Desert photovoltaic sand damage assessment method, device and equipment based on big data and medium

Abstract

The invention provides a desert photovoltaic sand damage assessment method, device and equipment based on big data and a medium, and the method comprises the steps: obtaining multi-source heterogeneous data of a desert area, carrying out the time-space fusion, and obtaining sand and dust activity features, the multi-source heterogeneous data comprising at least two items of wind speed, wind direction, sand grain granularity and sand and dust concentration; constructing a dynamic sand damage intensity evaluation model based on the sand and dust activity characteristics, calculating a dynamic quantized value of the sand and dust activity intensity, and generating a sand and dust intensity distribution diagram; according to the space-time heterogeneity of the sand and dust intensity distribution diagram, dynamic risk assessment is carried out in combination with desert landform difference characteristics, and risk assessment information is obtained. By adopting the method, the desert photovoltaic sand damage can be accurately evaluated by using multi-source heterogeneous big data, and information support is provided for targeted sand prevention and control measures.

Description

Technical Field

[0001] The present invention belongs to the technical field of ecological assessment, and particularly relates to a method, device, equipment and medium for desert photovoltaic sand hazard assessment based on big data. Background Art

[0002] Desert photovoltaic sand hazard assessment is a crucial research direction in the new energy field, which is directly related to the long-term stable operation and economic benefits of photovoltaic power stations. With the global energy structure transformation towards renewable energy, desert areas have become key regions for photovoltaic industry layout due to their rich solar energy resources. However, the erosion of photovoltaic equipment by sand and dust activities and the performance attenuation pose a non-negligible threat, and scientific assessment and countermeasures are urgently needed. The introduction of big data technology provides new possibilities for sand hazard analysis. Through the integration and analysis of multi-dimensional data, it can more accurately reveal the laws of sand and dust activities and their impacts on photovoltaic systems.

[0003] Currently, desert photovoltaic sand hazard assessment mostly relies on traditional meteorological observations and single-index analysis, such as simple statistics based only on wind speed or dust concentration. Although this method is easy to operate, it is often biased in complex desert environments and difficult to comprehensively reflect the dynamic characteristics and regional differences of sand and dust activities, resulting in the disconnection between assessment results and actual damage levels. In addition, existing solutions lack systematicness in data fusion and model construction and are difficult to meet the diverse needs of different desert landforms and climate conditions.

[0004] The core challenge in the research field lies in how to quantify the intensity of sand and dust activities and their potential impacts on photovoltaic equipment. Wind speed, wind direction, sand grain size and content, as key technical factors, jointly determine the occurrence mechanism and damage degree of sand hazards. However, the spatio-temporal distribution of these factors is highly heterogeneous and their interactions are complex. Traditional methods are difficult to effectively integrate multi-source data and establish a reliable intensity evaluation model. This not only limits the accurate prediction of sand hazard risks but also makes the targeted protection design insufficient. Especially in areas with high-intensity sand and dust activities, there are technical bottlenecks in the optimization of equipment material selection and dust prevention measures.

[0005] However, the current desert photovoltaic sand hazard assessment methods have insufficient use of big data, resulting in low assessment accuracy, difficulty in reflecting the real sand hazard situation, difficulty in correct site selection or targeted protection, and affecting the durability and operation efficiency of photovoltaic equipment. Summary of the Invention

[0006] Based on this, it is necessary to provide a method, device, equipment and medium for desert photovoltaic sand hazard assessment based on big data to accurately assess desert photovoltaic sand hazards using multi-source heterogeneous big data and provide information support for targeted sand prevention and control measures.

[0007] In a first aspect, the present application provides a method for evaluating desert photovoltaic sand hazards based on big data, the method comprising:

[0008] Obtaining multi-source heterogeneous data in a desert area and performing spatio-temporal fusion to obtain dust activity characteristics, the multi-source heterogeneous data including at least two of wind speed, wind direction, sand grain size, and dust concentration;

[0009] Constructing a dynamic sand hazard intensity evaluation model based on the dust activity characteristics, calculating a dynamic quantization value of the dust activity intensity, and generating a dust intensity distribution map;

[0010] Performing dynamic risk assessment according to the spatio-temporal heterogeneity of the dust intensity distribution map, and combining the desert geomorphic difference characteristics to obtain risk assessment information.

[0011] In a possible embodiment, obtaining multi-source heterogeneous data in a desert area and performing spatio-temporal fusion to obtain dust activity characteristics, including:

[0012] Performing spatio-temporal alignment processing on the wind speed and wind direction in the multi-source heterogeneous data to obtain a temporal wind speed field and a spatial wind direction vector, wherein the temporal wind speed field is generated by interpolating to unify the timestamps, and the spatial wind direction vector is generated by vector mapping to a geographic grid;

[0013] Generating a dynamic weight matrix based on the spatial distribution data of sand grain size and dust concentration, the dynamic weight matrix being used to characterize the coupling relationship between sand grain migration probability and wind speed;

[0014] Performing tensor fusion on the temporal wind speed field, the spatial wind direction vector, and the dynamic weight matrix to generate dust activity characteristics.

[0015] In a possible embodiment, constructing a dynamic sand hazard intensity evaluation model based on the dust activity characteristics, calculating a dynamic quantization value of the dust activity intensity, and generating a dust intensity distribution map, including:

[0016] Extracting the wind speed-wind direction spatio-temporal correlation in the dust activity characteristics through a spatio-temporal convolutional neural network to obtain a preliminary intensity quantization value;

[0017] Using the following formula, when the preliminary intensity quantization value exceeds a preset threshold, calculating a correction factor based on the probability density function of the grain size distribution of sand grains:

[0018]

[0019] where α is the correction factor, p is the sand grain size, f(p) is the grain size distribution probability density, p k is the non-linear weighting exponent, and Δp is the grain size distribution span;

[0020] Non-linearly weight the preliminary intensity quantization value according to the correction factor to generate a dust intensity distribution map. In a possible embodiment, according to the spatio-temporal heterogeneity of the dust intensity distribution map, combined with the characteristics of desert landform differences, perform dynamic risk assessment to obtain risk assessment information, including:

[0021] Perform density clustering analysis on the dust intensity distribution map to generate an initial risk level partition;

[0022] Based on the desert landform slope angle and vegetation cover index, calculate the deflection effect of the dust migration path between regions. Based on the deflection effect, obtain an adaptive risk distribution, and the deflection effect is generated through the projection operation of the terrain slope angle on the wind direction vector;

[0023] Generate risk assessment information according to the mapping relationship between the adaptive risk distribution and the spatial coordinates of the photovoltaic equipment.

[0024] In a possible embodiment, the method further includes:

[0025] Generate a protection parameter table according to the risk assessment information, and the protection parameter table includes the dust-proof coating thickness and the optimized value of the bracket inclination angle;

[0026] Input the protection parameter table into the digital twin model to simulate the dynamic wear process of dust and generate a wear rate field;

[0027] Calculate the durability attenuation curve of the photovoltaic equipment based on the wear rate field and output the durability distribution map.

[0028] In a possible embodiment, based on the spatial distribution data of sand grain size and dust concentration, generate a dynamic weight matrix, including:

[0029] Obtain the three-dimensional data of the desert terrain through lidar scanning to generate a terrain shielding coefficient matrix, and the terrain shielding coefficient matrix is generated by calculating the cosine value of the included angle between the terrain elevation and the wind direction;

[0030] Calculate the wind speed attenuation factor according to the terrain shielding coefficient matrix and the divergence distribution of the spatial wind direction vector, and the wind speed attenuation factor is used to quantify the weakening effect of the terrain on the wind speed;

[0031] Generate a dynamic weight matrix based on the wind speed attenuation factor and the spatial gradient distribution of dust concentration.

[0032] In a possible embodiment, the method further includes:

[0033] Perform Monte Carlo sampling on the dust intensity distribution map to generate a confidence heat map;

[0034] Through the following formula, calculate the safety redundancy factor according to the confidence heat map and the risk assessment information:

[0035]

[0036] Among them, γ is the safety redundancy coefficient, and σ s is the standard deviation of the dust storm intensity, and μ s is the mean value of the dust storm intensity, R is the risk level, and Φ(R) is the non-linear saturation function of the risk level;

[0037] The optimized coating thickness is obtained by multiplying the safety redundancy coefficient by the reference coating thickness, and the protection parameter table is updated according to the optimized coating thickness, where the reference coating thickness is obtained according to the material properties of the photovoltaic device.

[0038] In a second aspect, the present application also provides a desert photovoltaic sand hazard assessment device based on big data. The device includes:

[0039] A big data processing module, configured to obtain multi-source heterogeneous data in the desert area and perform spatio-temporal fusion to obtain dust storm activity characteristics. The multi-source heterogeneous data includes at least two of wind speed, wind direction, sand grain size, and dust concentration;

[0040] A sand hazard intensity evaluation module, configured to construct a dynamic sand hazard intensity evaluation model based on the dust storm activity characteristics, calculate the dynamic quantization value of the dust storm activity intensity, and generate a dust storm intensity distribution map;

[0041] A risk assessment module, configured to perform dynamic risk assessment according to the spatio-temporal heterogeneity of the dust storm intensity distribution map and combine the desert landform difference characteristics to obtain risk assessment information.

[0042] In a third aspect, the present application also provides a computer device, including a memory and a processor. The memory stores a computer program, and when the processor executes the computer program, the above-mentioned desert photovoltaic sand hazard assessment method based on big data is implemented.

[0043] In a fourth aspect, the present application also provides a computer-readable storage medium, on which a computer program is stored. When the computer program is executed by a processor, the above-mentioned desert photovoltaic sand hazard assessment method based on big data is implemented.

[0044] The above-mentioned big data-based desert photovoltaic sand hazard assessment method, device, equipment and medium obtain multi-source heterogeneous data in desert areas, including at least two of wind speed, wind direction, sand grain size and dust concentration, and perform spatio-temporal fusion on these data to obtain the characteristics of sand dust activities; based on these sand dust activity characteristics, a dynamic sand hazard intensity evaluation model is constructed, the dynamic quantification value of the sand dust activity intensity is calculated, and a sand dust intensity distribution map is generated; combined with the desert geomorphic difference characteristics, dynamic risk assessment is carried out according to the spatio-temporal heterogeneity of the sand dust intensity distribution map to obtain risk assessment information. The above technical solution realizes the accurate quantitative assessment of desert photovoltaic sand hazards through the synergistic effect of multi-source data fusion, dynamic modeling and geomorphic adaptive assessment, and provides reliable information support for targeted sand prevention and control measures. BRIEF DESCRIPTION OF THE DRAWINGS

[0045] In order to more clearly illustrate the technical solutions in the embodiments of the present application or related technologies, the following will briefly introduce the drawings required for use in the description of the embodiments or related technologies. Obviously, the drawings in the following description are only some embodiments of the present application. For those of ordinary skill in the art, without creative efforts, other drawings can be obtained based on these drawings.

[0046] Figure 1 It is a flowchart of the big data-based desert photovoltaic sand hazard assessment method provided by an embodiment of the present invention;

[0047] Figure 2 It is a schematic structural diagram of the big data-based desert photovoltaic sand hazard assessment device provided by an embodiment of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0048] In order to make the objectives, technical solutions and advantages of the present application clearer, the following further details the present application in conjunction with the drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present application and are not used to limit the present application.

[0049] First, a brief introduction is made to the nouns involved in the embodiments of the present application.

[0050] Desert photovoltaic sand hazard refers to the damage and impact on photovoltaic facilities caused by sand dust activities when building a photovoltaic power station in a desert area. This kind of sand hazard is mainly manifested as problems such as the reduction of power generation efficiency caused by sand dust covering the photovoltaic panels, the surface wear caused by sand grains impacting the photovoltaic components, and the influence of sand dust accumulation on equipment heat dissipation and normal operation. The severity of the sand hazard is closely related to the intensity and frequency of sand dust activities and the desert geomorphic characteristics, and is one of the key factors restricting the stable operation and efficient power generation of desert photovoltaic power stations.

[0051] Multi-source heterogeneous data refers to a collection of data from different sources, with different formats, structures, and characteristics. In practical applications, such data may include structured data (such as tabular data in a database), semi-structured data (such as XML and JSON files), and unstructured data (such as text, images, audio, and video). The integration and analysis of multi-source heterogeneous data can make full use of the advantages of different data sources, extract more comprehensive and in-depth information, and thus provide stronger support for solving complex problems. For example, in the assessment of desert photovoltaic sand hazards, by combining multi-source heterogeneous data such as wind speed and direction data from meteorological stations, dust concentration data from remote sensing, and sand grain size data from ground sensors, the characteristics of dust activities can be more accurately characterized, providing a basis for the precise assessment and prevention of sand hazards.

[0052] The Spatio-Temporal Convolutional Neural Network (STCNN) is a deep learning model that combines spatial and temporal information and is specifically designed to process data with spatio-temporal dependencies. By simultaneously applying spatial convolution and temporal convolution operations in the convolutional layer, it can effectively capture the local features of data in the spatial dimension and the dynamic changes in the temporal dimension. This network structure is widely used in fields such as video analysis, weather prediction, and traffic flow prediction, and can extract valuable patterns and rules from continuous spatio-temporal data, thus providing a powerful tool for the modeling and prediction of complex systems.

[0053] Based on the above noun explanations, the implementation environment of the big data-based desert photovoltaic sand hazard assessment method provided by the embodiments of this application is described. Schematically, this implementation environment includes: a sensor group, a terminal, and a processor. Among them, the terminal is signal-connected to the sensor group and the processor through a network; the sensor group includes, but is not limited to, synthetic aperture radar, optical cameras, infrared scanners, meteorological sensors, etc.; the processor can be a central processing unit, a neural network processor, or a multi-core processor, which is not limited here.

[0054] Combined with the above noun explanations and implementation environment, the application scenarios of the embodiments of this application are described. The big data-based desert photovoltaic sand hazard assessment method provided in the embodiments of this application can be applied to the following scenarios, including but not limited to:

[0055] In the site selection planning and layout optimization of desert photovoltaic power stations, when building large-scale photovoltaic power stations in desert areas, traditional site selection mainly relies on sunlight resource assessment, but the impact of dust activities on the equipment life is often underestimated. By integrating wind speed, wind direction, and sand grain size data, this technology can generate a dust intensity distribution map and accurately identify low-risk areas of sand hazards.

[0056] In the intelligent operation and maintenance and clean scheduling of a photovoltaic power station, this solution can be used for decision-making support in daily operation and maintenance. By real-time monitoring of data such as wind speed, wind direction, and dust concentration, combined with a dynamic sand hazard intensity evaluation model, changes in sand hazard risks can be detected in a timely manner. For example, when it is detected that the intensity of sand and dust activities in a certain area suddenly increases, maintenance work such as cleaning photovoltaic panels and strengthening windproof facilities can be arranged in advance to avoid a decline in the performance of photovoltaic modules caused by sand accumulation or sand particle erosion, ensure the stable operation and power generation efficiency of the photovoltaic power station, and reduce operation and maintenance costs.

[0057] When carrying out sand prevention and control projects around a desert photovoltaic power station, this solution can provide accurate data support for project design. According to the sand dust intensity distribution map and risk assessment information, combined with the desert geomorphic characteristics, key treatment areas and treatment measures can be determined. For example, in areas with frequent sand and dust activities and a greater impact on the photovoltaic power station, windbreak and sand fixation nets can be preferentially arranged, drought-tolerant sand-fixing plants can be planted, etc.; while in areas with relatively weak sand and dust activities, relatively simple protection measures can be adopted, so as to achieve the accurate design of sand prevention and control projects and the reasonable allocation of resources, and improve the effect and efficiency of sand prevention and control.

[0058] Schematically, the big data-based desert photovoltaic sand hazard assessment method provided by the embodiments of the present application can also be applied to other application scenarios. Only examples are given here, and the specific application scenarios are not limited.

[0059] In one embodiment, as Figure 1 shown, a big data-based desert photovoltaic sand hazard assessment method is provided. In this embodiment, taking the application of this method to a terminal as an example, it can be understood that this method can also be applied to a server, and can also be applied to a system including a terminal and a server, and is implemented through the interaction between the terminal and the server.

[0060] A big data-based desert photovoltaic sand hazard assessment method provided by the embodiments of the present application includes the following steps 101 to 103:

[0061] Step 101, obtain multi-source heterogeneous data in the desert area and perform spatio-temporal fusion to obtain sand and dust activity characteristics. The multi-source heterogeneous data includes at least two of wind speed, wind direction, sand grain size, and dust concentration.

[0062] Specifically, through a sensor network deployed in the desert area, multi-source heterogeneous data including wind speed, wind direction, sand grain size, and dust concentration is collected. The multi-source heterogeneous data has a wide range of sources, covering various means such as ground meteorological stations, satellite remote sensing equipment, and unmanned aerial vehicle monitoring. Through spatio-temporal fusion technology, data with different time scales and spatial resolutions are integrated to form unified sand and dust activity characteristics.

[0063] Step 102: Based on the characteristics of sand and dust activities, construct a dynamic sand hazard intensity evaluation model, calculate the dynamic quantification value of the sand and dust activity intensity, and generate a sand and dust intensity distribution map.

[0064] Exemplarily, use the fused sand and dust activity characteristics to construct a dynamic sand hazard intensity evaluation model. This model comprehensively considers the impacts of factors such as wind speed, wind direction, sand grain size, and sand and dust concentration on sand hazards, and obtains the dynamic quantification value of the sand and dust activity intensity through weighted calculation. By constructing the dynamic sand hazard intensity evaluation model, it is possible to calculate the dynamic quantification value of the sand and dust activity intensity in real time and generate a sand and dust intensity distribution map.

[0065] Step 103: According to the spatio-temporal heterogeneity of the sand and dust intensity distribution map, combine the desert landform difference characteristics to conduct dynamic risk assessment and obtain risk assessment information.

[0066] Specifically, the desert landform difference characteristics, including terrain slope, surface roughness, and vegetation cover index, can be extracted based on the digital elevation model (DEM) and surface cover type data. By analyzing the impacts of different landform types on sand and dust activities and the variation laws of sand and dust intensity in time and space, the sand hazard risk information of each region can be evaluated. The obtained risk assessment information can provide a scientific basis for the site selection, construction, and operation and maintenance of desert photovoltaic power stations, and help formulate targeted sand prevention and control measures.

[0067] The above-mentioned big data-based desert photovoltaic sand hazard assessment method, device, equipment, and medium obtain multi-source heterogeneous data in the desert area, including at least two of wind speed, wind direction, sand grain size, and sand and dust concentration, and perform spatio-temporal fusion on these data to obtain the characteristics of sand and dust activities; based on these sand and dust activity characteristics, construct a dynamic sand hazard intensity evaluation model, calculate the dynamic quantification value of the sand and dust activity intensity, and generate a sand and dust intensity distribution map; combine the desert landform difference characteristics, and conduct dynamic risk assessment according to the spatio-temporal heterogeneity of the sand and dust intensity distribution map to obtain risk assessment information. The above technical solution realizes the accurate quantitative assessment of desert photovoltaic sand hazards through the synergistic effect of multi-source data fusion, dynamic modeling, and landform adaptive assessment, and provides reliable information support for targeted sand prevention and control measures.

[0068] In a possible embodiment, obtaining multi-source heterogeneous data in the desert area and performing spatio-temporal fusion to obtain the characteristics of sand and dust activities includes:

[0069] Step 201: Perform spatio-temporal alignment processing on the wind speed and wind direction in the multi-source heterogeneous data to obtain a time-series wind speed field and a spatial wind direction vector. The time-series wind speed field is generated by unifying the timestamps through interpolation method, and the spatial wind direction vector is generated by vector mapping to a geographical grid.

[0070] Exemplarily, the time-series data of wind speed and the wind direction vector data in the desert area can be collected by deploying ground meteorological stations and satellite remote sensing devices; the linear interpolation method is used to align the wind speed data in the time dimension, and the wind speed observation values with discrete timestamps are interpolated into a unified time series to generate a time-series wind speed field with continuous time distribution; based on the Geographic Information System (GIS), the original wind direction vector data is mapped to a geographic grid of 0.1°×0.1°, and a spatial wind direction vector field is generated through vector normalization processing to solve the problems of resolution and coordinate system differences of different sensor data.

[0071] Step 202: Based on the spatial distribution data of sand grain size and dust concentration, generate a dynamic weight matrix, which is used to characterize the coupling relationship between the sand grain migration probability and the wind speed.

[0072] Specifically, a laser particle size analyzer and a particulate matter sensor can be used to obtain the sand grain size distribution data and the dust concentration gradient data in the desert area; the spatial distribution data of sand grain size and dust concentration is used to generate a dynamic weight matrix. This dynamic weight matrix is used to characterize the coupling relationship between the sand grain migration probability and the wind speed, that is, by analyzing the distribution of sand grain size and dust concentration in different regions and combining the influence of wind speed, the weight value is dynamically adjusted to reflect the sand grain migration probability under different wind speed and concentration conditions, providing key parameters for the accurate characterization of dust activity characteristics.

[0073] Step 203: Perform tensor fusion on the time-series wind speed field, the spatial wind direction vector, and the dynamic weight matrix to generate dust activity characteristics.

[0074] Specifically, the time-series wind speed field, the spatial wind direction vector field, and the dynamic weight matrix are constructed into a three-dimensional tensor data structure, where the time dimension, the spatial dimension, and the feature dimension correspond to the time-series change of wind speed, the spatial distribution of wind direction, and the dust migration weight respectively; the core correlation features of multi-dimensional data are extracted through a tensor decomposition algorithm to eliminate redundant noise; the decomposed feature matrices are normalized and spliced to generate a multi-dimensional dust activity feature matrix representing the spatio-temporal evolution law of dust activities, so as to fully consider the interaction of multiple factors and make the dust activity characteristics more comprehensive and real.

[0075] In a possible embodiment, based on the dust activity characteristics, a dynamic sand hazard intensity evaluation model is constructed, the dynamic quantization value of the dust activity intensity is calculated, and a dust intensity distribution map is generated, which may include:

[0076] Step 301: Extract the wind speed-wind direction spatio-temporal correlation in the dust activity characteristics through a spatio-temporal convolutional neural network to obtain a preliminary intensity quantization value.

[0077] Specifically, the sand and dust activity feature matrix can be input into a spatio-temporal convolutional neural network (ST-CNN). The convolutional kernel in the time dimension of the neural network is designed based on the autocorrelation of the wind speed time series, and the convolutional kernel in the spatial dimension is dynamically adjusted according to the spatial divergence distribution of the wind direction vector. Short-term fluctuation features and long-term trend features of sand and dust activities are extracted through multi-level convolution and pooling operations, and a preliminary intensity quantization value matrix is output. Each element in the matrix corresponds to the initial score of the sand hazard intensity of the geographical grid unit.

[0078] Step 302, use the following formula to calculate the correction factor based on the probability density function of the grain size distribution of sand grains when the preliminary intensity quantization value exceeds the preset threshold:

[0079]

[0080] where α is the correction factor, p is the sand grain size, f(p) is the probability density of the grain size distribution, p k is the non-linear weighting exponent, and Δp is the grain size distribution span.

[0081] Specifically, when the preliminary intensity quantization value exceeds the preset threshold, the correction factor is calculated using the probability density function of the sand grain size distribution. This correction factor is used to adjust the preliminary intensity quantization value to more accurately reflect the actual intensity of sand and dust activities.

[0082] Step 303, perform non-linear weighting on the preliminary intensity quantization value according to the correction factor to generate a sand and dust intensity distribution map.

[0083] Specifically, the correction factor and the preliminary intensity quantization value matrix can be multiplied point by point according to the grid unit to non-linearly amplify the high-intensity area. The weighted intensity values are mapped to the continuous geographical space through the Kriging spatial interpolation algorithm to quantify the impact and damage effect of large-particle sand and dust on the photovoltaic panels, and a sand and dust intensity distribution map is generated. Further, this distribution map characterizes the sand hazard intensity levels of different regions with color gradients and marks the geographical coordinate boundaries of high-risk areas.

[0084] In a possible embodiment, according to the spatio-temporal heterogeneity of the sand and dust intensity distribution map, combined with the desert landform difference characteristics, dynamic risk assessment is carried out to obtain risk assessment information, which may include:

[0085] Step 401, perform density clustering analysis on the sand and dust intensity distribution map to generate an initial risk level partition.

[0086] Exemplarily, perform density-based spatial clustering analysis on the dust intensity distribution map, and divide grid cells with similar dust intensity values and continuous space into the same risk level area; by setting density threshold and neighborhood radius parameters, identify high-risk core areas, medium-risk transition areas, and low-risk edge areas, and generate an initial risk level zoning map containing geographical boundary coordinates. The zoning map marks the spatial distribution of each level area with different color blocks.

[0087] Step 402, calculate the deflection effect of the dust migration path between regions based on the desert landform slope angle and vegetation coverage index, and obtain the adaptive risk distribution based on the deflection effect. The deflection effect is generated by the projection operation of the terrain slope angle on the wind direction vector.

[0088] Specifically, the desert landform slope angle data can be extracted based on the digital elevation model (DEM), combined with the vegetation coverage index retrieved from multi-spectral remote sensing images, to construct a terrain-vegetation composite influence factor; calculate the deflection correction amount of the terrain slope angle on the dominant wind direction through vector projection operation, adjust the boundary of the initial risk level zoning according to the deflection correction amount, and adjust the boundary of the initial risk level zoning according to the deflection angle to generate an adaptive risk distribution map integrating the influence of landform.

[0089] Step 403, generate risk assessment information according to the mapping relationship between the adaptive risk distribution and the spatial coordinates of the photovoltaic equipment.

[0090] Specifically, perform geocoding matching on the adaptive risk distribution map and the spatial coordinate data of the photovoltaic equipment, and establish a mapping relationship between the risk level and the equipment location through GIS overlay analysis; further, a risk assessment information table including equipment coordinates, the risk level it belongs to, the expected life attenuation rate, and the recommended protection level can be output. The information table supports CSV format and GIS layer export, and is used to guide the differential protection deployment of the photovoltaic power station.

[0091] In a possible embodiment, the method may further include:

[0092] Step 501, generate a protection parameter table according to the risk assessment information. The protection parameter table includes the thickness of the dust-proof coating and the optimized value of the bracket inclination angle.

[0093] Specifically, based on the obtained risk assessment information, generate a protection parameter table including the thickness of the dust-proof coating and the optimized value of the bracket inclination angle. The protection parameter table provides targeted protection parameter suggestions for the photovoltaic equipment according to the risk level of each region. For example, in high-risk areas, it is recommended to increase the thickness of the dust-proof coating and optimize the bracket inclination angle to reduce the impact of dust on the photovoltaic equipment.

[0094] Step 502, input the protection parameter table into the digital twin model, simulate the dynamic wear process of dust, and generate a wear rate field.

[0095] Exemplarily, input the protection parameter table into the digital twin model, which can be constructed based on the three-dimensional geometric structure of the photovoltaic device, material properties, and the sand dust impact dynamics equation; simulate the impact trajectories and energy distributions of sand grains with different particle sizes on the surface of the photovoltaic panel through the discrete element method (DEM), calculate the wear depth per unit time in combination with the Archard wear model, generate a wear rate field with a spatial resolution of 1m×1m, and label the annual wear rate for each grid unit of the rate field.

[0096] Step 503, calculate the durability decay curve of the photovoltaic device based on the wear rate field, and output the durability distribution map.

[0097] Specifically, based on the wear rate field, calculate the change relationship of the cumulative wear depth of each area on the surface of the photovoltaic device over time through integral operation to generate a durability decay curve; predict the time nodes when each area reaches the failure state according to the preset critical wear threshold; map the prediction results to the geographic coordinate system to generate a durability distribution map. Further, this distribution map labels the remaining years of the device life with a color gradient and marks the spatial coordinates of the high-risk failure areas, intuitively showing the durability levels of the photovoltaic device in different areas.

[0098] In a possible embodiment, generate a dynamic weight matrix based on the spatial distribution data of sand grain size and sand dust concentration, including:

[0099] Step 601, obtain the three-dimensional data of the desert terrain through lidar scanning to generate a terrain shielding coefficient matrix, which is calculated by the cosine value of the angle between the terrain elevation and the wind direction.

[0100] Specifically, the three-dimensional elevation data of the desert terrain can be obtained through lidar scanning of an unmanned aerial vehicle, and a terrain elevation matrix is generated based on geographic grid division; calculate the terrain shielding coefficient of each grid unit according to the cosine value of the angle between the terrain elevation and the dominant wind direction, where the value range of the terrain shielding coefficient is from 0 to 1, and the larger the value, the weaker the blocking effect of the terrain on the sand wind flow; arrange the terrain shielding coefficients according to the geographic grid to generate a terrain shielding coefficient matrix.

[0101] Step 602, calculate the wind speed attenuation factor according to the terrain shielding coefficient matrix and the divergence distribution of the spatial wind direction vector, and the wind speed attenuation factor is used to quantify the weakening effect of the terrain on the wind speed.

[0102] Specifically, calculate the wind speed divergence distribution of each grid unit based on the spatial wind direction vector data, and the divergence distribution is used to characterize the convergence or divergence characteristics of the wind speed in space; combine the terrain shielding coefficient matrix with the wind speed divergence distribution, and generate a wind speed attenuation factor through weighted superposition operation. The attenuation factor quantifies the weakening effect of the terrain undulation on the local wind speed, and the higher the value of the wind speed attenuation factor in the area with a lower terrain shielding coefficient or a larger wind speed divergence.

[0103] Step 603: Generate a dynamic weight matrix based on the spatial gradient distributions of the wind speed attenuation factor and the dust concentration.

[0104] Specifically, a dynamic weight matrix is generated based on the spatial gradient distributions of the wind speed attenuation factor and the dust concentration. This matrix comprehensively considers the effects of wind speed attenuation and dust concentration changes on sand grain migration, can dynamically reflect the migration probability and intensity changes of dust activities under different terrain conditions, and provides an important basis for accurately characterizing the characteristics of dust activities.

[0105] In a possible embodiment, the method may further include:

[0106] Step 701: Perform Monte Carlo sampling on the dust intensity distribution map to generate a confidence heat map.

[0107] Specifically, perform Monte Carlo random sampling on the dust intensity distribution map to simulate the uncertainty of sand hazard intensity by generating multiple groups of dust intensity distribution scenarios; each group of scenarios is constructed based on historical dust activity data and meteorological disturbance parameters, and the standard deviation and mean of the dust intensity of each geographical grid unit are statistically calculated to generate a confidence heat map, which characterizes the credibility level of sand hazard intensity prediction in different regions.

[0108] Step 702: Calculate the safety redundancy coefficient according to the following formula based on the confidence heat map and the risk assessment information:

[0109]

[0110] where γ is the safety redundancy coefficient, σ s is the standard deviation of the dust intensity, μ s is the mean of the dust intensity, R is the risk level, and Φ(R) is a non-linear saturation function of the risk level.

[0111] Specifically, according to the standard deviation and mean of the dust intensity in the confidence heat map, combined with the risk level in the risk assessment information, calculate the safety redundancy coefficient through a non-linear saturation function, comprehensively considering the volatility of the dust intensity (the ratio of the standard deviation to the mean) and the influence of the risk level. The introduction of the non-linear saturation function makes the contribution of the risk level to the safety redundancy coefficient more reasonable, avoids the over-amplification of the safety redundancy in high-risk areas, and improves the practicality and adaptability of the safety redundancy coefficient.

[0112] Step 703: Multiply the safety redundancy coefficient by the reference coating thickness to obtain the optimized coating thickness, and update the protection parameter table according to the optimized coating thickness, where the reference coating thickness is obtained based on the material properties of the photovoltaic device.

[0113] Specifically, set the reference dust-proof coating thickness according to the material properties of the photovoltaic device, multiply the reference thickness by the safety redundancy factor to obtain the dynamically optimized coating thickness value; update the optimized coating thickness value to the protection parameter table, associate the device coordinates with the risk level information, and generate an optimized protection parameter table containing differential protection parameters for guiding the zonal protection construction of the photovoltaic power station.

[0114] In summary, the big data-based desert photovoltaic sand hazard assessment method provided by the embodiments of the present application integrates dynamic monitoring data such as wind speed, wind direction, sand grain size, and dust concentration, and uses spatio-temporal fusion technology to construct a sand dust activity feature matrix reflecting the sand dust migration law; based on a spatio-temporal convolutional neural network, extracts the spatio-temporal correlation of wind speed-wind direction and introduces a sand grain size correction factor to generate a high-resolution sand dust intensity distribution map; combines the desert landform slope angle and vegetation coverage index to dynamically correct the risk zoning boundary, and outputs device-level risk assessment information through spatial coordinate mapping; further simulates the sand dust abrasion process through a digital twin model and calculates the durability distribution, and dynamically optimizes the protection parameter table in combination with the confidence heat map. The above technical solutions achieve the accurate quantification of the sand dust activity intensity and the equipment damage risk through the fusion mechanism of multi-source data collaborative analysis, dynamic modeling, and geomorphic adaptive assessment, break through the limitations of traditional static assessment methods, and can provide full-chain data support for the differential sand prevention design, clean scheduling optimization, and dynamic adjustment of protection projects of photovoltaic power stations, effectively improving the pertinence and economy of sand prevention and control measures.

[0115] It should be understood that although the steps in the flowcharts involved in the above-described embodiments are shown in sequence according to the arrows, these steps do not necessarily have to be executed in the order indicated by the arrows. Unless there is a clear indication in this article, the execution of these steps is not strictly limited in order, and these steps can be executed in other orders. Moreover, at least a part of the steps in the flowcharts involved in the above-described embodiments may include multiple steps or multiple stages, and these steps or stages do not necessarily have to be executed at the same time, but can be executed at different times, and the execution order of these steps or stages is not necessarily sequential, but can be executed alternately or alternately with at least a part of other steps or steps or stages in other steps.

[0116] Based on the same inventive concept, the embodiments of the present application also provide a big data-based desert photovoltaic sand hazard assessment device for implementing the above-mentioned big data-based desert photovoltaic sand hazard assessment method. The solution provided by the device for solving the problem is similar to the solution described in the above method. Therefore, the specific limitations in one or more embodiments of the big data-based desert photovoltaic sand hazard assessment device provided below can refer to the limitations on the big data-based desert photovoltaic sand hazard assessment method in the above text, and will not be repeated here.

[0117] In an exemplary embodiment, as Figure 2 shown, a desert photovoltaic sand hazard assessment device 10 based on big data is provided, including:

[0118] A big data processing module 11, configured to obtain multi-source heterogeneous data in the desert area and perform spatio-temporal fusion to obtain dust activity characteristics, and the multi-source heterogeneous data includes at least two of wind speed, wind direction, sand grain size, and dust concentration.

[0119] A sand hazard intensity evaluation module 12, configured to construct a dynamic sand hazard intensity evaluation model based on the dust activity characteristics, calculate a dynamic quantization value of the dust activity intensity, and generate a dust intensity distribution map.

[0120] A risk assessment module 13, configured to perform dynamic risk assessment according to the spatio-temporal heterogeneity of the dust intensity distribution map and in combination with the desert landform difference characteristics to obtain risk assessment information.

[0121] In a possible embodiment, the big data processing module 11 may include:

[0122] A spatio-temporal alignment unit 111, configured to perform spatio-temporal alignment processing on the wind speed and wind direction in the multi-source heterogeneous data to obtain a time-series wind speed field and a spatial wind direction vector, where the time-series wind speed field is generated by interpolating to unify the time stamps, and the spatial wind direction vector is generated by vector mapping to a geographic grid.

[0123] A dynamic weight unit 112, configured to generate a dynamic weight matrix based on the spatial distribution data of the sand grain size and the dust concentration, and the dynamic weight matrix is used to characterize the coupling relationship between the sand grain migration probability and the wind speed.

[0124] A dust activity characteristic generation unit 113, configured to perform tensor fusion on the time-series wind speed field, the spatial wind direction vector, and the dynamic weight matrix to generate dust activity characteristics.

[0125] In a possible embodiment, the sand hazard intensity evaluation module 12 may include:

[0126] A spatio-temporal convolutional neural network unit 121, configured to extract the wind speed-wind direction spatio-temporal correlation in the dust activity characteristics through a spatio-temporal convolutional neural network to obtain a preliminary intensity quantization value.

[0127] A correction factor calculation unit 122, configured to use the following formula to calculate a correction factor based on the probability density function of the grain size distribution when the preliminary intensity quantization value exceeds a preset threshold:

[0128]

[0129] Where α is the correction factor, p is the sand grain size, f(p) is the grain size distribution probability density, p kis a non-linear weighted exponent, and Δp is the particle size distribution span.

[0130] The dust intensity distribution generation unit 123 is used to non-linearly weight the preliminary intensity quantization value according to the correction factor to generate a dust intensity distribution map.

[0131] In a possible embodiment, the risk assessment module 13 may include:

[0132] The clustering analysis unit 131 is used to perform density clustering analysis on the dust intensity distribution map to generate an initial risk level partition.

[0133] The geomorphological analysis unit 132 is used to calculate the deflection effect of the dust migration path between regions based on the desert geomorphological slope angle and the vegetation coverage index, and obtain an adaptive risk distribution based on the deflection effect. The deflection effect is generated by the projection operation of the terrain slope angle on the wind direction vector.

[0134] The data mapping unit 133 is used to generate risk assessment information according to the spatial coordinate mapping relationship between the adaptive risk distribution and the photovoltaic device.

[0135] In a possible embodiment, the device further includes:

[0136] The protection parameter table generation module 14 is used to generate a protection parameter table according to the risk assessment information. The protection parameter table includes the dust-proof coating thickness and the optimized value of the bracket inclination angle.

[0137] The dust dynamic wear simulation module 15 is used to input the protection parameter table into the digital twin model to simulate the dust dynamic wear process and generate a wear rate field.

[0138] The durability calculation module 16 is used to calculate the durability decay curve of the photovoltaic device based on the wear rate field and output a durability distribution map.

[0139] In a possible embodiment, the dynamic weight unit 112 may include:

[0140] The terrain shielding coefficient acquisition sub-unit 1121 is used to obtain the three-dimensional desert terrain data through lidar scanning to generate a terrain shielding coefficient matrix. The terrain shielding coefficient matrix is generated by calculating the cosine value of the angle between the terrain elevation and the wind direction.

[0141] The wind speed attenuation factor calculation sub-unit 1122 is used to calculate the wind speed attenuation factor according to the terrain shielding coefficient matrix and the divergence distribution of the spatial wind direction vector. The wind speed attenuation factor is used to quantify the weakening effect of the terrain on the wind speed.

[0142] The dynamic weight calculation sub-unit 1123 is used to generate a dynamic weight matrix based on the wind speed attenuation factor and the spatial gradient distribution of the dust concentration.

[0143] In a possible embodiment, the device further includes:

[0144] A Monte Carlo sampling module 17, configured to perform Monte Carlo sampling on the dust intensity distribution map to generate a confidence heat map.

[0145] A safety redundancy factor calculation module 18, configured to calculate a safety redundancy factor according to the confidence heat map and risk assessment information through the following formula:

[0146]

[0147] where γ is the safety redundancy factor, σ s is the standard deviation of the dust intensity, μ s is the mean value of the dust intensity, R is the risk level, and Φ(R) is a non-linear saturation function of the risk level.

[0148] A coating thickness optimization module 19, configured to multiply the safety redundancy factor by a reference coating thickness to obtain an optimized coating thickness, and update the protection parameter table according to the optimized coating thickness, where the reference coating thickness is obtained according to the material properties of the photovoltaic device.

[0149] In one embodiment, a computer device is provided, including a memory and a processor. The memory stores a computer program, and when the processor executes the computer program, the steps of a method for evaluating desert photovoltaic sand hazards based on big data as described above are implemented.

[0150] In one embodiment, a computer-readable storage medium is provided, on which a computer program is stored. When the computer program is executed by a processor, the steps in the above method embodiments are implemented.

[0151] For the device embodiments, since they basically correspond to the method embodiments, the relevant parts can be referred to the partial descriptions of the method embodiments. The device embodiments described above are merely illustrative. The components described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place, or may be distributed to multiple network units. Some or all of the modules can be selected according to actual needs to achieve the purpose of the solution of the present disclosure. Those of ordinary skill in the art can understand and implement it without creative efforts.

[0152] The above-described embodiments merely represent several implementation manners of the embodiments of the present application. The description is relatively specific and detailed, but it should not be construed as a limitation to the patent scope of the embodiments of the application. It should be noted that for those of ordinary skill in the art, without departing from the concept of the embodiments of the present application, several modifications and improvements can still be made, and these all fall within the protection scope of the embodiments of the present application.

Claims

1. A method for evaluating desert photovoltaic sand hazards based on big data, characterized in that The method includes: Obtaining multi-source heterogeneous data in the desert area and performing spatio-temporal fusion to obtain sand dust activity characteristics, where the multi-source heterogeneous data includes at least two of wind speed, wind direction, sand grain size, and sand dust concentration; Constructing a dynamic sand hazard intensity evaluation model based on the sand dust activity characteristics, calculating the dynamic quantization value of the sand dust activity intensity, and generating a sand dust intensity distribution map; Performing dynamic risk assessment according to the spatio-temporal heterogeneity of the sand dust intensity distribution map, and combining with the desert landform difference characteristics to obtain risk assessment information.

2. The method according to claim 1, wherein The obtaining multi-source heterogeneous data in the desert area and performing spatio-temporal fusion to obtain sand dust activity characteristics includes: Performing spatio-temporal alignment processing on the wind speed and the wind direction in the multi-source heterogeneous data to obtain a time series wind speed field and a spatial wind direction vector, where the time series wind speed field is generated by interpolating to unify the time stamps, and the spatial wind direction vector is generated by vector mapping to a geographical grid; Generating a dynamic weight matrix based on the spatial distribution data of the sand grain size and the sand dust concentration, where the dynamic weight matrix is used to characterize the coupling relationship between the sand grain migration probability and the wind speed; Performing tensor fusion on the time series wind speed field, the spatial wind direction vector, and the dynamic weight matrix to generate the sand dust activity characteristics.

3. The method according to claim 2, wherein The constructing a dynamic sand hazard intensity evaluation model based on the sand dust activity characteristics, calculating the dynamic quantization value of the sand dust activity intensity, and generating a sand dust intensity distribution map includes: Extracting the wind speed-wind direction spatio-temporal correlation in the sand dust activity characteristics through a spatio-temporal convolutional neural network to obtain a preliminary intensity quantization value; Using the following formula, when the preliminary intensity quantization value exceeds a preset threshold, calculating a correction factor based on the probability density function of the grain size distribution of the sand grains: where α is the correction factor, p is the sand grain size, f(p) is the probability density of grain size distribution, p k is the non-linear weighted exponent, and Δp is the span of grain size distribution; Non-linearly weighting the preliminary intensity quantization value according to the correction factor to generate the sand dust intensity distribution map.

4. The method according to claim 1, wherein The performing dynamic risk assessment according to the spatio-temporal heterogeneity of the sand dust intensity distribution map, and combining with the desert landform difference characteristics to obtain risk assessment information includes: Performing density clustering analysis on the sand dust intensity distribution map to generate an initial risk level partition; Calculating the deflection effect of the sand dust migration path between regions based on the desert landform slope angle and the vegetation coverage index, and obtaining an adaptive risk distribution based on the deflection effect, where the deflection effect is generated by the projection operation of the terrain slope angle on the wind direction vector; Generating the risk assessment information according to the mapping relationship between the adaptive risk distribution and the spatial coordinates of the photovoltaic equipment.

5. The method according to claim 1, characterized in that, The method further includes: Generating a protection parameter table according to the risk assessment information, where the protection parameter table includes the dust-proof coating thickness and the optimized value of the support inclination angle; Inputting the protection parameter table into a digital twin model to simulate the dynamic sand dust wear process and generate a wear rate field; Calculating the durability decay curve of the photovoltaic equipment based on the wear rate field and outputting a durability distribution map.

6. The method according to claim 2, characterized in that, The generating a dynamic weight matrix based on the spatial distribution data of the sand grain size and the sand dust concentration includes: Obtaining three-dimensional desert terrain data through lidar scanning to generate a terrain shielding coefficient matrix, where the terrain shielding coefficient matrix is generated by calculating the cosine value of the included angle between the terrain elevation and the wind direction; Calculate a wind speed attenuation factor based on the terrain shielding coefficient matrix and the divergence distribution of the spatial wind direction vector, where the wind speed attenuation factor is used to quantify the weakening effect of the terrain on the wind speed; Generate the dynamic weight matrix based on the spatial gradient distribution of the wind speed attenuation factor and the dust concentration.

7. The method according to claim 5, wherein The method further includes: Performing Monte Carlo sampling on the dust intensity distribution map to generate a confidence heat map; Calculate a safety redundancy factor according to the confidence heat map and the risk assessment information through the following formula: where γ is the safety redundancy factor, σ s is the standard deviation of dust intensity, μ s is the mean value of dust intensity, R is the risk level, and Φ(R) is the non-linear saturation function of the risk level; Multiply the safety redundancy factor by a reference coating thickness to obtain an optimized coating thickness, and update the protection parameter table according to the optimized coating thickness, where the reference coating thickness is obtained based on the material properties of the photovoltaic device.

8. A desert photovoltaic sand hazard assessment device based on big data, characterized in that, The device includes: A big data processing module, configured to obtain multi-source heterogeneous data in the desert area and perform spatio-temporal fusion to obtain dust activity characteristics, where the multi-source heterogeneous data includes at least two of wind speed, wind direction, sand grain size, and dust concentration; A sand hazard intensity evaluation module, configured to construct a dynamic sand hazard intensity evaluation model based on the dust activity characteristics, calculate a dynamic quantization value of the dust activity intensity, and generate a dust intensity distribution map; A risk assessment module, configured to perform dynamic risk assessment according to the spatio-temporal heterogeneity of the dust intensity distribution map and in combination with the desert landform difference characteristics to obtain risk assessment information.

9. A computer device, comprising a memory and a processor, the memory storing a computer program, characterized in that, When the processor executes the computer program, the method according to any one of claims 1 to 7 is implemented.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by the processor, the method according to any one of claims 1 to 7 is implemented.

Images