the USLE model are general standard techniques in the fields of soil erosion monitoring and remote sensing geographic information analysis. , , , , The acquisition of values ​​is a matter of existing technology and will not be elaborated upon here.

[0009] S4. After completing the environmental monitoring in step S3, acquire the sensor data of the wind turbine. Fusion analysis is performed on the sensor data and the remote sensing data from step S1 to assess the health status and performance of the wind turbine. The fusion analysis employs a spatiotemporal dual-dimensional adaptive weighted data fusion mechanism based on time series analysis and spatial analysis. Time series analysis and spatial analysis are parallel analysis methods that can be performed simultaneously without any temporal order. The two methods work together to achieve comprehensive analysis of multi-source data. Specifically: 401. Sensor Data Acquisition: The acquired sensor data of the wind turbine includes key operating parameters such as wind speed, wind direction, rotational speed, vibration, and temperature, providing basic data for equipment operation for unit health assessment; 402. Parallel Analysis Implementation: Time series analysis uses the autoregressive integral moving average model (ARIMA) or long short-term memory network (LSTM) to model the time trend of sensor operating parameters, predict equipment status changes, and detect anomalies; spatial analysis uses the Kriging interpolation method to predict spatial data, and combines buffer analysis and spatial overlay analysis to assess the potential impact of the wind turbine's surrounding environment on equipment operation. Both are implemented simultaneously to achieve in-depth data analysis from both time and spatial dimensions. 403. Spatiotemporal dual-dimensional adaptive weighted fusion: The formula for calculating the fusion weight is: = ; in: For the first Typhoon turbine units, indexed by wind turbine; For the first The data source / feature type is the basic unit index for fusion analysis, representing a certain type of original data source or a certain feature indicator that is a general index of the same dimension as m. t: Current monitoring time, representing the time attribute of the data; The spatial dimension weight of the j-th wind turbine corresponding to the k-th data source / feature type is determined by the geographical location characteristics of the wind turbine. The time / context dimension weight of the j-th wind turbine corresponding to the k-th data source / feature type at time t is dynamically adjusted based on real-time meteorological conditions and historical operating data. m: A general traversal index for data source / feature type, which is the same dimension index as k. It replaces k to complete the summation calculation and is only used to traverse all fusion units during the summation calculation to avoid confusion with k in the molecule. M: The total number of data sources / feature types participating in this fusion analysis, that is, a total of M types of original data sources / feature indicators were included in this fusion; Summation symbol This indicates that for all M types of data sources / feature types participating in the fusion, m = 1, 2, ... M is used for summation calculation; Denominator: Normalization term, ensuring that the sum of the weights of all data sources / feature types participating in the fusion for the j-th wind turbine at time t is 1, so that the weight allocation meets the probability distribution requirements and avoids the distortion of analysis results caused by the weight superposition exceeding 1; The data source / feature type includes two categories: first, remote sensing raw data sources and sensor raw data sources; second, environmental / equipment feature indicators extracted from the raw data sources. The data source feature type is the basic unit of fusion analysis, encompassing both raw data sources and feature indicators. It is a general term for remote sensing data, sensor data, and various analytical indicators extracted from them. Raw data sources include satellite remote sensing imagery, UAV radar imagery, wind speed sensor data, vibration sensor data, etc. Feature indicators include NDVI index, NDWI index, and soil erosion rate extracted from remote sensing data, and vibration intensity and mean rotational speed extracted from sensor data. 'k' is not a single data source / feature type, but a collective term for both: this invention uses data source / feature type as a unified technical term, encompassing raw data sources and feature indicators extracted from them. 'k' is the index of this unified term, representing the k-th type of raw data source / feature indicator. In the fusion analysis of this invention, both raw data sources and feature indicators are basic fusion units, and the unified index simplifies formula calculations.

[0010] S5. After completing the unit assessment in step S4, based on the environmental monitoring results in step S3 and the unit assessment results in step S4, construct an environmental-equipment coupling risk quantification model and complete the coupling risk assessment. Then, display the wind farm's operating status, environmental monitoring results, unit assessment results, and coupling risk assessment results in real time through the GIS platform to provide decision support for wind farm operation and maintenance. The platform's display function and decision support function are parallel functions implemented simultaneously.

[0011] The construction and evaluation of the environment-equipment coupling risk quantification model are carried out in a progressive order, with the results of the previous step serving as the basis for the implementation of the next step. Specifically: 501. Calculation of Wind Turbine Environmental Exposure Index: Based on the soil erosion rate calculated using the USLE model (Universal Soil Loss Equation), and combined with the spatial location of the wind turbine, calculate the wind turbine environmental exposure index. : = ; Where j represents a specific wind turbine, and i represents the environmental grid or area surrounding the wind turbine. This refers to a buffer zone, such as a radius of 100 meters, centered on the wind turbine. The soil erosion rate of region i at time t; this index characterizes the environmental pressure borne by a single wind turbine and provides an environmental data basis for subsequent correlation analysis; This means that only when the spatial distance to the j-th wind turbine is less than a preset threshold... Soil erosion rate for all environmental grids / regions i. Summation calculation.

[0012] 502. Calculate the environmental-equipment time correlation coefficient: After calculating the wind turbine environmental exposure index, perform a correlation analysis between the wind turbine environmental exposure index time series and the wind turbine vibration intensity time series, and calculate the Pearson correlation coefficient. :

[0013] in, The vibration intensity time series of wind turbine j was obtained from the SCADA system. Let the covariance of the two be , for standard deviation for Standard deviation; when When the absolute value is greater than the preset value, it is determined that there is a significant correlation between changes in the wind farm environment and the corresponding wind turbine operating status. This coefficient provides a basis for the coupling relationship in the subsequent calculation of the dynamic coupling risk index; regarding The absolute value is greater than a preset value, which is generally set to 0.7. The preset value is a threshold set based on common sense in the field of statistical analysis and in combination with the actual situation of wind power monitoring projects. | 0.7 indicates a strong linear correlation, i.e., a significant correlation; 0.3... | | 0.7, moderate linear correlation, | | 0.3, no significant correlation.

[0014] The SCADA (Supervisory Control and Data Acquisition) system is an automated system used in the industrial field, especially in the energy sector, for real-time data acquisition, equipment monitoring, and remote control. The SCADA system provides core equipment parameters such as vibration intensity for the time series analysis in step S4, and provides equipment-side data for the environment-equipment correlation analysis in step S5. The SCADA system, together with the remote sensing data in step S1 and the environmental monitoring data such as soil erosion and vegetation change in step S3, constitutes the dual-dimensional data foundation of environment and equipment, supporting the subsequent coupled risk assessment. Let covariance be the variance of two time series, for the first series: The environmental exposure index time series of wind turbines refers to the set of environmental exposure indices of the j-th wind turbine at different times t, such as the daily indices over the past month. Value, second sequence In other words, the time series of wind turbine vibration intensity refers to the set of vibration intensities of the j-th wind turbine at the same time dimension, and... The monitoring time t corresponds one-to-one, such as the daily vibration intensity values ​​within the past month; covariance is a fundamental statistic for measuring the degree of linear correlation between two variables or sequences, and its calculation formula is... = Standard deviation = ,in , Let n be the mean of the two time series. and The total number of paired data or the total number of data in a sequence.

[0015] absolute value of Pearson correlation coefficient | | 0.7 indicates a significant correlation between environmental changes, such as increased soil erosion, and changes in equipment operating conditions, such as increased vibration intensity. This suggests that environmental risks may trigger equipment failures, such as soil erosion causing tower foundation loosening, which in turn leads to increased vibration. | A value of 0.7 indicates a low correlation between the two factors; changes in equipment condition are more likely caused by other factors such as equipment wear and tear, or instantaneous weather fluctuations, and are not significantly related to environmental changes. If no calculation is performed... The value only indicates that severe soil erosion and large unit vibration, but it is impossible to determine whether the two are related or whether the large vibration is caused by soil erosion. After determining a significant correlation, it can be clearly identified that soil erosion is the potential root cause of large unit vibration. Subsequently, the DCRI index can accurately quantify this coupled risk and provide targeted solutions for operation and maintenance, such as foundation reinforcement and soil erosion control, rather than blind maintenance.

[0016] 503. Calculate the dynamic coupling risk index: After calculating the correlation coefficient, construct a dynamic coupling risk index model based on the wind turbine environmental exposure index and the time correlation coefficient. = + + ; in, Let be the dynamic coupling risk index of the j-th wind turbine. The normalized wind turbine environmental exposure index. The normalized equipment vibration level, , , These are the weighting coefficients. , The weights of environmental factors are parallel weights that are set as needed, without any sequential order. Equipment factor weights Based on fixed weights, environmental risk and equipment status are the two core sources of coupled risk. Weighting coefficients. Based on the time correlation coefficient The size is dynamically adjusted. The larger the absolute value, The larger the value, the stronger the correlation between the environment and the equipment. , The setting and The adjustment is to be implemented as a parallel operation carried out independently. If The low value indicates that environmental risk is not significantly related to the unit's operating status, even if Even with a high value, the dynamic coupling risk index will not be significantly increased. in: = , This represents the initial environmental exposure index of the j-th wind turbine; The original structure of the wind farm unit Minimum value; Original units in wind farms Maximum value. = , This represents the initial vibration intensity of the j-th fan. The original structure of the wind farm unit Minimum value; Original units in wind farms Maximum value.

[0017] The GIS platform provides integrated services for displaying monitoring results and supporting decision-making. The platform offers tools for dynamic trend analysis, environmental change monitoring, wind turbine health status assessment, predictive maintenance, spatial analysis, multi-source data fusion, and risk management. Each tool is a parallel functional module that can be called independently without any time sequence. Users can call any tool independently according to their needs. The spatial analysis includes buffer analysis, overlay analysis, and 3D visualization. Buffer analysis, overlay analysis, and 3D visualization are parallel analysis operations that can be performed simultaneously without any time sequence. Buffer analysis uses GIS software to generate polygonal areas of a certain width around spatial objects, such as wind turbine locations, to define the scope of influence, such as the environmental risk zone within 100 meters of a wind turbine. Overlay analysis overlays multiple spatial data layers, such as soil erosion layers, wind turbine location layers, and vegetation cover layers, and identifies the spatial relationships between layers through logical operations such as intersection and union, such as the overlapping area between high-risk soil erosion areas and wind turbine locations. 3D visualization is based on DEM data and remote sensing imagery to construct a 3D terrain model of the wind farm, showing the spatial relationship between wind turbines, terrain, and environmental risk zones, such as the vertical height difference between hillside wind turbines and soil erosion areas. It is a basic module of the GIS platform's display function.

[0018] The platform supports real-time data updates and dynamic monitoring, enabling multi-level and multi-dimensional data display and user interaction. Wind farm operation and maintenance personnel can formulate scientific operation and maintenance plans and ecological protection measures based on the monitoring results and analysis data displayed on the platform, achieving predictive maintenance and accurate decision-making.

[0019] To implement the above-mentioned onshore wind power monitoring method based on remote sensing and GIS, this invention also provides a supporting system. This system, through modular design, enables the coordinated execution of each step, with clear logical relationships between modules and a complete data transmission link. The specific architecture is as follows: Data transmission logic: Remote sensing image acquisition module → GIS preprocessing module → Environmental risk monitoring module → Multi-source data fusion module / coupled risk assessment module; Sensor data acquisition module → Multi-source data fusion module; Multi-source data fusion module → Coupled risk assessment module; Coupled risk assessment module / Environmental risk monitoring module / Multi-source data fusion module → GIS display and decision-making module; Execution sequence logic: The remote sensing image acquisition module and the sensor data acquisition module can start data acquisition synchronously; the GIS preprocessing module starts after the remote sensing image acquisition module completes data acquisition; the environmental risk monitoring module starts after the GIS preprocessing module generates the geographic information system model; the multi-source data fusion module starts after both the sensor data acquisition module and the environmental risk monitoring module complete data output; the coupled risk assessment module starts after the multi-source data fusion module completes the unit assessment; the GIS display and decision-making module can display the results synchronously in real time after the coupled risk assessment module outputs its results. Module collaboration logic: Each module independently implements its corresponding function steps, forming a collaborative closed loop through data transmission. The GIS preprocessing module provides the data foundation for the environmental risk monitoring module. The multi-source data fusion module needs to receive sensor data, remote sensing data, and environmental monitoring results simultaneously to complete the fusion analysis. The coupled risk assessment module relies on environmental monitoring results and unit assessment results to quantify risks. The GIS display and decision-making module integrates the output results of all modules to provide operation and maintenance support.

[0020] The beneficial effects of this invention are: 1. This invention combines satellite, UAV, and aerial remote sensing technologies to acquire multi-type image data of the entire wind farm area, and uses GIS technology to complete image preprocessing in the steps of geometric correction, spectral correction, radiometric correction, and data integration, so as to realize the full-area, real-time monitoring of the wind farm environment. This solves the problems of incomplete data acquisition, poor real-time performance, and high cost of traditional monitoring methods, and provides a comprehensive and accurate spatial data foundation for wind farm environmental risk identification.

[0021] 2. This invention proposes a spatiotemporal dual-dimensional adaptive weighted data fusion mechanism, which clarifies the parallel logic of spatial dimension weights and time / context dimension weights, as well as the parallel logic of time series analysis and spatial analysis. The data fusion weights are dynamically adjusted according to the spatial geographical location characteristics of wind turbines and real-time meteorological conditions, replacing the traditional static weight setting. This makes the data fusion process more in line with the actual working conditions of wind farms and improves the accuracy of wind turbine health status and performance evaluation.

[0022] 3. This invention constructs an environment-equipment coupled risk quantification model, which couples macro-level geographical environmental evolution, such as soil erosion, with micro-level equipment operating conditions, such as vibration intensity, through an algorithm. This achieves a breakthrough from listing phenomena to diagnosing root causes, and can accurately locate the correlation between environmental risks and equipment failures, providing a basis for predictive maintenance of wind farms.

[0023] 4. This invention uses a GIS platform to achieve integrated and visualized display of wind farm operation status, environmental monitoring results, unit evaluation results, and coupled risk assessment results. It clarifies the parallel logic of each functional module and analysis operation of the platform and is equipped with a variety of independently callable analysis tools, providing comprehensive decision support for wind farm operation and maintenance and ecological protection, and improving the operation management level and ecological protection capabilities of wind farms.

[0024] 5. The monitoring method of the present invention is not only applicable to onshore wind power monitoring, but can also be extended to the monitoring and management of other energy facilities such as photovoltaics and hydropower, as well as infrastructure such as roads and bridges, and has broad application prospects. Attached Figure Description

[0025] Figure 1 This is a schematic diagram of the method flow of the present invention. Detailed Implementation

[0026] To make the purpose, technical solution, and advantages of the invention clearer, the invention will be further described below with reference to the accompanying drawings.

[0027] like Figure 1 As shown, an onshore wind power monitoring method based on remote sensing and GIS specifically includes: S1. Remote sensing image acquisition: A combination of satellite remote sensing, such as Sentinel-2 satellite, UAV remote sensing, and aerial remote sensing is used to acquire multispectral and high-resolution images of the entire mountain wind farm and its surrounding 5km range. Satellite remote sensing acquires visible light and infrared light images, UAV remote sensing acquires radar images, and aerial remote sensing acquires high-resolution visible light images. The images cover the wind farm's generating units and infrastructure.

[0028] S2. GIS Image Preprocessing and Model Building: After completing step S1, the acquired images are preprocessed using ArcGIS software in a progressive manner, including geometric correction, spectral correction, radiometric correction, and data integration. 201. Geometric Correction: Geometric correction is performed based on the DEM data and sensor observation data of the area to eliminate geometric distortions caused by terrain undulations and drone aerial photography angles, so that the image is accurately aligned with the WGS84 geographic coordinate system. 202. Spectral Correction: After completing geometric correction, spectral correction is performed using the atmospheric correction model Flaash to eliminate spectral distortion caused by atmospheric scattering and absorption. 203. Radiometric Correction: After completing spectral correction, radiometric correction is performed in conjunction with the sensor response characteristics to ensure the consistency of image radiance. 204. Data Integration: After completing the radiometric correction, the corrected image data is integrated with the DEM data and wind farm geographic coordinate data. In other words, multi-source spatial data such as images, DEM, and geographic coordinates are integrated into the GIS system to generate a comprehensive GIS model that includes topography, land use, and vegetation cover.

[0029] S3. Environmental Change and Potential Risk Monitoring: After completing step S2, conduct simultaneous parallel analysis of vegetation, hydrology, and soil erosion on the GIS integrated model to improve monitoring efficiency. Vegetation change monitoring: The NDVI index of multispectral images in the GIS model was calculated using ENVI software. By comparing the NDVI index data of the past year, it was identified that the vegetation coverage around wind farm units 3 and 7 decreased by about 20%, which was found to be caused by the construction of the maintenance road for the units. Hydrological change monitoring: The water body index NDWI was calculated to extract the water body range within the wind farm. It was found that during the flood season, there was water accumulation in the depression where Units 10 and 15 are located, with a water accumulation area of ​​about 500 square meters. Soil erosion monitoring: Based on the USLE model, combined with DEM data and high-resolution imagery, the soil erosion rate in various areas of the wind farm was calculated, and the eastern slope of the wind farm was identified as a high-risk area for soil erosion. 15t / (hm) 2 •a) This is far higher than the average level of 8t / (hm²) for wind farms. 2 •a).

[0030] Normalized Difference Vegetation Index (NDVI) and Normalized Difference Water Index (NDWI) are common indicators for remote sensing monitoring of vegetation and water bodies. Their definitions and calculation methods are existing technologies in the field and will not be elaborated here.

[0031] S4. Multi-source data fusion and unit health assessment: After completing step S3, perform this step, with time series analysis and spatial analysis implemented simultaneously and in parallel. 401. Sensor Data Acquisition: Acquire SCADA sensor data from wind farm units, including parameters such as wind speed, wind direction, rotational speed, tower vibration, and generator temperature; 402. Parallel Analysis Implementation: Time series analysis and spatial analysis were carried out simultaneously. The time series analysis used an LSTM network to model the sensor data and detected a slight upward trend in the tower vibration intensity of Units 7 and 10. The spatial analysis used Kriging interpolation to analyze the remote sensing data and found that vegetation degradation around Unit 7 and water accumulation around Unit 10 both affected the operating environment of the units. This invention employs the Autoregressive Integral Moving Average (ARIMA) model or the Long Short-Term Memory (LSTM) network to model the time trends of sensor operating parameters. ARIMA / LSTM is a standard method for equipment status prediction and time series anomaly detection, and is commonly used in the wind power industry for trend analysis and fault early warning of wind turbine sensor parameters such as vibration, speed, and temperature. Its modeling process and parameter selection methods are existing technologies. Kriging interpolation is a common method for spatial data interpolation, widely used in spatial data completion and distribution prediction in the field of remote sensing / GIS. Its interpolation application for spatial heterogeneity of wind farms, such as topography, wind speed, and soil erosion, is a routine operation. Buffer analysis and spatial overlay analysis are basic spatial analysis tools of GIS platforms, used in wind power monitoring to assess the spatial impact range of the surrounding environment of wind turbines, such as vegetation degradation, water accumulation, and soil erosion. Their operation methods are existing technologies. Comparing the NDVI index of the area surrounding Unit 7 at different times, such as the monthly / quarterly data of the past year, when the NDVI index continues to decline and the magnitude is significant, such as exceeding 10%, it can be determined as vegetation degradation. For the depression where Unit 10 is located, the DEM shows a low altitude. The NDWI extraction results show that there is a significant water body during the flood season. Combining the monitoring period, such as the flood season and the depression, it is determined that there is water accumulation in the surrounding area.

[0032] 403. Spatiotemporal Adaptive Weighted Fusion: Spatial and temporal / contextual dimension weights were collaboratively determined. A higher spatial weight was assigned to the vegetation NDVI index of Unit 7, and a higher spatial weight to the NDWI index of Unit 10. Simultaneously, due to the flood season, the temporal / contextual dimension weights were independently adjusted, increasing the temporal weight of remote sensing soil moisture data for all units and decreasing the weight of ground anemometer data. The weights of each data source were calculated according to the fusion weight formula. The post-fusion assessment concluded that the health status of Units 7 and 10 was "warning," while the health status of the remaining units was "normal."

[0033] Spatial Dimension Weight Based on the geographical location characteristics of the wind turbine, such as the terrain, slope, and windward / leeward side, the spatial weighting is set. For example, turbines located in depressions are assigned a higher spatial weight to the NDWI index (a characteristic indicator); turbines located on the windward side of ridges are assigned a higher spatial weight to the raw data source, such as wind speed sensor data. Time / contextual dimension weighting is also considered. The weighting is dynamically adjusted based on real-time weather conditions and historical data. For example, during periods of continuous heavy rainfall, the time weight of remote sensing soil moisture data (original data source) and soil erosion rate (feature indicator) is increased, while the weight of ground anemometer data is decreased. During windy weather, the weight of wind speed sensor data (original data source) and tower vibration intensity (feature indicator) is increased. In actual engineering, the weight values ​​for both types are assigned in a 0-1 range, such as 0.7-1.0 for high weight, 0.3-0.6 for medium weight, and 0-0.2 for low weight. The sum of the fused weights of all data sources for a single wind turbine is 1.

[0034] S5. Coupled Risk Assessment and GIS Visualization: After completing step S4, construct the coupled risk model according to the progressive steps, and then realize the parallel functions of display and decision support through the GIS platform: 501. Calculate the wind turbine environmental exposure index: Based on the soil erosion data in S3, and setting the spatial proximity range D to 100 meters, the environmental exposure index of units 2 and 3 on the east slope is calculated. They are 12.5 t / (hm) 2 •a) 13.2t / (hm 2 •a) represents the highest value for a wind farm; 502. Calculate the time correlation coefficient: After completing step 501, calculate the time correlation coefficient for Units 2 and 3. The values ​​were 0.78 and 0.82 respectively, both greater than 0.7, indicating a significant positive correlation between soil erosion and unit vibration intensity. 503. Calculate the dynamic coupling risk index: After completing step 502, set the weighting coefficients as needed. , ,according to Dynamically adjust γ, set =0.3、 =0.3, Unit 2 γ=0.4, Unit 3 γ=0.45, and the calculated DCRI2 of Unit 2 is 0.81 and the DCRI3 of Unit 3 is 0.85, both of which are at the high risk level.

[0035] In this invention, calculation The primary goal is to achieve a quantitative assessment of the coupled risks of the environment and equipment, addressing the limitation of existing technologies that can only independently monitor environmental risks and equipment failures, failing to quantify their coupled impact. Traditional monitoring can only identify environmental risks such as severe soil erosion or equipment problems such as excessive unit vibration, but cannot assess the comprehensive risk resulting from their combined effect. Calculate the normalized environmental exposure index Normalized equipment vibration level and the correlation coefficient between environment and equipment By integrating them into a single quantitative index, a single value can reflect the coupled impact of two types of risks; secondly, through... The system accurately distinguishes risk levels, providing a clear basis for operational and maintenance decisions. The magnitude of the value directly corresponds to the risk level of the wind turbine, with high risk such as Unit 2. =0.8, Unit 3 =0.85, Units 2 and 3 require priority maintenance, such as reinforcing the tower foundations combined with soil erosion control; Normal risk: If the value is below the high-risk threshold, no emergency maintenance is required; routine monitoring is sufficient. This type of quantitative result eliminates the need for subjective judgment by maintenance personnel, allowing them to directly assess the situation. Prioritizing numerical values ​​in the development of operation and maintenance plans resolves the issues of ambiguous risks and disorganized operations; furthermore, The value calculation reflects dynamic coupling characteristics and is adapted to real-time operating conditions of wind farms. The dynamics of the value are reflected in two dimensions, ensuring that the risk assessment is in line with the actual working conditions in real time, and the time dynamics. With soil erosion rate Dynamic updates The vibration data changes in real time with the SCADA system; the coupled dynamics, i.e., the Pearson correlation coefficient. It dynamically adjusts according to the strength of the correlation between the environment and the equipment, and The larger the value, the higher the weighting coefficient. The higher, the more The value can reflect the impact of the environment-equipment correlation on risk in real time.

[0036] Risk grading falls under the category of flexible adjustments in engineering practice. Different wind farms have significantly different operating conditions, such as terrain complexity, turbine models, and varying maintenance standards. Therefore, the risk level thresholds need to be set based on the actual scenario. Mountain wind farms can High risk; plain wind farms have lower environmental risk. This is considered high-risk. In actual engineering, the following steps are used to establish an adaptation. The numerical value-risk level correspondence is collected from wind farms with generating units. Calculation results, such as for 10 units The values ​​are 0.32, 0.45, 0.58, 0.63, 0.71, 0.76, 0.81, 0.85, 0.88, and 0.92, respectively. Based on the actual operating status of the unit, such as whether there is abnormal vibration and whether the environmental risk is clear, thresholds are divided into high / medium / low levels. For example, using 0.5 and 0.8 as the dividing points, less than 0.5 is low risk, 0.5-0.8 is medium risk, and greater than 0.8 is high risk. If subsequent monitoring... If the corresponding threshold is exceeded, the corresponding level of operation and maintenance response will be triggered, such as prioritizing the reinforcement of the foundation of high-risk units in conjunction with environmental governance.

[0037] GIS Platform Application: Environmental monitoring results, unit health assessment results, and coupled risk assessment results are imported into the GIS platform, simultaneously realizing display and decision support functions: Vegetation degradation areas, high-risk soil erosion areas, and waterlogging areas are marked on the platform, and different colors are used to indicate the unit health status and coupled risk level; The predictive maintenance tools of the platform are independently invoked to formulate operation and maintenance plans for vegetation restoration and waterlogging drainage for Units 7 and 10, and for tower foundation reinforcement and slope soil and water conservation for Units 2 and 3; The predictive maintenance tools are a standard functional module of the wind power operation and maintenance management system. When calling the spatial analysis tools, buffer zone analysis, overlay analysis, and 3D visualization are carried out simultaneously to accurately locate the spatial relationship between risk areas and units.

[0038] After six months of operation and maintenance, the vegetation cover of the wind farm has been restored to its original level, and the soil erosion rate has been reduced to 9 t / (hm²). 2 •a) The problem of water accumulation in the depression was resolved, and the vibration intensity of Units 2, 3, 7, and 10 all returned to normal range. This embodiment, by strictly following the logical relationships of the steps defined in this invention, makes the monitoring process clear and efficient, verifying the accuracy, practicality, and operability of the monitoring method of this invention.

[0039] To implement the above monitoring method, this embodiment provides a supporting system, and the specific configuration and operating logic of each module are as follows: The remote sensing image acquisition module is equipped with Sentinel-2 satellite remote sensing equipment, multi-rotor UAV remote sensing equipment, and fixed-wing aerial remote sensing equipment. The satellite remote sensing equipment acquires visible light and infrared light images, the UAV remote sensing equipment acquires radar images, and the aerial remote sensing equipment acquires high-resolution visible light images. The images cover the wind farm's generating units and a surrounding 5km range. After acquisition, the image data is transmitted to the GIS preprocessing module via a communication network. The GIS preprocessing module, which is a server device that deploys ArcGIS software, receives remote sensing image data and performs preprocessing in the order of geometric correction → spectral correction → radiometric correction → data integration: geometric correction and alignment with the WGS84 coordinate system are completed based on regional DEM data and sensor observation data; spectral correction is completed using the Flaash atmospheric correction model; radiometric correction is completed by combining sensor response characteristics; and finally, the corrected image is integrated with DEM data and wind farm geographic coordinate data to generate a comprehensive GIS model, which is stored in the GIS database and transmitted to the environmental risk monitoring module. The sensor data acquisition module deploys wind speed sensors, wind direction sensors, speed sensors, vibration sensors, and temperature sensors in each wind turbine. The sensor data is collected through the wind farm SCADA system and uploaded to the multi-source data fusion module at a frequency of minutes. The environmental risk monitoring module deploys ENVI software and USLE model calculation tools. After receiving the GIS integrated model, it simultaneously conducts three types of analysis: using ENVI software to calculate the NDVI index to compare vegetation changes, calculating the NDWI index to extract water body range and analyze hydrological changes, and quantifying soil erosion intensity based on the USLE model to identify high-risk areas such as vegetation degradation around Units 3 and 7, water accumulation around Unit 10, and soil erosion on the eastern hillside. The above environmental monitoring results are then transmitted to the multi-source data fusion module and the coupled risk assessment module. The multi-source data fusion module deploys a data fusion server to receive sensor data, remote sensing image data, and environmental monitoring results from the SCADA system. Simultaneously, it performs time series analysis and spatial analysis through LSTM network modeling. The spatial analysis uses Kriging interpolation combined with buffer analysis. The weights of each data source are calculated through a spatiotemporal dual-dimensional adaptive weighted fusion formula. The assessment results show that the health status of Units 7 and 10 is in the warning stage, while the health status of the remaining units is normal. The unit assessment results are then transmitted to the coupled risk assessment module. The risk assessment module is coupled with a risk assessment server. After receiving environmental monitoring results and unit assessment results, the server calculates the environmental exposure index of the wind turbines of units 2 and 3 in a progressive order, setting D=100 meters. 2 •a) 13.2t / (hm 2 •a) Calculate the time correlation coefficients of Units 2 and 3. The values ​​are 0.78 and 0.82 respectively; [Settings] =0.3、 =0.3, Unit 2 γ=0.4, Unit 3 γ=0.45, resulting in DCRI2 and DCRI3 being 0.81 and 0.85 respectively, which are high risks. The coupled risk assessment results are then transmitted to the GIS display and decision-making module. The GIS display and decision-making module deploys GIS display terminals, including a monitoring screen and client. After receiving various assessment results, it marks vegetation degradation areas, waterlogging areas, and high-risk areas for soil erosion on the platform. Different colors are used to indicate the health status and coupled risk level of the units. Maintenance personnel can independently call predictive maintenance tools to formulate vegetation restoration and waterlogging drainage plans for units 7 and 10, and foundation reinforcement and soil and water conservation plans for units 2 and 3, thus achieving precise operation and maintenance decisions.

[0040] The system of this invention, through modular design, ensures that the execution logic and methods of each step are fully matched, and that data transmission between modules is real-time and reliable. It can realize the full-area, real-time and accurate environmental and equipment monitoring of wind farms, effectively improving the operation and management level of wind farms.

[0041] The onshore wind power monitoring method based on remote sensing and GIS of this invention achieves industrial application through remote sensing equipment, GIS software, data processing server, wind turbine sensors, and GIS display platform. The execution process is clear and highly operable, and it can provide onshore wind farms with full-area, real-time, and accurate environmental and equipment monitoring services. It effectively improves the operation and management level and ecological protection capabilities of wind farms, reduces operation and maintenance costs, and reduces equipment failure rate, and has significant industrial application value and market prospects.

[0042] Finally, it should be noted that the contents not described in detail in this specification belong to the prior art known to those skilled in the art. The above descriptions are merely preferred embodiments of the present invention and are not intended to limit the present invention. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art can still modify the technical solutions described in the foregoing embodiments or make equivalent substitutions for some of the technical features. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A method for monitoring onshore wind power based on remote sensing and GIS, characterized in that, Includes the following steps: S1. Obtain multispectral and high-resolution images of the wind farm area through remote sensing technology, wherein the remote sensing technology includes at least one of satellite remote sensing, UAV remote sensing and airborne remote sensing, and the images include at least one of visible light images, infrared light images and radar images. S2. Use GIS technology to preprocess the images obtained in step S1 to generate a geographic information system model containing topographic, land use and vegetation cover information. The preprocessing includes geometric correction, spectral correction and radiometric correction. The preprocessed data is stored in the GIS system. S3. Analyze the geographic information system model generated in step S2 to identify and monitor environmental changes and potential risks of the wind farm, including vegetation changes, hydrological changes and soil erosion. S4. Acquire sensor data of the wind turbine, and perform fusion analysis on the sensor data and the remote sensing data in step S1 to evaluate the health status and performance of the wind turbine. The fusion analysis adopts spatiotemporal dual-dimensional adaptive weighted data fusion based on time series analysis and spatial analysis. S5. Based on the environmental monitoring results of step S3 and the unit assessment results of step S4, construct an environment-equipment coupled risk quantification model for risk assessment. Use a GIS platform to display the wind farm's operating status, environmental monitoring results, unit assessment results, and coupled risk assessment results in real time, providing decision support for wind farm operation and maintenance.

2. The onshore wind power monitoring method based on remote sensing and GIS according to claim 1, characterized in that: In step S2, the geometric correction includes aligning the remote sensing image with the geographic coordinate system based on the high-resolution digital elevation model (DEM) and sensor observation data, eliminating geometric distortions caused by terrain undulations and sensor angle changes. The spectral correction includes using an atmospheric correction model to eliminate spectral distortions in the image caused by changes in atmospheric conditions, ensuring the accuracy of the image's spectral data. The radiometric correction includes combining sensor response characteristic adjustment algorithms to correct sensor response inhomogeneities and changes in radiometric intensity in the image, achieving consistent radiometric brightness. The data integration includes unifying the geometrically, spectrally, and radiometrically corrected image data with the DEM and geographic coordinate data into a GIS system to generate a comprehensive geographic information system model.

3. The onshore wind power monitoring method based on remote sensing and GIS according to claim 1, characterized in that, The analysis methods for the geographic information system model in step S3 include: Vegetation change monitoring utilizes multispectral images to calculate the Normalized Difference Vegetation Index (NDVI) or Enhanced Vegetation Index (EVI), and compares the index values ​​at different times to identify vegetation cover changes and areas of vegetation reduction. Hydrological change monitoring combines high-resolution remote sensing images and digital elevation models (DEM) to calculate the water body index NDWI to extract the water body range and analyze the water body area, river flow direction and water quality changes in the wind farm area. Soil erosion monitoring, based on the USLE model of soil loss equation, combined with high-resolution imagery and DEM data, quantifies soil erosion intensity and locates high-risk areas of soil erosion.

4. The onshore wind power monitoring method based on remote sensing and GIS according to claim 3, characterized in that, The USLE model for soil loss equation is as follows: = , For the region At any moment t The rate of soil erosion, As the erosivity factor of rainfall, As a soil erodibility factor, For slope length and slope factor, For vegetation cover factor, Factors related to soil and water conservation measures.

5. The onshore wind power monitoring method based on remote sensing and GIS according to claim 1, characterized in that: The sensor data of the wind turbine in step S4 includes wind speed, wind direction, rotational speed, vibration, and temperature parameters; the time series analysis includes using an autoregressive integral moving average model (ARIMA) or a long short-term memory network (LSTM) to model the time trend of sensor operating parameters, predict equipment status changes, and detect anomalies; the spatial analysis includes using the Kriging interpolation method to predict spatial data, and combining buffer analysis and spatial overlay analysis to assess the environmental impact around the wind turbine.

6. The onshore wind power monitoring method based on remote sensing and GIS according to claim 1, characterized in that, The formula for calculating the fusion weight of the spatiotemporal dual-dimensional adaptive weighted data fusion mechanism in step S4 is as follows: = ； in: Indicates the first Typhoon generator sets; Indicates the first Data source / feature type, serving as the basic unit index for fusion analysis, and... m It is a general index for the same dimension; t represents the current monitoring time; This represents the spatial dimension weight of the k-th data source / feature type corresponding to the j-th wind turbine; This represents the time / contextual dimension weight of the j-th wind turbine corresponding to the k-th data source / feature type at time t; m represents the general traversal index of the data source / feature type, replacing k to complete the summation calculation; M represents the total number of data sources and feature types participating in this fusion analysis.

7. The onshore wind power monitoring method based on remote sensing and GIS according to claim 1, characterized in that, The construction and evaluation of the environment-equipment coupling risk quantification model in step S5 includes:

501. Calculation of Wind Turbine Environmental Exposure Index: Based on the soil erosion rate calculated using the USLE model (Universal Soil Loss Equation), and combined with the spatial location of the wind turbine, calculate the wind turbine environmental exposure index. : = ； Where j represents a specific wind turbine, and i represents the environmental grid or area surrounding the wind turbine. The area surrounding the wind turbine. Let be the soil erosion rate of region i at time t; 502. Calculate the environment-equipment time correlation coefficient: After completing step 501, perform a correlation analysis between the wind turbine environmental exposure index time series and the wind turbine vibration intensity time series to calculate the environment-equipment time correlation coefficient. ； in, The vibration intensity time series of wind turbine j is given. For covariance, for standard deviation for Standard deviation; 503. Calculate the dynamic coupling risk index: After completing step 502, calculate the dynamic coupling risk index based on the wind turbine environmental exposure index and time correlation coefficient. = + + ; in, Let be the dynamic coupling risk index of the j-th wind turbine. The numerical value corresponds to the risk level of the wind turbine; The normalized wind turbine environmental exposure index. The normalized equipment vibration level, , , These are the weighting coefficients.

8. A system for implementing the onshore wind power monitoring method based on remote sensing and GIS as described in any one of claims 1-7, characterized in that, The system includes: The remote sensing image acquisition module is used to perform step S1 as described in claim 1, and to acquire multispectral images and high-resolution images of the wind farm area through at least one of satellite remote sensing equipment, UAV remote sensing equipment and airborne remote sensing equipment. The images include at least one of visible light images, infrared light images and radar images. The acquired image data is transmitted to the GIS preprocessing module. The GIS preprocessing module is communicatively connected to the remote sensing image acquisition module and is used to execute the S2 step described in claim 1. It receives image data transmitted by the remote sensing image acquisition module, preprocesses the image in a progressive order of geometric correction, spectral correction, radiometric correction, and data integration, generates a geographic information system model containing topographic, land use, and vegetation cover information, stores the preprocessed data in the GIS system, and transmits the geographic information system model to the environmental risk monitoring module. The sensor data acquisition module is used to perform the sensor data acquisition operation in step S4 of claim 1, to acquire wind speed, wind direction, rotation speed, vibration and temperature parameters of the wind turbine, and to transmit the acquired sensor data to the multi-source data fusion module. The environmental risk monitoring module is connected to the GIS preprocessing module and is used to perform the S3 step as described in claim 1. It receives the geographic information system model transmitted by the GIS preprocessing module, performs parallel analysis of vegetation change monitoring, hydrological change monitoring and soil erosion monitoring on the model, identifies environmental changes and potential risks of the wind farm, and transmits the environmental monitoring results to the multi-source data fusion module and the coupled risk assessment module. The multi-source data fusion module is communicatively connected to the sensor data acquisition module, the environmental risk monitoring module, and the remote sensing image acquisition module, respectively. It is used to execute the S4 step described in claim 1, receive sensor data from the sensor data acquisition module, remote sensing data from the remote sensing image acquisition module, and environmental monitoring results from the environmental risk monitoring module, and perform fusion analysis using a spatiotemporal dual-dimensional adaptive weighted data fusion mechanism based on time series analysis and spatial analysis to evaluate the health status and performance of the wind turbine unit, and transmit the unit evaluation results to the coupled risk assessment module. The coupling risk assessment module is communicatively connected to the environmental risk monitoring module and the multi-source data fusion module, respectively. It is used to execute the S5 step described in claim 1, receive the environmental monitoring results from the environmental risk monitoring module and the unit assessment results from the multi-source data fusion module, perform the calculation of the wind turbine environmental exposure index, the calculation of the environmental-equipment time correlation coefficient and the calculation of the dynamic coupling risk index, complete the quantitative assessment of environmental-equipment coupling risk, and transmit the coupling risk assessment results to the GIS display and decision-making module. The GIS display and decision-making module is communicatively connected to the coupled risk assessment module, the environmental risk monitoring module, and the multi-source data fusion module. It is used to execute the display and decision support function in step S5 of claim 1, receive the coupled risk assessment results, environmental monitoring results, and unit assessment results, and display the wind farm's operating status, environmental monitoring results, unit assessment results, and coupled risk assessment results in real time through the GIS platform.

Images