Distributed photovoltaic power generation potential assessment method, device and equipment

Abstract

The embodiment of the invention provides a distributed photovoltaic power generation potential assessment method, device and equipment. The method comprises the steps of obtaining multi-source scene data of a distributed photovoltaic scene, and performing cross-modal registration processing on the multi-source scene data according to a preset fusion dimension to obtain a data cube corresponding to the distributed photovoltaic scene; according to the data cube, determining a candidate laying surface attribute set corresponding to the distributed photovoltaic scene; determining an effective irradiation sequence of each candidate laying surface according to the irradiation information of each candidate laying surface; according to the electrical load sequence, the effective irradiation sequence and the candidate laying surface of the distributed photovoltaic scene, determining the acceptability of the transformer area under the distributed photovoltaic scene; and determining a power generation potential index according to the effective irradiation sequence, the acceptability of the transformer area and the candidate laying surface attribute set, thereby evaluating the photovoltaic power generation potential in the distributed photovoltaic scene according to the power generation potential index, and improving the accuracy, reliability and implementability of power generation potential evaluation.

Description

Technical Field

[0001] This application relates to the field of power system technology, and in particular to a method, apparatus and equipment for assessing the potential of distributed photovoltaic power generation. Background Technology

[0002] With the acceleration of the new energy transition, distributed photovoltaic (PV) power generation, as an important component in promoting a low-carbon energy structure, is rapidly becoming widespread in urban building clusters, industrial parks, and rural areas. To fully utilize the power generation performance of distributed PV, its potential is typically assessed before its implementation.

[0003] Traditional potential assessment methods mainly utilize the area and sunlight conditions of building rooftops or vacant land to evaluate the potential of distributed photovoltaic power generation through static irradiance and photovoltaic module arrangement.

[0004] This implementation method results in a relatively singular assessment factor for the potential of distributed photovoltaic (PV) power generation, thus affecting the accuracy of the assessment. Furthermore, this assessment method neglects the impact of the output characteristics of distributed PV power generation on the distribution network after it is connected to the substation, further impacting the accuracy of the distributed PV power generation potential assessment. Summary of the Invention

[0005] This application provides a method, apparatus, and equipment for assessing the potential of distributed photovoltaic power generation, which can improve the accuracy, reliability, and feasibility of power generation potential assessment.

[0006] In a first aspect, embodiments of this application provide a method for assessing the potential of distributed photovoltaic power generation, including:

[0007] Acquire multi-source scene data of distributed photovoltaic scenarios, and perform cross-modal registration processing on the multi-source scene data according to a preset fusion dimension to obtain a data cube corresponding to the distributed photovoltaic scenario;

[0008] Based on the data cube, a set of candidate paving surface attributes corresponding to the distributed photovoltaic scenario is determined; wherein, the set of candidate paving surface attributes includes candidate paving surfaces and paving surface attribute information for each candidate paving surface;

[0009] Based on the irradiation information of each candidate paving surface, determine the effective irradiation sequence for each candidate paving surface;

[0010] Based on the electricity load sequence of the distributed photovoltaic scenario, the effective irradiance sequence, and the candidate installation area, determine the acceptability of the lower transformer area of ​​the distributed photovoltaic scenario;

[0011] A power generation potential index is determined based on the effective irradiance sequence, the acceptability of the distribution area, and the attribute set of the candidate paving surface; wherein, the power generation potential index is used to evaluate the photovoltaic power generation potential under the distributed photovoltaic scenario.

[0012] In one possible implementation, the multi-source scene data includes at least remote sensing image data, UAV orthophoto data, and airborne lidar point cloud data; the preset fusion dimension includes at least a spatial index dimension and a temporal index dimension.

[0013] Based on a preset fusion dimension, cross-modal registration processing is performed on the multi-source scene data to obtain a data cube corresponding to the distributed photovoltaic scene, including:

[0014] The remote sensing image data is subjected to a first preprocessing process to obtain processed remote sensing image data; and the airborne lidar point cloud data is subjected to a second preprocessing process to obtain processed airborne lidar point cloud data.

[0015] Based on the UAV orthophoto data and key point matching method, the processed remote sensing image data is corrected to obtain corrected remote sensing image data. Based on the UAV orthophoto data and three-dimensional projection method, the processed airborne lidar point cloud data is corrected to obtain corrected airborne lidar point cloud data.

[0016] Based on the ground points in the corrected airborne lidar point cloud data, a digital terrain model of the distributed photovoltaic scene is determined, and based on the non-ground points in the corrected airborne lidar point cloud data, a digital surface model of the distributed photovoltaic scene is determined.

[0017] Based on the constructed spatial index dimension and temporal index dimension, the corrected remote sensing image data, the UAV orthophoto data, the corrected airborne lidar point cloud data, the digital terrain model, and the digital surface model are written into tiles to obtain a tiled data set.

[0018] Determine at least one quality mask for the tiled data set, and determine the data cube based on the tiled data set and the at least one quality mask.

[0019] In one possible implementation, determining the candidate paving surface attribute set corresponding to the distributed photovoltaic scenario based on the data cube includes:

[0020] Multi-scale semantic segmentation processing is performed on the corrected remote sensing image data and the UAV orthophoto data to obtain semantic segmentation results; wherein, the semantic segmentation results include the first paving surface;

[0021] Based on the non-ground points in the corrected airborne lidar point cloud data, a plane fitting process is performed on the distributed photovoltaic scene to obtain a plane fitting result; wherein, the plane fitting result includes a second paving surface;

[0022] After cross-verification of the first paving surface and the second paving surface, candidate paving surfaces for the distributed photovoltaic scenario are obtained.

[0023] Based on the roof type information, structural load-bearing level information and paving surface boundary information corresponding to each candidate paving surface, the paving surface attribute information of each candidate paving surface is determined.

[0024] The candidate paving surface attribute set is determined based on the candidate paving surfaces and the paving surface attribute information of each candidate paving surface.

[0025] In one possible implementation, the method further includes:

[0026] Based on the solar trajectory calculation method, the solar driving information of each candidate paving surface is determined;

[0027] Based on the digital surface model of each candidate paving surface, the solar drive information is processed by shading ray tracing to obtain shading constraint information, and the sky view factor and ground time domain factor are determined based on the paving surface tilt angle of each candidate paving surface.

[0028] Acquire meteorological observation data; wherein, the meteorological observation data includes global horizontal irradiance data and surface albedo.

[0029] The irradiance information of each candidate paving surface is determined based on the global horizontal irradiance data, the surface albedo, the solar drive information, the sky view factor, and the ground time domain factor.

[0030] In one possible implementation, the meteorological observation data also includes ambient temperature information;

[0031] Based on the irradiation information of each candidate paving surface, an effective irradiation sequence for each candidate paving surface is determined, including:

[0032] Based on the ambient temperature information and the irradiation information, a temperature correction factor is determined;

[0033] The effective irradiation sequence for each candidate paving surface is determined based on the irradiation information of each candidate paving surface, the temperature correction coefficient, the preset incident angle correction coefficient, the preset surface contamination factor, and the shading constraint information.

[0034] In one possible implementation, the acceptability of the distributed photovoltaic (PV) substation is determined based on the electricity load sequence of the distributed PV scenario, the effective irradiance sequence, and the candidate installation area, including:

[0035] After performing data alignment processing on the power load sequence and the effective irradiation sequence, an aligned power load sequence and an aligned effective irradiation sequence are obtained.

[0036] Determine the transformer substation topology information and the device status information of the deployed equipment in the distributed photovoltaic scenario;

[0037] Based on the mapping relationship table between the candidate laying surface and the injection node of the distribution network in the distributed photovoltaic scenario, and the parameter constraint information of each injection node, the acceptance capacity prediction model is determined.

[0038] The aligned power load sequence, the aligned effective irradiance sequence, the transformer area topology information, and the equipment status information are input into the acceptable capacity prediction model to determine the acceptable upper bound of the transformer area under the distributed photovoltaic scenario.

[0039] After applying a weighted smoothing process to the upper bound of the acceptable capacity, the acceptable capacity of the transformer area is obtained.

[0040] In one possible implementation, before determining the power generation potential index based on the effective irradiation sequence, the acceptability of the substation, and the candidate paving surface attribute set, the method further includes:

[0041] Based on the candidate paving surface attribute set, determine the candidate paving mesh cell set, and determine the attitude candidate set for each candidate paving mesh cell in the candidate paving mesh cell set;

[0042] Based on the acceptable capacity of the transformer area, an hourly capacity quota table is determined, and based on the hourly capacity quota table, capacity pre-allocation and local filling processing are performed to obtain a grid selection sequence for candidate grid cells in the candidate grid cell set.

[0043] Based on the grid selection sequence and the attitude candidate set, the component layout scheme of the photovoltaic modules in the distributed photovoltaic scenario is determined.

[0044] In one possible implementation, a power generation potential index is determined based on the effective irradiation sequence, the acceptability of the substation area, and the set of candidate paving surface attributes, including:

[0045] Based on the effective irradiance sequence, determine the annual cumulative effective irradiance and shading risk factor within the available boundary range under each of the component layout schemes;

[0046] Based on the acceptable capacity of the aforementioned transformer area, determine the average acceptable capacity under each of the aforementioned component layout schemes;

[0047] The component layout scheme is verified based on the candidate paving surface attribute set to obtain a verification result identifier;

[0048] The power generation potential index is determined based on the annual cumulative effective irradiance, the shading risk factor, the average acceptable capacity, and the verification result identifier.

[0049] Secondly, embodiments of this application provide a distributed photovoltaic power generation potential assessment device, comprising:

[0050] The acquisition unit is used to acquire multi-source scene data of the distributed photovoltaic scenario, and perform cross-modal registration processing on the multi-source scene data according to a preset fusion dimension to obtain the data cube corresponding to the distributed photovoltaic scenario.

[0051] The first determining unit is configured to determine the candidate paving surface attribute set corresponding to the distributed photovoltaic scenario based on the data cube; wherein the candidate paving surface attribute set includes candidate paving surfaces and paving surface attribute information for each candidate paving surface.

[0052] The second determining unit is used to determine the effective irradiation sequence of each candidate paving surface based on the irradiation information of each candidate paving surface.

[0053] The third determining unit is used to determine the acceptability of the lower transformer area of ​​the distributed photovoltaic scenario based on the electricity load sequence of the distributed photovoltaic scenario, the effective irradiance sequence and the candidate laying surface;

[0054] An evaluation unit is used to determine a power generation potential index based on the effective irradiance sequence, the acceptability, and the candidate paving surface attribute set; wherein the power generation potential index is used to evaluate the photovoltaic power generation potential under the distributed photovoltaic scenario.

[0055] In one possible implementation, the multi-source scene data includes at least remote sensing image data, UAV orthophoto data, and airborne lidar point cloud data; the preset fusion dimension includes at least a spatial index dimension and a temporal index dimension; in this case, the acquisition unit is used for:

[0056] The remote sensing image data is subjected to a first preprocessing process to obtain processed remote sensing image data; and the airborne lidar point cloud data is subjected to a second preprocessing process to obtain processed airborne lidar point cloud data.

[0057] Based on the UAV orthophoto data and key point matching method, the processed remote sensing image data is corrected to obtain corrected remote sensing image data. Based on the UAV orthophoto data and three-dimensional projection method, the processed airborne lidar point cloud data is corrected to obtain corrected airborne lidar point cloud data.

[0058] Based on the ground points in the corrected airborne lidar point cloud data, a digital terrain model of the distributed photovoltaic scene is determined, and based on the non-ground points in the corrected airborne lidar point cloud data, a digital surface model of the distributed photovoltaic scene is determined.

[0059] Based on the constructed spatial index dimension and temporal index dimension, the corrected remote sensing image data, the UAV orthophoto data, the corrected airborne lidar point cloud data, the digital terrain model, and the digital surface model are written into tiles to obtain a tiled data set.

[0060] Determine at least one quality mask for the tiled data set, and determine the data cube based on the tiled data set and the at least one quality mask.

[0061] In one possible implementation, the first determining unit is configured to:

[0062] Multi-scale semantic segmentation processing is performed on the corrected remote sensing image data and the UAV orthophoto data to obtain semantic segmentation results; wherein, the semantic segmentation results include the first paving surface;

[0063] Based on the non-ground points in the corrected airborne lidar point cloud data, a plane fitting process is performed on the distributed photovoltaic scene to obtain a plane fitting result; wherein, the plane fitting result includes a second paving surface;

[0064] After cross-verification of the first paving surface and the second paving surface, candidate paving surfaces for the distributed photovoltaic scenario are obtained.

[0065] Based on the roof type information, structural load-bearing level information and paving surface boundary information corresponding to each candidate paving surface, the paving surface attribute information of each candidate paving surface is determined.

[0066] The candidate paving surface attribute set is determined based on the candidate paving surfaces and the paving surface attribute information of each candidate paving surface.

[0067] In one possible implementation, the device is also used for:

[0068] Based on the solar trajectory calculation method, the solar driving information of each candidate paving surface is determined;

[0069] Based on the digital surface model of each candidate paving surface, the solar drive information is processed by shading ray tracing to obtain shading constraint information, and the sky view factor and ground time domain factor are determined based on the paving surface tilt angle of each candidate paving surface.

[0070] Acquire meteorological observation data; wherein, the meteorological observation data includes global horizontal irradiance data and surface albedo.

[0071] The irradiance information of each candidate paving surface is determined based on the global horizontal irradiance data, the surface albedo, the solar drive information, the sky view factor, and the ground time domain factor.

[0072] In one possible implementation, the meteorological observation data further includes ambient temperature information; in this case, the second determining unit is used to:

[0073] Based on the ambient temperature information and the irradiation information, a temperature correction factor is determined;

[0074] The effective irradiation sequence for each candidate paving surface is determined based on the irradiation information of each candidate paving surface, the temperature correction coefficient, the preset incident angle correction coefficient, the preset surface contamination factor, and the shading constraint information.

[0075] In one possible implementation, the third determining unit is configured to:

[0076] After performing data alignment processing on the power load sequence and the effective irradiation sequence, an aligned power load sequence and an aligned effective irradiation sequence are obtained.

[0077] Determine the transformer substation topology information and the device status information of the deployed equipment in the distributed photovoltaic scenario;

[0078] Based on the mapping relationship table between the candidate laying surface and the injection node of the distribution network in the distributed photovoltaic scenario, and the parameter constraint information of each injection node, the acceptance capacity prediction model is determined.

[0079] The aligned power load sequence, the aligned effective irradiance sequence, the transformer area topology information, and the equipment status information are input into the acceptable capacity prediction model to determine the acceptable upper bound of the transformer area under the distributed photovoltaic scenario.

[0080] After applying a weighted smoothing process to the upper bound of the acceptable capacity, the acceptable capacity of the transformer area is obtained.

[0081] In one possible implementation, the device is also used for:

[0082] Before determining the power generation potential index based on the effective irradiation sequence, the acceptability of the substation area, and the candidate paving surface attribute set, a candidate paving grid cell set is determined based on the candidate paving surface attribute set, and the attitude candidate set of each candidate paving grid cell in the candidate paving grid cell set is determined.

[0083] Based on the acceptable capacity of the transformer area, an hourly capacity quota table is determined, and based on the hourly capacity quota table, capacity pre-allocation and local filling processing are performed to obtain a grid selection sequence for candidate grid cells in the candidate grid cell set.

[0084] Based on the grid selection sequence and the attitude candidate set, the component layout scheme of the photovoltaic modules in the distributed photovoltaic scenario is determined.

[0085] In one possible implementation, the evaluation unit is used for:

[0086] Based on the effective irradiance sequence, determine the annual cumulative effective irradiance and shading risk factor within the available boundary range under each of the component layout schemes;

[0087] Based on the acceptable capacity of the aforementioned transformer area, determine the average acceptable capacity under each of the aforementioned component layout schemes;

[0088] The component layout scheme is verified based on the candidate paving surface attribute set to obtain a verification result identifier;

[0089] The power generation potential index is determined based on the annual cumulative effective irradiance, the shading risk factor, the average acceptable capacity, and the verification result identifier.

[0090] Thirdly, embodiments of this application provide a computer device, including: a memory and a processor;

[0091] The memory stores computer-executed instructions;

[0092] The processor executes computer execution instructions stored in the memory, causing the processor to perform the first aspect and / or various possible implementations of the first aspect as described above.

[0093] Fourthly, embodiments of this application provide a computer-readable storage medium storing computer-executable instructions, which, when executed by a processor, are used to implement the first aspect and / or various possible implementations of the first aspect.

[0094] Fifthly, embodiments of this application provide a computer program product, including a computer program that, when executed by a processor, implements the first aspect and / or various possible implementations of the first aspect.

[0095] The distributed photovoltaic (PV) power generation potential assessment method, apparatus, and equipment provided in this application can acquire multi-source scene data of distributed PV scenarios and perform cross-modal registration processing on the multi-source scene data according to a preset fusion dimension to obtain a data cube corresponding to the distributed PV scenario. Then, based on the data cube, candidate laying surfaces corresponding to the distributed PV scenario and the laying surface attribute information of each candidate laying surface can be determined. This enables accurate identification of candidate laying surfaces under the distributed PV scenario, thereby improving the accuracy of the laying surface attribute information of candidate laying surfaces and providing a real and reliable physical constraint basis for subsequent power generation potential assessment. Next, based on the irradiance information of each candidate laying surface, the effective irradiance sequence of each candidate laying surface can be determined, thereby determining the hourly and effective irradiance information of each candidate laying surface. This allows for the improvement of the accuracy of the irradiance information of candidate laying surfaces through dynamic irradiance information. Then, based on the load sequence, effective irradiance sequence, and candidate installation areas of the distributed photovoltaic (PV) scenario, the acceptability of the distribution substations under the distributed PV scenario can be determined. Furthermore, based on the effective irradiance sequence, the acceptability of the distribution substations, and the attribute set of the candidate installation areas, a power generation potential index can be determined to assess the PV power generation potential under the distributed PV scenario. This implementation method comprehensively considers the temporal characteristics of PV output and the grid acceptability, thereby organically coupling the distributed PV power generation potential with the actual operation of the distribution network, thus improving the accuracy, reliability, and feasibility of the power generation potential assessment. Attached Figure Description

[0096] The accompanying drawings, which are incorporated in and form part of this specification, illustrate embodiments consistent with this application and, together with the description, serve to explain the principles of this application.

[0097] Figure 1 A flowchart illustrating a method for assessing the potential of distributed photovoltaic power generation, provided in an embodiment of this application;

[0098] Figure 2 A flowchart illustrating another method for assessing the potential of distributed photovoltaic power generation provided in this application embodiment;

[0099] Figure 3 This is a schematic diagram of the structure of a distributed photovoltaic power generation potential assessment device provided in an embodiment of this application;

[0100] Figure 4 This is a schematic diagram of the structure of a computer device provided in an embodiment of this application.

[0101] The accompanying drawings illustrate specific embodiments of this application, which will be described in more detail below. These drawings and descriptions are not intended to limit the scope of the concept in any way, but rather to illustrate the concept of this application to those skilled in the art through reference to particular embodiments. Detailed Implementation

[0102] Exemplary embodiments will now be described in detail, examples of which are illustrated in the accompanying drawings. When the following description relates to the drawings, unless otherwise indicated, the same numbers in different drawings denote the same or similar elements. The embodiments described in the following exemplary embodiments do not represent all embodiments consistent with this application. Rather, they are merely examples of apparatuses and methods consistent with some aspects of this application as detailed in the appended claims.

[0103] First, some terms used in the embodiments of this application will be explained.

[0104] A distribution area (or transformer substation) is the smallest management unit of a distribution network, centered on a transformer, with a limited power supply radius, serving a certain number of users. It is the spatial carrier for calculating the time series and average capacity of a distribution area. For example, a distribution area containing several low-voltage outgoing lines and dozens of smart meters is a distribution area.

[0105] With the acceleration of the new energy transition, distributed photovoltaic (PV) power generation, as an important component in promoting a low-carbon energy structure, is rapidly becoming widespread in urban building clusters, industrial parks, and rural areas. To fully utilize the power generation performance of distributed PV, its potential is typically assessed before its implementation.

[0106] Traditional potential assessment methods mainly utilize the area and sunlight conditions of building rooftops or vacant land to evaluate the potential of distributed photovoltaic power generation through static irradiance and photovoltaic module arrangement.

[0107] This implementation method results in a relatively simple set of evaluation factors for the potential of distributed photovoltaic power generation, which fails to effectively reflect the dynamic fluctuation characteristics of distributed photovoltaic modules on an hourly scale, thus affecting the accuracy of the evaluation of the potential of distributed photovoltaic power generation.

[0108] In addition, the connection of distributed photovoltaic power directly affects the distribution network of the transformer substation, and the output characteristics of distributed photovoltaic power may cause problems such as voltage over-limit, line overload or reverse power flow at the nodes of the distribution network.

[0109] Therefore, traditional potential assessment methods neglect the impact of distributed photovoltaic (PV) output characteristics on the distribution network after it is connected to the distribution area, thus ignoring the impact of dynamic changes in the distribution network's capacity on the potential of distributed PV power generation. This leads to a mismatch between the planned power generation capacity and the actual power generation capacity of distributed PV, further affecting the accuracy of distributed PV power generation potential assessment.

[0110] The distributed photovoltaic (PV) power generation potential assessment method provided in this application can determine candidate deployment surfaces by using data cubes obtained after registering and fusing multi-source scenario data, thus improving the accuracy of the determined candidate deployment surfaces. Based on this, by determining the effective irradiance sequence of the candidate deployment surfaces, hourly and effective irradiance information of the candidate deployment surfaces is determined, thereby improving the accuracy of irradiance information of the candidate deployment surfaces through dynamic irradiance information. Simultaneously, by determining the acceptability of the distribution area under the distributed PV scenario and combining it with the effective irradiance sequence of the candidate deployment surfaces, a power generation potential index is determined to assess the PV power generation potential under the distributed PV scenario. This improves the reliability and accuracy of distributed PV scenario assessment, making the planned scenarios of distributed PV more closely match the actual application scenarios, thereby solving the aforementioned technical problems.

[0111] The technical solution of this application and how the technical solution of this application solves the above-mentioned technical problems are described in detail below with specific embodiments. These specific embodiments can be combined with each other, and the same or similar concepts or processes may not be described again in some embodiments. The embodiments of this application will now be described with reference to the accompanying drawings.

[0112] Figure 1 This is a flowchart illustrating a method for assessing the potential of distributed photovoltaic power generation, as provided in an embodiment of this application. Figure 1 As shown, the method includes:

[0113] S101. Obtain multi-source scene data of distributed photovoltaic scenarios, and perform cross-modal registration processing on the multi-source scene data according to the preset fusion dimension to obtain the data cube corresponding to the distributed photovoltaic scenario.

[0114] Optionally, the distributed photovoltaic power generation potential assessment method provided in this application embodiment can be applied to a server.

[0115] Specifically, a server is a computing device used to centrally perform data acquisition, data processing, data storage and service publishing. It is usually deployed in a data center or cloud platform and has a multi-core central processing unit, high-speed solid-state storage, graphics processing unit or tensor processing acceleration, 10 Gigabit network and redundant power supply. It can carry out parallel processing of multi-source data and long-cycle task scheduling.

[0116] Optionally, distributed photovoltaic (PV) scenarios refer to PV power generation environments that utilize building rooftops and small ground-mounted sites as carriers and are distributed across transformer substations or the low-voltage side of the distribution network. These scenarios are characterized by dispersed installation locations, complex shading factors, varying load curves, and limited capacity. For example, a transformer substation covering residential areas, commercial buildings, and schools typically has hundreds of candidate rooftop installation surfaces, each with significantly different shading and structural conditions.

[0117] For example, in a city-level distributed photovoltaic scenario, the server runs data acquisition services, orthophoto reconstruction services, and point cloud preprocessing services, and provides query capabilities to the upper-level planning platform through application programming interfaces.

[0118] In the embodiments of this application, multi-source scene data can indicate scene data obtained according to various remote sensing observation methods. For example, multi-source scene data can be obtained based on scene data obtained from aerial platforms and / or satellite platforms.

[0119] Optionally, the preset fusion dimension indicates the common dimension of multi-source scene data. For example, the preset fusion dimension may include, but is not limited to, spatial index dimension and time index dimension.

[0120] Based on this, the data cube can be a storage structure for fusion of multimodal and multitemporal data constructed according to a unified spatial index dimension and a unified temporal index dimension, thereby enabling the aligned storage of multi-source scene data on the same grid and the same time axis.

[0121] S102. Based on the data cube, determine the candidate paving surface attribute set corresponding to the distributed photovoltaic scenario; wherein, the candidate paving surface attribute set includes candidate paving surfaces and paving surface attribute information for each candidate paving surface.

[0122] In one example, a candidate paving surface refers to a set of usable roof areas that meet the requirements of geometric flatness, reasonable orientation, controllable shading, and maintenance access. It is the direct target for subsequent hourly effective irradiance estimation and component layout optimization. For example, the connected polygonal areas remaining after removing skylights, inner parapet buffer zones, and densely populated equipment areas are candidate paving surfaces.

[0123] Optionally, the paving surface attribute information of the candidate paving surface can indicate the physical set attributes of the candidate paving surface. For example, the paving surface attribute information may include, but is not limited to: the building to which the candidate paving surface belongs, its location, planar shape, planar area, planar slope, and planar structural strength, etc.

[0124] S103. Based on the irradiation information of each candidate paving surface, determine the effective irradiation sequence for each candidate paving surface.

[0125] Optionally, the irradiation information for each candidate paving surface can indicate the planar irradiation of the candidate paving surface.

[0126] In one example, the effective radiation sequence indicates hourly effective irradiance information for candidate paving surfaces. For instance, under the effective radiation sequence, the acceptability of the same substation at midday on sunny days in spring and autumn is higher than that in the afternoon in winter, and the time series curves differ significantly with seasonal and load structure variations.

[0127] S104. Based on the electricity load sequence, effective irradiance sequence, and candidate installation areas of the distributed photovoltaic scenario, determine the acceptability of the distribution area under the distributed photovoltaic scenario.

[0128] In one example, the electricity load sequence in a distributed photovoltaic (PV) scenario refers to the power and voltage operation data recorded at fixed time steps, with each transformer substation as a unit. For example, the electricity load sequence may include, but is not limited to, active power, reactive power, voltage, and power factor. Furthermore, it may also include the sampling period, timestamps, and data quality identifiers.

[0129] Optionally, the electricity load sequence for a distributed photovoltaic (PV) scenario can be obtained from smart meters or distribution transformer monitoring devices aggregated by the concentrator. In this case, the time granularity for obtaining the electricity load sequence for the distributed PV scenario can be 15 minutes or 1 hour, etc. Based on this time granularity, an hourly changing electricity load sequence can be obtained. For example, a residential area experiences peak electricity consumption in the summer evening, with an increase in the active power curve and a decrease in the power factor. In this case, the hourly changing electricity load sequence can affect the allowable PV injection space after the distributed PV modules are connected to the grid.

[0130] In one example, the acceptability of a distribution area in a distributed photovoltaic scenario can be used to reflect the time series of the upper limit of the photovoltaic injection that the distribution network can accept throughout the entire period. At the same time, it can reflect the grid-connected nodes and phases, as well as the corresponding constraint labels and data quality identifiers.

[0131] S105. Determine the power generation potential index based on the effective irradiance sequence, the acceptability of the transformer area, and the attribute set of the candidate paving surface; wherein, the power generation potential index is used to evaluate the photovoltaic power generation potential in the distributed photovoltaic scenario.

[0132] Optionally, the power generation potential index can indicate the comprehensive development potential of the candidate installation surface; a higher value indicates higher photovoltaic power generation potential. In this case, the power generation potential of photovoltaic modules arranged in different ways can be compared based on the power generation potential index.

[0133] Optionally, the higher the value of the power generation potential index, the more likely the corresponding candidate installation area will be included in the project implementation list, so as to improve the performance of distributed photovoltaic power generation.

[0134] Optionally, if the power generation potential index is low, the corresponding candidate paving surface can be marked as a verification area to ensure the power generation performance of distributed photovoltaics.

[0135] As described above, this embodiment of the application can acquire multi-source scene data of a distributed photovoltaic (PV) scenario and perform cross-modal registration processing on the multi-source scene data according to a preset fusion dimension to obtain a data cube corresponding to the distributed PV scenario. Then, based on the data cube, candidate laying surfaces corresponding to the distributed PV scenario and the laying surface attribute information of each candidate laying surface can be determined. This enables accurate identification of candidate laying surfaces under the distributed PV scenario, thereby improving the accuracy of the laying surface attribute information of the candidate laying surfaces and providing a real and reliable physical constraint basis for subsequent power generation potential assessment. Next, based on the irradiance information of each candidate laying surface, the effective irradiance sequence of each candidate laying surface can be determined, thereby determining the hourly and effective irradiance information of each candidate laying surface. This allows for the improvement of the accuracy of the irradiance information of the candidate laying surfaces through dynamic irradiance information. Then, based on the load sequence, effective irradiance sequence, and candidate installation areas of the distributed photovoltaic (PV) scenario, the acceptability of the distribution substations under the distributed PV scenario can be determined. Furthermore, based on the effective irradiance sequence, the acceptability of the distribution substations, and the attribute set of the candidate installation areas, a power generation potential index can be determined to assess the PV power generation potential under the distributed PV scenario. This implementation method comprehensively considers the temporal characteristics of PV output and the grid acceptability, thereby organically coupling the distributed PV power generation potential with the actual operation of the distribution network, thus improving the accuracy, reliability, and feasibility of the power generation potential assessment.

[0136] Figure 2 A flowchart illustrating another method for assessing the potential of distributed photovoltaic power generation provided in this application embodiment is shown below. Figure 2 As shown, in this embodiment... Figure 1 Based on the examples, a detailed description of the method for assessing the potential of distributed photovoltaic power generation is provided. This method includes:

[0137] S201. Obtain multi-source scene data of distributed photovoltaic scenarios, and perform cross-modal registration processing on the multi-source scene data according to the preset fusion dimension to obtain the data cube corresponding to the distributed photovoltaic scenario.

[0138] In this application embodiment, multi-source scene data may include, but is not limited to: remote sensing image data, UAV orthophoto data, and airborne lidar point cloud data.

[0139] Remote sensing image data refers to surface image data acquired by imaging sensors mounted on satellites or high-altitude airborne platforms. It includes visible light, multispectral, or panchromatic bands and features wide coverage and stable spatiotemporal revisiting. For example, multispectral satellite imagery with a resolution better than one meter can be used to identify roof materials and rough roof outlines, providing prior information for subsequent high-precision reconstruction and verification.

[0140] UAV orthophoto data refers to surface image stitching data obtained through multi-view aerial photography by UAVs, followed by aerial triangulation and terrain correction. This data possesses a uniform scale and geometrically distortion-free characteristics on a plane, exhibiting centimeter-level spatial resolution and accurately reflecting the planar position of objects. For example, executing a gridded flight path in densely populated urban areas can yield UAV orthophotos with a resolution of three to five centimeters, which can be used to precisely identify parapet walls, skylights, air conditioner units, and maintenance access routes.

[0141] Airborne lidar point cloud data refers to a collection of discrete points in space obtained by emitting pulses and receiving echoes from lidar sensors mounted on a flight platform, followed by positioning and attitude determination calculations. Each point cloud data contains attribute information such as three-dimensional coordinates and echo intensity. Airborne lidar point cloud data can be further classified into ground points and non-ground points. For example, airborne lidar point cloud data with a point density of ten to twenty points per square meter can be used to reconstruct digital surface models and digital terrain models, and can be used to extract information such as roof plan, slope, orientation, and obstacle boundaries.

[0142] In practice, once the target area for evaluation in the distributed photovoltaic scenario is determined, a satellite mission queue and UAV flight path planning can be established in the mission management platform based on the target area and the boundary of the transformer substation. Constraints on the same time window and maximum cloud cover threshold are set, and imaging mode, resolution, and coverage redundancy are specified. Ground control points are deployed and coordinate reference, elevation reference, and marking method are recorded. UAV flight altitude, lateral overlap, and forward overlap are configured, as are airborne lidar point density, scanning frequency, and flight path interval. Equipment verification, time synchronization, and sensor parameter archiving are completed before fieldwork. Remote sensing images, UAV orthophotos, and airborne lidar point clouds are acquired according to plan, and a collection registration list containing acquisition time, platform attitude, sensor model, and weather conditions is generated as input for subsequent processing.

[0143] Based on this, if the preset fusion dimension includes at least the spatial index dimension and the temporal index dimension, then for the above steps: according to the preset fusion dimension, cross-modal registration processing is performed on the multi-source scenario data to obtain the data cube corresponding to the distributed photovoltaic scenario. For details, please refer to the process described below.

[0144] S2011. Perform a first preprocessing on the remote sensing image data to obtain processed remote sensing image data; and perform a second preprocessing on the airborne lidar point cloud data to obtain processed airborne lidar point cloud data.

[0145] Optionally, the first preprocessing step for the remote sensing image data can be described as follows: performing radiometric calibration, geometric correction, and stripe joining on the remote sensing image data, followed by seam optimization and color balancing. At this point, the resulting processed remote sensing image data is remote sensing image data with a unified georeferenced identity.

[0146] Optionally, the process of performing a second preprocessing on the airborne lidar point cloud data can be described as follows: after performing time synchronization, attitude calculation, echo demixing, noise point removal and ground feature classification on the airborne lidar point cloud data, and distinguishing between ground points and non-ground points, the quality-checked point cloud data is output.

[0147] S2012. Based on the UAV orthophoto data and key point matching method, the processed remote sensing image data is corrected to obtain corrected remote sensing image data; and based on the UAV orthophoto data and three-dimensional projection method, the processed airborne lidar point cloud data is corrected to obtain corrected airborne lidar point cloud data.

[0148] Optionally, key point matching methods can include ground control point matching and feature point matching. In this case, the UAV orthophoto can be used as a high-precision reference base map. The processed remote sensing image is corrected and reprojected through ground control point matching and feature point matching. Residual evaluation and local fine-tuning are performed on high-identity elements such as building edges and road centerlines, thereby realizing the correction processing of the processed remote sensing image data and obtaining corrected remote sensing image data.

[0149] Optionally, when correcting the processed airborne lidar point cloud data, the airborne lidar point cloud data can be projected onto the UAV orthophoto coordinate system, and the local geometric deformation of the processed airborne lidar point cloud data can be constrained and corrected according to the spatial consistency between the building outline and the roof ridge line, so as to obtain the corrected airborne lidar point cloud data.

[0150] Subsequently, cross-modal registration processing of remote sensing image data, UAV orthophoto data, and airborne lidar point cloud data can be completed based on residual vectors, control point errors, and feature overlap.

[0151] Optionally, in addition to completing the cross-modal registration process, this embodiment of the application can also generate a registration quality report for subsequent review.

[0152] S2013. Based on the ground points in the corrected airborne lidar point cloud data, determine the digital terrain model of the distributed photovoltaic scene, and based on the non-ground points in the corrected airborne lidar point cloud data, determine the digital surface model of the distributed photovoltaic scene.

[0153] Optionally, the ground points in the calibrated airborne lidar point cloud data can be interpolated to obtain a digital terrain model, and the non-ground points in the calibrated airborne lidar point cloud data can be rasterized to obtain a digital surface model.

[0154] S2014. Based on the constructed spatial index dimension and temporal index dimension, write the corrected remote sensing image data, UAV orthophoto data, corrected airborne lidar point cloud data, digital terrain model and digital surface model into tiles to obtain a tiled data set.

[0155] Optionally, after writing the tiles, you can also add the corresponding acquisition time, processing version and source identifier to each tile so that the spatial index and time index can be retrieved in a consistent manner.

[0156] S2015. Determine at least one quality mask for the tiled data set, and determine a data cube based on the tiled data set and at least one quality mask.

[0157] Optionally, the quality mask may include, but is not limited to: cloud mask, shadow mask, occlusion mask, and registration quality mask. In this case, a cloud mask can be generated based on cloud cover discrimination and spectral thresholding; a shadow mask can be generated based on illuminance abrupt changes and solar geometric constraints; an occlusion mask can be generated based on digital surface model viewpoint analysis; and a registration quality mask can be generated based on registration residual thresholding. These four quality masks can then be used as data quality identifiers for the corresponding tiles.

[0158] Afterwards, the tiled data set (i.e., the tile sequence) with data quality identifiers is entered into the database and checked for consistency. Then, it is written into the data cube, and search catalogs and metadata indexes are built by tile, by time, and by layer to facilitate data retrieval.

[0159] S202. Based on the data cube, determine the candidate paving surface attribute set corresponding to the distributed photovoltaic scenario; wherein, the candidate paving surface attribute set includes candidate paving surfaces and paving surface attribute information for each candidate paving surface.

[0160] For specific implementation, please refer to the process described in S2021 to S2025 below.

[0161] S2021. Perform multi-scale semantic segmentation processing on the corrected remote sensing image data and UAV orthophoto data to obtain semantic segmentation results; wherein, the semantic segmentation results include the first paving surface.

[0162] S2022. Based on the non-ground points in the corrected airborne lidar point cloud data, perform plane fitting processing on the distributed photovoltaic scene to obtain the plane fitting result; wherein, the plane fitting result includes the second paving surface.

[0163] S2023. After cross-verification of the first and second paving surfaces, candidate paving surfaces for distributed photovoltaic scenarios are obtained.

[0164] S2024. Based on the roof type information, structural load-bearing level information and paving surface boundary information corresponding to each candidate paving surface, determine the paving surface attribute information of each candidate paving surface.

[0165] S2025. Determine the candidate paving surface attribute set based on the candidate paving surfaces and the paving surface attribute information of each candidate paving surface.

[0166] In one example, multi-scale semantic segmentation refers to a semantic segmentation method that classifies each pixel at multiple spatial scales within an image pyramid. By simultaneously capturing the overall shape of the roof and the details of small components, it improves the robustness of recognition under complex backgrounds and shadow interference. For example, it can distinguish between roofs and non-roofs at a large scale, identify the boundaries between pitched and flat roofs at a medium scale, and separate equipment obstacles such as air conditioner outdoor units, antennas, and vents at a small scale.

[0167] In one example, plane fitting refers to estimating locally optimal planes among non-ground points in calibrated airborne lidar point cloud data to determine the geometry of roof areas. For instance, normal vector statistics and region growing methods can be used to ensure boundary continuity and robustness, resulting in better plane fitting results. Exemplarily, plane fitting can be used to identify clusters of sloping planes consistent with the roof surface on pitched roofs, or clusters of horizontal planes on flat roofs. In these cases, the facade and abnormally high-slope areas can be separated based on the plane fitting results.

[0168] Optionally, the roof type information indicates the classification of the roof structure to which the candidate paving surface belongs. For example, the roof type information can be concrete flat roof, metal panel pitched roof, tile roof, corrugated steel roof, and roof garden, etc. In this case, the roof type information can be used to constrain the fixing method, maintenance path, and load assumptions of photovoltaic modules. For example, it can be used to constrain clamp-type fixing and slope-following arrangement for metal panel pitched roofs, or to constrain cement block or ballast-type fixing and orientation for concrete flat roofs.

[0169] Optionally, structural load-bearing capacity rating information refers to a classification of the additional load capacity that the candidate paving surface can withstand, based on information such as roof type, beam and column span, age, point cloud flatness, and equipment density. In this case, structural load-bearing capacity rating information can be used to limit the maximum layout density and concentrated load threshold. For example, a concrete flat roof with a higher structural load-bearing capacity rating allows for higher component density and fewer ventilation gaps, while a color steel roof with a lower structural load-bearing capacity rating requires reduced component density and additional maintenance access.

[0170] Optionally, the surface boundary information refers to a vectorized description of the geometric range of the candidate surface. In this case, the surface boundary information may include, but is not limited to, information such as: outer boundary, usable boundary, internal hole maintenance channel buffer zone, and boundary setback distance. This surface boundary information can be used to directly constrain the position and spacing of photovoltaic modules during the layout phase. For example, within the outer boundary of a flat roof, internal holes and buffer zones are set according to drainage outlets and maintenance channels to determine the usable boundary that meets construction and operation and maintenance requirements.

[0171] Optionally, after determining the paving surface attribute information of the candidate paving surface, corresponding semantic tags can be added to the paving surface attribute information of the candidate paving surface and stored in the data cube.

[0172] Then, the candidate paving surface attribute set can be determined based on the candidate paving surfaces and the paving surface attribute information of each candidate paving surface.

[0173] In one possible implementation, when performing multi-scale semantic segmentation on corrected remote sensing image data and UAV orthophoto data, a multi-scale pyramid hierarchy can be built first, and multi-scale semantic segmentation can be performed at each scale to obtain the category probability raster and instance outline at each scale.

[0174] Subsequently, the low-confidence areas of clouds and shadows can be suppressed according to the quality mask, and the results of each scale can be fused across scales under a unified spatial index to obtain the image area set corresponding to the first paving surface. The fusion weight is jointly adjusted by edge confidence and texture stability, as shown in the following formula (1).

[0175] (1)

[0176] in, Indicates pixel position; This indicates a pixel category label, limited to rooftops, device obstacles, and non-rooftop backgrounds; Represents the pyramid scale index; Indicated in scale Category The probability of segmentation; Representing scale In pixels The fusion weights are obtained by normalizing the edge confidence and texture stability, ensuring that edge regions are biased towards high-resolution layers and homogeneous regions are biased towards low-resolution layers. The above fusion achieves cross-scale consistency through probability weighting, avoiding single-scale missed detections and oversegmentation.

[0177] Optionally, when performing plane fitting processing on the distributed photovoltaic scene based on the non-ground points in the corrected airborne lidar point cloud data, normal estimation and flatness assessment can be performed on the non-ground points by constructing a local neighborhood, and a set of roof plane clusters corresponding to the second paving surface can be generated based on robust plane fitting. At the same time, facades and high-slope abnormal areas are removed to obtain the plane fitting result. In this case, the plane fitting is based on minimizing the point-to-plane residual, as shown in the formula (2) below.

[0178] (2)

[0179] in, This represents the set of point indices within the current candidate cluster; Represents the three-dimensional coordinates of point cloud data; Represents the unit normal vector; This represents the transpose of the unit normal vector; Indicates plane offset; constraint Used to eliminate scale indeterminacy.

[0180] At this point, the stability of the planar cluster under noise and sparsity conditions can be guaranteed by least squares fitting based on the above formula (2).

[0181] The image region set corresponding to the first paving surface and the roof plane cluster set corresponding to the second paving surface are aligned across modes under a unified spatial index. The category coverage, normal consistency and boundary overlap are calculated, and a pairing candidate set is generated by screening with weighted scores. Low-confidence pairings are corrected by multi-temporal completion and boundary extrapolation rules, thereby realizing the cross-verification of the first paving surface and the second paving surface, and obtaining the final candidate paving surface. The weighted score can be referred to the formula (3) below.

[0182] (3)

[0183] in, A polygon representing an image region; The projected polygon representing a planar cluster; Indicates area measurement; and This represents the unit normal vector corresponding to the image region and the plane cluster. This represents the weighting coefficient, which is obtained after calibration based on historical scenarios. Indicates the weighting coefficients for category coverage. Weighting coefficients indicating boundary overlap. Weighting coefficients for indicating normal consistency.

[0184] At this point, the above formula (3) can be used to select the optimal pair under both geometric and semantic consistency, and use the time consistency rule to repair the occasional gaps caused by occlusion and shadow, thereby improving the accuracy of the obtained candidate paving surfaces.

[0185] In one example, after the candidate paving surface is determined, the paving surface boundary information can be generated based on polygon calculation, including but not limited to the outer boundary, available boundary and internal holes. The constructable space is constructed with the maintenance channel buffer zone and the boundary setback distance. Equipment obstacles and elevation change lines are deducted from the available boundary in the form of buffer to form an effective area that meets the construction and maintenance constraints. For details, please refer to the formula (4) below.

[0186] (4)

[0187] in, Represents the outer boundary polygon; Indicates the inward distance applied to the outer boundary. Offset operation; This represents the set of objects that need to be avoided, including equipment obstacles and instances of elevation change lines; Representation Object By distance The generated buffer; Indicates the available boundary polygon.

[0188] At this point, the workable space can be constructed using the set difference according to the above formula (4), ensuring geometric feasibility and reserving maintenance access.

[0189] In this embodiment of the application, the roof type of the candidate paving surface can be determined based on image texture features, point cloud data normal statistics and building ledger records. On this basis, the structural load-bearing level information can be determined by combining flatness index, equipment density and structural data mapping. At this time, the structural load-bearing level information can be mapped by weighted scoring and grading threshold. For details, please refer to the following formula (5).

[0190] (5)

[0191] in, Indicates the structural load-bearing capacity score; The flatness index is obtained by normalizing the plane fitting residuals; The equipment density is obtained by normalizing the number of equipment obstacles per unit area. The reliability factor of the building ledger is derived by mapping the degree of matching between age information and structural type. The weighting coefficient is determined by expert prior knowledge and historical verification. This represents a monotonic piecewise mapping function from rating to discrete levels.

[0192] At this point, based on the integration of geometric information and archival information, the structural load-bearing rating corresponding to the structural load-bearing level information can be obtained.

[0193] At this point, the roof type information, structural load-bearing level information and paving surface boundary information are bound to the candidate paving surface at the instance level to generate a candidate paving surface attribute set, and a registration quality score is calculated for each instance to reflect cross-modal consistency and data completeness. The registration quality score can be determined based on the registration residual and the quality mask coverage, as shown in the following formula (6).

[0194] (6)

[0195] in, Indicates the registration quality score; The normalized mean of the cross-modal registration residuals is obtained by combining the errors of control points and feature points; This indicates the quality mask coverage rate, and counts the proportion of cloud and fog masks, shadow masks, and registration quality masks within the instance range; This represents the residual sensitivity coefficient, used to adjust the sensitivity of the score to changes in the residual. At this point, the above scoring can be used to measure instance credibility across both geometric alignment and quality completeness, ultimately forming a candidate paving surface attribute set containing instance identifiers, geometric information, roof type, structural load-bearing capacity, and registration quality score, which is then written into the data cube.

[0196] S203. Based on the irradiation information of each candidate paving surface, determine the effective irradiation sequence for each candidate paving surface.

[0197] In practice, the irradiation information for each candidate paving surface can be determined by referring to the process described in S2031 to S2034 below.

[0198] S2031. Based on the solar trajectory calculation method, determine the solar driving information for each candidate paving surface.

[0199] In one example, the solar trajectory calculation method refers to the process of calculating the solar azimuth and solar altitude angles in a local horizontal coordinate system based on geographical latitude, longitude, and time. This is used to determine the incident geometry and the period of sunshine attainability at different time steps. For example, within a time window from 9:00 AM to 5:00 PM in a certain area during summer, the solar azimuth and solar altitude angles are calculated for each hour to determine the incident direction and potential direct irradiance accessibility of the candidate paving surface at each time.

[0200] In one example, the solar azimuth and solar altitude angles at each time step can be determined according to the solar trajectory calculation method, thereby obtaining solar driving information. For details, please refer to formulas (7) and (8) below.

[0201] (7)

[0202] (8)

[0203] in, The solar zenith angle; The latitude of the center of the distribution area; The solar declination can be approximated by the day number; The hour angle; The unit solar direction vector; The solar altitude angle satisfies ; This is the solar azimuth angle.

[0204] After determining the solar drive information, the solar drive information can be bound to the corresponding candidate paving surface instance as a unified input for shading ray tracing and irradiance transpose.

[0205] S2032. Based on the digital surface model of each candidate paving surface, perform shading ray tracing processing on the solar drive information to obtain shading constraint information, and determine the sky view factor and ground time domain factor based on the paving surface tilt angle of each candidate paving surface.

[0206] Optionally, a digital surface model refers to regular raster data containing the elevations of the ground and the surfaces of buildings, trees, and structures. It is used to depict occlusion relationships and slope morphology in real space and is typically reconstructed from airborne LiDAR point clouds or multi-view imagery. For example, in an old residential area, a digital surface model can reflect the height differences between the varying rooftops and adjacent tree canopies, which can be used to subsequently identify which time periods are characterized by shadows cast by tree canopies onto the rooftops.

[0207] Optionally, shading ray tracing refers to emitting detection rays along the sun's direction from a candidate paving surface onto a digital surface model, determining whether the rays are blocked by surrounding features, and thus obtaining the shading status, duration, and percentage of shading. In this case, shading ray tracing can be used to determine the reduction effect of shadows on irradiance. For example, on the roof of a low-rise factory building south of a high-rise office building, shading ray tracing can identify the duration of shading by the building's projection during the winter afternoon, thereby reducing the estimated available irradiance during that period.

[0208] In practice, the shading light tracking processing of solar driving information can be carried out according to the digital surface model, the direct irradiance accessibility can be determined hourly and the shading duration can be counted; at the same time, the sky view factor and the ground view factor can be calculated to establish the shading and view constraint set, as shown in formulas (9) to (11) below.

[0209] (9)

[0210] (10)

[0211] (11)

[0212] in, For time step The occlusion percentage, that is, the occlusion constraint information; Sky view factor; Ground view factor; The paving surface tilt angle for candidate paving surfaces is derived from normal statistics. At this point, the visibility of shadow reduction and the contributions of scattering and reflection can be quantified using the formula described above.

[0213] S2033. Obtain meteorological observation data; among which, the meteorological observation data includes global horizontal irradiance data and surface albedo.

[0214] Optional meteorological observation data may also include, but are not limited to: direct normal irradiance, diffuse horizontal irradiance, cloud cover, aerosol optical thickness, precipitable water, near-surface temperature and near-surface wind speed, along with station numbers, spatial interpolation weights and quality labels.

[0215] At this point, meteorological observation data can be used to integrate ground-based radiation station observations, numerical weather prediction products, and remote sensing data such as cloud cover, aerosols, and precipitable water into irradiance estimation. This allows for dynamic adjustments to irradiance under clear-sky conditions to reflect the attenuation and fluctuations of irradiance caused by actual weather conditions. Simultaneously, by combining near-surface temperature and wind speed information, the impact of photovoltaic module temperature rise can be corrected to more closely approximate the actual usable irradiance level. For example, in situations where cloud cover changes rapidly at different times in the same city, the effective irradiance for the corresponding period can be reduced based on meteorological observation data, while the effective irradiance corresponding to photovoltaic modules can be appropriately increased during periods of strong winds and temperature drops.

[0216] S2034. Based on global horizontal irradiance data, surface albedo, solar drive information, sky view factor, and ground time domain factor, determine the irradiance information for each candidate paving surface.

[0217] Optionally, the irradiation information of the candidate paving surface can indicate the planar irradiation of the candidate paving surface. In this case, the irradiation information of the candidate paving surface can be determined according to the following formulas (12) to (13).

[0218] (12)

[0219] (13)

[0220] in, For time step Global horizontal irradiance data; Direct normal irradiation; For scattering horizontal radiation; The solar zenith angle; Irradiation information for candidate paving surfaces; The incident angle is defined as the unit normal vector of the candidate paving surface. With unit solar direction vector The included angle satisfies ; It is the surface albedo; and As above.

[0221] At this point, the above formula can be transposed to adapt to different poses and surrounding scenes under the set of occlusion and view constraints.

[0222] In one possible implementation, the meteorological observation data also includes ambient temperature information. In this case, the effective irradiation sequence of each candidate paving surface is determined based on the irradiation information of each candidate paving surface, as can be seen in the following process.

[0223] First, the temperature correction factor is determined based on the ambient temperature and irradiance information.

[0224] Then, based on the irradiation information, temperature correction coefficient, preset incident angle correction coefficient, preset surface contamination factor, and shading constraint information of each candidate paving surface, the effective irradiation sequence of each candidate paving surface is determined.

[0225] For specific implementation, please refer to formulas (14) to (16) below.

[0226] (14)

[0227] (15)

[0228] (16)

[0229] in, For time step The temperature of the component cells; This refers to ambient temperature information. The nominal operating temperature constant; The component energy efficiency after temperature correction; This is a reference energy efficiency under standard testing conditions; This is a temperature correction factor (negative value); For effective irradiation sequences; This is a preset incident angle correction factor used to determine the angle loss caused by glass encapsulation and edge reflection; These are preset surface contamination factors, determined based on maintenance records or empirical statistics; The above refers to the occlusion constraint information.

[0230] At this point, attitude, occlusion, weather, and component physical response can be unified to an hourly scale, outputting an effective irradiation sequence that corresponds one-to-one with candidate paving surface instances. Furthermore, data quality identifiers derived from the mask and data source can be attached for direct use during subsequent joint optimization of substation acceptability and component layout.

[0231] S204. Based on the electricity load sequence, effective irradiance sequence, and candidate installation areas of the distributed photovoltaic scenario, determine the acceptability of the distribution area under the distributed photovoltaic scenario.

[0232] For specific implementation, please refer to the steps described in S2041 to S2045 below.

[0233] S2041. After performing data alignment processing on the power load sequence and the effective irradiation sequence, an aligned power load sequence and an aligned effective irradiation sequence are obtained.

[0234] Optionally, when performing data alignment processing on the power load sequence and the effective irradiance sequence, time step consistency, missing measurement interpolation, and abnormal amplitude limiting processing can be completed based on a unified time index. During the alignment process, a weighted alignment function can be used to remap sequences with different sampling periods to a common time step, as shown in the following formula (17).

[0235] (17)

[0236] in, Indicates common time step A unified value on; Represents the original time step Observations on; Indicates surrounding A collection of windows; This indicates the weight of time proximity and quality identifier.

[0237] S2042. Determine the transformer substation topology information and the equipment status information of deployed equipment in the distributed photovoltaic scenario.

[0238] Optionally, the transformer area topology information can indicate the transformer area topology.

[0239] Optionally, the equipment status information may include, but is not limited to: line parameters, transformer parameters, voltage regulating device status, parallel capacitor status, and switch status.

[0240] S2043. Based on the mapping relationship table between the candidate laying surface and the injection node of the distribution network in the distributed photovoltaic scenario, and the parameter constraint information of each injection node, determine the acceptance capacity prediction model.

[0241] In one example, the mapping relationship table can be represented by a mapping matrix. In this case, after determining the mapping matrix, the output of the candidate laying surface can be aggregated to the grid node and phase, as shown in the formula (18) below.

[0242] (18)

[0243] in, Indicates time Potential active power injection vector for each candidate paving surface; Indicates time Aggregated active power injection vector for each phase of each grid-connected node; It is a sparse mapping matrix from candidate paving surface instances to nodes and phases.

[0244] Optionally, constraint parameter information includes, but is not limited to, node voltage thresholds. Upper limit of line and transformer current Reverse current allowable threshold Permissible range of phase imbalance and inverter reactive power strategy set .

[0245] Optionally, the acceptability prediction model can predict the upper limit of photovoltaic injection that can be safely accepted at different time steps under constraints such as topology, line and transformer parameters, voltage regulation device and parallel capacitor status, grid-connected inverter reactive power strategy, and phase imbalance tolerance range.

[0246] S2044. Input the aligned power load sequence, aligned effective irradiance sequence, transformer area topology information, and equipment status information into the acceptability prediction model to determine the upper limit of acceptability for the transformer area in the distributed photovoltaic scenario.

[0247] In practical implementation, under the constraints of transformer area topology information, equipment status information, and parameter constraint information, the rolling power flow solution and sensitivity scan can be performed through the acceptability prediction model. The injection power is increased proportionally along the candidate injection set, and the moment when the constraint is first triggered is recorded as the upper bound of the acceptability of the transformer area, and the constraint label is output. At this time, the analytical expression of the upper bound based on linearized sensitivity can be seen in the following formula (19).

[0248] (19)

[0249] in, Indicates time The acceptable upper bound for proportional injection; , , These are the baseline node voltage, branch current, and tie line power, respectively. For nodes Voltage allowable boundaries (upper or lower limit is determined by the first triggered side); , , These represent the sensitivity projections of the injection direction on node voltage, branch current, and tie-line power, respectively. The physical quantity corresponding to the minimum value among the three sets of data within the curly braces is the constraint label (e.g., voltage exceeding limits, line overload, or reverse power flow). For example, at noon on a weekday in a low-load suburban area, voltage exceeding limits may be the dominant constraint; while during the high-load evening peak, line thermal capacity may be the dominant constraint. In this case, this acceptable upper bound can be used as the starting point for rolling power flow, and verified by the full power flow check, thereby ensuring the accuracy of the determined acceptable upper bound.

[0250] S2045. After applying a weighted smoothing process to the upper bound of acceptability, the acceptability of the transformer area is obtained.

[0251] Specifically, the acceptable upper bound of the distribution area can be checked for time series consistency and anomaly suppression based on the confidence index of the effective irradiation sequence and the data quality index of the distribution area's power load time series data to obtain the acceptable capacity of the distribution area. For example, weighted robust smoothing can be used to suppress occasional spikes to obtain the acceptable capacity of the distribution area, as shown in formula (20).

[0252] (20)

[0253] in, This refers to the smoothed-out acceptability, which is also the acceptability of the station area; For Huber loss or Tukey loss; For time The overall weight is determined by the irradiation confidence weight. With load quality weight Multiplying them together yields the result; This is a local time window.

[0254] Optionally, after determining the acceptability of the operating area, extended verification can be performed based on the operating conditions to obtain uncertainty labeling information covering multiple operating conditions.

[0255] Optionally, multiple operating conditions may include, but are not limited to: high temperature and high load days, low temperature and low load days, cloudy / sunny transition days, and equipment maintenance days. In this case, the rolling power flow solution and sensitivity scan are repeated under each operating condition to synthesize the interval and statistics, as shown in formulas (21) to (24) below.

[0256] (twenty one)

[0257] (twenty two)

[0258] (twenty three)

[0259] (twenty four)

[0260] in, For working conditions Next time Acceptability; and These are the lower and upper bounds of the uncertainty interval; and The mean and standard deviation of the scenario are used to generate uncertainty labels and robustness assessments.

[0261] Optionally, after determining the acceptability of the substation, the uncertainty labeling information, the acceptability of the substation, the constraint labels, and the acceptability upper bound of the substation can be written into the acceptability layer of the data cube to form the acceptability time series of the substation, and a bidirectional index can be established with the effective irradiation sequence.

[0262] S205. Based on the candidate paving surface attribute set, determine the candidate paving mesh cell set, and determine the attitude candidate set for each candidate paving mesh cell in the candidate paving mesh cell set.

[0263] In practice, the available boundary is discretized regularly with a fixed grid step size, and grid cells that fall into holes, maintenance channel buffer zones and obstacle avoidance buffer zones are removed to obtain a set of candidate laying grid cells.

[0264] Then, the attitude candidate set corresponding to each candidate mesh cell can be determined. At this time, the attitude candidate set includes at least one attitude candidate cell, and the attitude parameters corresponding to the attitude under each attitude candidate cell are composed of discrete combinations of tilt angle and azimuth.

[0265] In the embodiments of this application, a pose availability score and an occlusion sensitivity score can be calculated for each pose candidate unit, and the pose availability label and occlusion sensitivity label of each pose candidate unit can be determined accordingly.

[0266] Optionally, the attitude availability score can be determined based on the hourly effective irradiance weighted average on an annual scale, as shown in formula (25) below.

[0267] (25)

[0268] in, Indicates the candidate grid cell index; Indicates the attitude index; This represents the set of time steps within the evaluation period; This indicates the weight of each time step, which can be set according to season or business focus. Indicates at time step Grid cells ,attitude Effective irradiation at specific times.

[0269] Optionally, shadow instability can be determined based on the occlusion sensitivity score, which can be determined according to the following formula (26).

[0270] (26)

[0271] in, This represents the time series of the occlusion percentage; Indicates the 95th percentile; Indicates variance; These are the weighting coefficients.

[0272] Optionally, corresponding thresholds can be set for the pose availability score and the occlusion sensitivity score based on experience. For example, a first threshold can be set for the pose availability score, and if the pose availability score is less than the first threshold, the pose candidate unit is considered unavailable; and / or, a second threshold can be set for the occlusion sensitivity score, and if the pose availability score is greater than the second threshold, the pose candidate unit is considered unavailable.

[0273] S206. Based on the acceptable capacity of the transformer area, determine the hourly capacity quota table, and based on the hourly capacity quota table, perform capacity pre-allocation and local filling processing to obtain the grid selection sequence for candidate laying grid units in the candidate laying grid unit set.

[0274] Optionally, the hourly capacity quota table can indicate the capacity limit for each grid-connected node in the distribution area.

[0275] At this point, when determining the hourly capacity quota table, the acceptable capacity of the transformer area can be expanded according to the grid-connected nodes and phases to obtain the capacity limit of each grid-connected node phase at each time step. Then, a mapping from candidate deployment grid cells to grid-connected node phases can be established, forming a sparse mapping matrix. Based on this sparse mapping matrix, the hourly capacity quota table can be determined, which can then be used to statistically analyze the hourly grid-connected occupancy of any chosen combination.

[0276] Based on this, after pre-allocating and partially filling the available candidate cells according to the hourly capacity quota table, a selection sequence of candidate laying grid cells can be obtained.

[0277] Optionally, capacity pre-allocation is based on time-by-time constraints, prioritizing candidate units with high annual production capacity per unit area. In this case, the revenue index can be determined according to the following formula (27).

[0278] (27)

[0279] in, Indicates profitability metrics; Indicates from grid cell The wiring path cost to the nearest inverter site candidate is a weighted value that combines path length, roof crossing penalty, and construction difficulty. It represents the normalization coefficient of the structural bearing capacity margin; the smaller the margin, the greater the penalty.

[0280] Optionally, after the capacity pre-allocation process, the local filling process can be carried out in order of the revenue indicators from high to low, and the hourly capacity quota table can be checked at each step. The specific check method can be found in the following formula (28).

[0281] (28)

[0282] in, The elements of the mapping matrix represent grid cells. Whether to be integrated into the grid node phase ; For grid cells In posture The nominal power below; For time step The selection variable can be set when the layout is static. ; For grid connection nodes phase At time step The maximum capacity.

[0283] Optionally, if a constraint is triggered, roll back the current selection and skip to the next candidate.

[0284] Through the above process, the component selection sequence for each candidate grid cell can be determined.

[0285] S207. Based on the grid selection sequence and attitude candidate set, determine the module layout scheme of photovoltaic modules in the distributed photovoltaic scenario.

[0286] Optionally, after completing the component layout and attitude determination based on the grid selection sequence and the available attitudes in the attitude candidate set, the inverter site candidate point set can be determined first, and the connectivity can be verified based on the cable path reachability and branch length constraints, thereby generating an inverter model, site and branch grouping table.

[0287] Optionally, cable path accessibility can be determined based on the set of accessible cable paths. In this case, the accessible cable path can satisfy the requirements of not crossing prohibited areas and meeting the minimum turning radius.

[0288] Optionally, the branch length constraint can indicate that the path length of each bus branch does not exceed the upper limit. In this case, for each inverter site candidate point, the shortest path distance and expected current of the connected grid cell set can be calculated to verify the voltage drop. The voltage drop verification method can be found in the following formula (29).

[0289] (29)

[0290] in, branch road Pressure drop estimate; For branch current estimation, calculations are made based on the series and parallel connection of components and representative irradiance values; The equivalent resistance of the line is calculated based on wire diameter and length. This is the upper limit of the pressure reduction.

[0291] At this point, while ensuring connectivity, voltage drop, and branch length, the model and location of each inverter are determined, and a branch grouping table and phase access suggestions are output.

[0292] Then, the inverter model, station location and branch grouping table can be bound with the attitude candidate set and grid selection sequence to form a component layout scheme, and written into the layout layer of the data cube.

[0293] Optionally, the component layout scheme includes a component installation vector layer, candidate laying grid cells and attitude parameter table, inverter model and location, combiner branch grouping table and cable path, grid connection node and phase allocation table, as well as hourly capacity verification report and trigger constraint statistics.

[0294] Optionally, in this embodiment, component layout schemes under different priority dimensions can be generated to obtain a component layout scheme set. For example, the component layout scheme set may include component layout schemes under capacity priority, component layout schemes under robustness priority, and component layout schemes under construction priority. Then, the association between instance identifiers and unified spatial indexes and candidate paving surface attribute sets, effective irradiation sequences, and the acceptability of the substation area can be established.

[0295] Based on this, for step S205 above: determining the power generation potential index according to the effective irradiation sequence, the acceptability of the substation area and the attribute set of the candidate laying surface, the following process can be referred to.

[0296] S208. Based on the effective irradiance sequence, determine the annual cumulative effective irradiance and shading risk factor within the available boundary range under each component layout scheme.

[0297] Optionally, annual cumulative effective irradiance refers to the annual-scale irradiance obtained by accumulating the effective irradiance over an hourly time step during the assessment period. It can be used to determine the long-term usable effective irradiance level of candidate paving surfaces. For example, candidate paving surfaces with less shading and reasonable orientation have higher annual cumulative effective irradiance.

[0298] Optionally, when determining the annual cumulative effective irradiance, when mapping the effective irradiance sequence within the available boundary range under the component layout scheme, the candidate paving surface can be used as the instance unit, and the effective irradiance sequence can be weighted and accumulated in the actual occupied area of ​​the component according to the grid cell and attitude parameters to obtain the annual cumulative effective irradiance weighted by the component density and attitude parameters, as shown in the formula (30) below.

[0299] (30)

[0300] in, Indicates a candidate paving surface instance; Representation of instances The set of candidate laying grid cells within; Represents grid cells The set of attitude parameters (discrete combination of tilt angle and azimuth). Indicates whether the layout selects a pose. Grid cells ; Indicates component density (component coverage or equivalent power density weight). Represents the area of ​​a grid cell; This represents the set of hourly time steps within the evaluation period; Indicates the weight of each time step; Indicates at time step Grid cells ,attitude Effective irradiation below.

[0301] At this point, the calculation results of the cumulative effective irradiance for that year can be written back to the instance-level feature table as annual-scale descriptive information of the energy supply intensity on the energy consumption side.

[0302] Optionally, the shading risk factor is used as a comprehensive indicator to quantify the adverse effects of shading on power generation. For example, the shading risk factor can be used to combine shading duration, shading intensity, and shading instability to reflect the risk level of shading on the effective irradiance sequence and power output fluctuations. For instance, candidate paving surfaces that are adjacent to tall buildings and cause long-term projections have significantly higher shading risk factors.

[0303] Therefore, when determining shading risk factors, shading duration, shading intensity distribution, and shading instability can be calculated based on the effective irradiation sequence to obtain the shading risk factors. Specifically, the shading proportion within an instance can be weighted according to the grid selection sequence and component density to obtain the instance-level shading trajectory. Then, determine the duration of occlusion, the distribution of occlusion intensity, and the instability of occlusion, as detailed in formulas (31) to (33) below.

[0304] (31)

[0305] (32)

[0306] (33)

[0307] in, For the duration of occlusion; The 95th percentile of the shading intensity distribution; Collection of time periods for weather transition. The instability of occlusion on the surface.

[0308] Based on this, the occlusion risk factor can be determined according to the following formula (34).

[0309] (34)

[0310] in, These are pre-calibrated weighting coefficients.

[0311] S209. Based on the acceptable capacity of the transformer area, determine the average acceptable capacity under each component layout scheme.

[0312] Optionally, the average acceptability of a distribution area is used to characterize the average acceptability of the area during the assessment period. In this case, the average acceptability can be used to quantify the long-term average load capacity of the distribution area for photovoltaic injection, and can be refined to the grid connection node and phase. For example, distribution areas with higher loads and better voltage regulation conditions usually have a larger average acceptability.

[0313] Optionally, when determining the average acceptable capacity, the average acceptable capacity of the transformer substation can be obtained by statistically analyzing the time coverage ratio and average value of the capacity quota based on the time sequence of the acceptable capacity of the transformer substation and the mapping table of candidate laying surfaces, grid connection nodes and phases. For details, please refer to the formulas (35) to (36) below.

[0314] (35)

[0315] (36)

[0316] in, Indicates grid connection node phase At time step Maximum capacity; For example The normalized weights of the phases to the nodes (determined by the shortest path, grid connection distance, or selected station topology, and) ); This represents the capacity availability threshold. This represents the average capacity of the area to accommodate passengers. This refers to the time coverage ratio of the capacity quota.

[0317] S210. Based on the candidate paving surface attribute set, verify the component layout scheme and obtain the verification result identifier.

[0318] Optionally, the candidate paving surface attribute set may also include structural margin, where structural margin indicates the ratio between the maximum load-bearing capacity of the structure and the actual working load.

[0319] Optionally, the verification result identifier may include, but is not limited to: a structural load-bearing capacity pass indicator and a structural composite recommendation identifier.

[0320] Based on this, when identifying the verification results, the consistency of the component layout scheme set can be checked according to the structural load-bearing level and structural margin in the candidate paving surface attribute set, and a structural load-bearing pass mark and a structural review recommendation mark can be generated.

[0321] Optionally, the verification result identifier can be determined based on the load utilization rate, as shown in formulas (37) to (39) below.

[0322] (37)

[0323] (38)

[0324] (39)

[0325] in, Indicates the carrying capacity utilization rate; The areal density constant of the component (including supports and ballast); , representing the instance area-weighted component density; The area of ​​the instance; For additional permanent load; Representative snow load or climate-related load; This refers to the allowable bearing capacity given based on the structural bearing capacity level and structural margin. For structural support, identification is required; Structural review recommendation identifier; Indicates an indicator function; This is to verify the bandwidth threshold.

[0326] S211. Determine the power generation potential index based on annual cumulative effective irradiance, shading risk factor, average acceptable capacity, and verification result labeling.

[0327] Optionally, a power generation potential index can be obtained by integrating annual cumulative effective irradiance, average acceptable capacity, verification result indicators, and shading risk factors. In this case, the potential index can indicate the comprehensive development potential of candidate paving surfaces; a higher value indicates sufficient irradiance, network availability, structural feasibility, and lower shading risk, facilitating horizontal comparison and prioritization. For example, candidate paving surfaces ranking high in the potential index are given priority for inclusion in the project implementation list.

[0328] Optionally, when assessing the photovoltaic power generation potential in distributed photovoltaic scenarios based on the power generation potential index, candidate installation surfaces with high potential indices can be marked as priority access areas, while candidate installation surfaces with low potential indices and high shading risk factors can be marked as verification areas, etc.

[0329] Optionally, when integrating the annual cumulative effective irradiance, the average acceptable capacity, the verification result identifier, and the shading risk factor, normalization can be performed first, followed by weighted fusion to obtain the power generation potential index. In this case, when performing normalization, refer to the formulas (40) to (42) below, and determine the power generation potential index according to the formula (43) below.

[0330] (40)

[0331] (41)

[0332] (42)

[0333] (43)

[0334] in, , , These are the normalized values ​​of annual cumulative effective irradiance, average acceptable capacity, and shading risk factor, respectively. For fusion weights; This is the structural verification penalty coefficient; It is a potential index.

[0335] like The potential index can then be set to zero or significantly reduced according to the rules; if A review penalty will then be imposed to reflect the uncertainty of the project.

[0336] Optionally, you can Write the potential layer into the data cube for potential assessment and prioritization.

[0337] Figure 3 This is a schematic diagram of the structure of a distributed photovoltaic power generation potential assessment device provided in an embodiment of this application, as shown below. Figure 3As shown, the distributed photovoltaic power generation potential assessment device 30 provided in this embodiment includes:

[0338] The acquisition unit 301 is used to acquire multi-source scene data of the distributed photovoltaic scenario, and perform cross-modal registration processing on the multi-source scene data according to the preset fusion dimension to obtain the data cube corresponding to the distributed photovoltaic scenario.

[0339] The first determining unit 302 is used to determine the candidate paving surface attribute set corresponding to the distributed photovoltaic scenario based on the data cube; wherein, the candidate paving surface attribute set includes candidate paving surfaces and paving surface attribute information of each candidate paving surface.

[0340] The second determining unit 303 is used to determine the effective irradiation sequence of each candidate paving surface based on the irradiation information of each candidate paving surface.

[0341] The third determining unit 304 is used to determine the acceptability of the distribution photovoltaic (PV) substation area based on the power load sequence, effective irradiance sequence, and candidate installation surfaces of the distributed PV scenario.

[0342] Evaluation unit 305 is used to determine the power generation potential index based on the effective irradiance sequence, acceptability and candidate paving surface attribute set; wherein, the power generation potential index is used to evaluate the photovoltaic power generation potential in distributed photovoltaic scenarios.

[0343] In one possible implementation, the multi-source scene data includes at least remote sensing image data, UAV orthophoto data, and airborne lidar point cloud data; the preset fusion dimensions include at least a spatial index dimension and a temporal index dimension; in this case, the acquisition unit 301 is used for:

[0344] The remote sensing image data is subjected to a first preprocessing process to obtain processed remote sensing image data; and the airborne lidar point cloud data is subjected to a second preprocessing process to obtain processed airborne lidar point cloud data.

[0345] Based on UAV orthophoto data and key point matching methods, the processed remote sensing image data is corrected to obtain corrected remote sensing image data. Based on UAV orthophoto data and three-dimensional projection methods, the processed airborne lidar point cloud data is corrected to obtain corrected airborne lidar point cloud data.

[0346] Based on the ground points in the corrected airborne lidar point cloud data, the digital terrain model of the distributed photovoltaic scene is determined, and based on the non-ground points in the corrected airborne lidar point cloud data, the digital surface model of the distributed photovoltaic scene is determined.

[0347] Based on the constructed spatial index dimension and temporal index dimension, the corrected remote sensing image data, UAV orthophoto data, corrected airborne lidar point cloud data, digital terrain model and digital surface model are written into tiles to obtain a tiled data set.

[0348] Determine at least one quality mask for the tiled dataset, and determine a data cube based on the tiled dataset and the at least one quality mask.

[0349] In one possible implementation, the first determining unit 302 is configured to:

[0350] Multi-scale semantic segmentation processing is performed on the corrected remote sensing image data and UAV orthophoto data to obtain semantic segmentation results; the semantic segmentation results include the first paving surface;

[0351] Based on the non-ground points in the corrected airborne lidar point cloud data, a plane fitting process is performed on the distributed photovoltaic scene to obtain the plane fitting result; the plane fitting result includes the second paving surface;

[0352] After cross-verification of the first and second paving surfaces, candidate paving surfaces for distributed photovoltaic scenarios are obtained.

[0353] Based on the roof type information, structural load-bearing level information and paving surface boundary information corresponding to each candidate paving surface, the paving surface attribute information of each candidate paving surface is determined.

[0354] Based on the candidate paving surfaces and the paving surface attribute information of each candidate paving surface, the candidate paving surface attribute set is determined.

[0355] In one possible implementation, the device is also used for:

[0356] Based on the solar trajectory calculation method, the solar driving information of each candidate paving surface is determined;

[0357] Based on the digital surface model of each candidate paving surface, the shading ray tracing process is performed on the solar driving information to obtain shading constraint information, and the sky view factor and ground time domain factor are determined according to the paving surface tilt angle of each candidate paving surface.

[0358] Acquire meteorological observation data; including global horizontal irradiance data and surface albedo.

[0359] Irradiance information for each candidate paving surface is determined based on global horizontal irradiance data, surface albedo, solar drive information, sky view factor, and ground time factor.

[0360] In one possible implementation, the meteorological observation data also includes ambient temperature information; in this case, the second determining unit 303 is used to:

[0361] The temperature correction factor is determined based on ambient temperature and irradiance information;

[0362] The effective irradiation sequence for each candidate paving surface is determined based on the irradiation information, temperature correction factor, preset incident angle correction factor, preset surface contamination factor, and shading constraint information of each candidate paving surface.

[0363] In one possible implementation, the third determining unit 304 is configured to:

[0364] After data alignment processing of the electricity load sequence and the effective irradiance sequence, aligned electricity load sequence and aligned effective irradiance sequence are obtained;

[0365] Determine the transformer topology information and equipment status information of deployed devices in a distributed photovoltaic scenario;

[0366] Based on the mapping relationship table between the candidate laying surface and the injection node of the distribution network in the distributed photovoltaic scenario, as well as the parameter constraint information of each injection node, the acceptance capacity prediction model is determined.

[0367] The aligned power load sequence, aligned effective irradiance sequence, transformer area topology information, and equipment status information are input into the acceptance capacity prediction model to determine the upper limit of acceptance for the transformer area in the distributed photovoltaic scenario.

[0368] After applying a weighted smoothing process to the upper bound of acceptability, the acceptability of the transformer area is obtained.

[0369] In one possible implementation, the device is also used for:

[0370] Before determining the power generation potential index based on the effective irradiation sequence, the acceptability of the substation area, and the attribute set of the candidate paving surface, the set of candidate paving grid cells is determined based on the attribute set of the candidate paving surface, and the attitude candidate set of each candidate paving grid cell in the set of candidate paving grid cells is determined.

[0371] Based on the acceptability of the transformer area, an hourly capacity quota table is determined, and based on the hourly capacity quota table, capacity pre-allocation and local filling processing are performed to obtain a grid selection sequence for candidate grid cells in the candidate grid cell set.

[0372] Based on the grid selection sequence and attitude candidate set, the module layout scheme of photovoltaic modules in the distributed photovoltaic scenario is determined.

[0373] In one possible implementation, the evaluation unit 305 is used for:

[0374] Based on the effective irradiance sequence, determine the annual cumulative effective irradiance and shading risk factor within the available boundary range under each component layout scheme;

[0375] Based on the acceptable capacity of the transformer area, determine the average acceptable capacity under each component layout scheme;

[0376] Based on the candidate paving surface attribute set, the component layout scheme is verified to obtain the verification result identifier;

[0377] The power generation potential index is determined based on the annual cumulative effective irradiance, shading risk factor, average acceptable capacity, and verification result labeling.

[0378] The distributed photovoltaic power generation potential assessment device provided in this embodiment can execute the method provided in the above method embodiment. Its implementation principle and technical effect are similar, and will not be described in detail here.

[0379] Figure 4 This is a schematic diagram of the structure of a computer device provided in an embodiment of this application. Figure 4 As shown, the computer device 40 provided in this embodiment includes at least one processor 401 and a memory 402. Optionally, the computer device 40 further includes a communication component 403. The processor 401, memory 402, and communication component 403 are connected via a bus 404.

[0380] In a specific implementation, at least one processor 401 executes computer execution instructions stored in memory 402, causing at least one processor 401 to perform the above-described method.

[0381] The specific implementation process of processor 401 can be found in the above method embodiments, and its implementation principle and technical effect are similar. It will not be repeated here.

[0382] In the above embodiments, it should be understood that the processor can be a Central Processing Unit (CPU), or other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), etc. The general-purpose processor can be a microprocessor or any conventional processor. The steps of the method disclosed in this invention can be directly implemented by a hardware processor, or implemented by a combination of hardware and software modules within the processor.

[0383] The memory may include random access memory (RAM) and may also include non-volatile memory (NVM), such as at least one disk storage device.

[0384] The bus can be an Industry Standard Architecture (ISA) bus, a Peripheral Component Interconnect (PCI) bus, or an Extended Industry Standard Architecture (EISA) bus, etc. Buses can be categorized as address buses, data buses, control buses, etc. For ease of illustration, the buses shown in the accompanying drawings are not limited to a single bus or a single type of bus.

[0385] This application also provides a computer program product, including a computer program that, when executed by a processor, implements the above-described method.

[0386] This application also provides a computer-readable storage medium storing computer-executable instructions, which, when executed by a processor, implement the above-described method.

[0387] The aforementioned readable storage medium can be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as static random access memory (SRAM), electrically erasable programmable read-only memory (EEPROM), erasable programmable read-only memory (EPROM), programmable read-only memory (PROM), read-only memory (ROM), magnetic storage, flash memory, magnetic disk, or optical disk. The readable storage medium can be any available medium accessible to a general-purpose or special-purpose computer.

[0388] An exemplary readable storage medium is coupled to a processor, enabling the processor to read information from and write information to the readable storage medium. Of course, the readable storage medium can also be a component of the processor. The processor and the readable storage medium can reside in an Application Specific Integrated Circuit (ASIC). Alternatively, the processor and the readable storage medium can exist as discrete components in the device.

[0389] The division of units is merely a logical functional division; in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be indirect coupling or communication connection through some interfaces, devices, or units, and may be electrical, mechanical, or other forms.

[0390] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.

[0391] In addition, the functional units in the various embodiments of the present invention can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit.

[0392] If a function is implemented as a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this invention, or the part that contributes to the prior art, or a part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods of the various embodiments of this invention. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0393] Those skilled in the art will understand that all or part of the steps of the above-described method embodiments can be implemented by hardware related to program instructions. The aforementioned program can be stored in a computer-readable storage medium. When executed, the program performs the steps of the above-described method embodiments; and the aforementioned storage medium includes various media capable of storing program code, such as ROM, RAM, magnetic disks, or optical disks.

[0394] Finally, it should be noted that other embodiments of the invention will readily occur to those skilled in the art upon consideration of the specification and practice of the invention disclosed herein. This invention is intended to cover any variations, uses, or adaptations of the invention that follow the general principles of the invention and include common knowledge or customary techniques in the art not disclosed herein, and is not limited to the precise structures described above and shown in the accompanying drawings, and various modifications and changes can be made without departing from its scope. The scope of the invention is limited only by the appended claims.

Claims

1. A method for assessing the potential of distributed photovoltaic power generation, characterized in that, include: Acquire multi-source scene data of distributed photovoltaic scenarios, and perform cross-modal registration processing on the multi-source scene data according to a preset fusion dimension to obtain a data cube corresponding to the distributed photovoltaic scenario; Based on the data cube, a set of candidate paving surface attributes corresponding to the distributed photovoltaic scenario is determined; wherein, the set of candidate paving surface attributes includes candidate paving surfaces and paving surface attribute information for each candidate paving surface; Based on the irradiation information of each candidate paving surface, determine the effective irradiation sequence for each candidate paving surface; Based on the electricity load sequence of the distributed photovoltaic scenario, the effective irradiance sequence, and the candidate installation area, determine the acceptability of the lower transformer area of ​​the distributed photovoltaic scenario; A power generation potential index is determined based on the effective irradiance sequence, the acceptability of the distribution area, and the attribute set of the candidate paving surface; wherein, the power generation potential index is used to evaluate the photovoltaic power generation potential under the distributed photovoltaic scenario.

2. The method according to claim 1, characterized in that, The multi-source scene data includes at least remote sensing image data, UAV orthophoto data, and airborne lidar point cloud data; the preset fusion dimension includes at least a spatial index dimension and a temporal index dimension. Based on a preset fusion dimension, cross-modal registration processing is performed on the multi-source scene data to obtain a data cube corresponding to the distributed photovoltaic scene, including: The remote sensing image data is subjected to a first preprocessing process to obtain processed remote sensing image data; and the airborne lidar point cloud data is subjected to a second preprocessing process to obtain processed airborne lidar point cloud data. Based on the UAV orthophoto data and key point matching method, the processed remote sensing image data is corrected to obtain corrected remote sensing image data. Based on the UAV orthophoto data and three-dimensional projection method, the processed airborne lidar point cloud data is corrected to obtain corrected airborne lidar point cloud data. Based on the ground points in the corrected airborne lidar point cloud data, a digital terrain model of the distributed photovoltaic scene is determined, and based on the non-ground points in the corrected airborne lidar point cloud data, a digital surface model of the distributed photovoltaic scene is determined. Based on the constructed spatial index dimension and temporal index dimension, the corrected remote sensing image data, the UAV orthophoto data, the corrected airborne lidar point cloud data, the digital terrain model, and the digital surface model are written into tiles to obtain a tiled data set. Determine at least one quality mask for the tiled data set, and determine the data cube based on the tiled data set and the at least one quality mask.

3. The method according to claim 2, characterized in that, Based on the data cube, the set of candidate paving surface attributes corresponding to the distributed photovoltaic scenario is determined, including: Multi-scale semantic segmentation processing is performed on the corrected remote sensing image data and the UAV orthophoto data to obtain semantic segmentation results; wherein, the semantic segmentation results include the first paving surface; Based on the non-ground points in the corrected airborne lidar point cloud data, a plane fitting process is performed on the distributed photovoltaic scene to obtain a plane fitting result; wherein, the plane fitting result includes a second paving surface; After cross-verification of the first paving surface and the second paving surface, candidate paving surfaces for the distributed photovoltaic scenario are obtained. Based on the roof type information, structural load-bearing level information and paving surface boundary information corresponding to each candidate paving surface, the paving surface attribute information of each candidate paving surface is determined. The candidate paving surface attribute set is determined based on the candidate paving surfaces and the paving surface attribute information of each candidate paving surface.

4. The method according to claim 1, characterized in that, The method further includes: Based on the solar trajectory calculation method, the solar driving information of each candidate paving surface is determined; Based on the digital surface model of each candidate paving surface, the solar drive information is processed by shading ray tracing to obtain shading constraint information, and the sky view factor and ground time domain factor are determined based on the paving surface tilt angle of each candidate paving surface. Acquire meteorological observation data; wherein, the meteorological observation data includes global horizontal irradiance data and surface albedo. The irradiance information of each candidate paving surface is determined based on the global horizontal irradiance data, the surface albedo, the solar drive information, the sky view factor, and the ground time domain factor.

5. The method according to claim 4, characterized in that, The meteorological observation data also includes ambient temperature information; Based on the irradiation information of each candidate paving surface, an effective irradiation sequence for each candidate paving surface is determined, including: Based on the ambient temperature information and the irradiation information, a temperature correction factor is determined; The effective irradiation sequence for each candidate paving surface is determined based on the irradiation information of each candidate paving surface, the temperature correction coefficient, the preset incident angle correction coefficient, the preset surface contamination factor, and the shading constraint information.

6. The method according to claim 1, characterized in that, Based on the electricity load sequence of the distributed photovoltaic scenario, the effective irradiance sequence, and the candidate installation areas, the acceptable capacity of the distribution area under the distributed photovoltaic scenario is determined, including: After performing data alignment processing on the power load sequence and the effective irradiation sequence, an aligned power load sequence and an aligned effective irradiation sequence are obtained. Determine the transformer substation topology information and the device status information of the deployed equipment in the distributed photovoltaic scenario; Based on the mapping relationship table between the candidate laying surface and the injection node of the distribution network in the distributed photovoltaic scenario, and the parameter constraint information of each injection node, the acceptance capacity prediction model is determined. The aligned power load sequence, the aligned effective irradiance sequence, the transformer area topology information, and the equipment status information are input into the acceptable capacity prediction model to determine the acceptable upper bound of the transformer area under the distributed photovoltaic scenario. After applying a weighted smoothing process to the upper bound of the acceptable range, the acceptable capacity of the transformer area is obtained.

7. The method according to any one of claims 1-6, characterized in that, Before determining the power generation potential index based on the effective irradiation sequence, the acceptability of the substation area, and the candidate paving surface attribute set, the method further includes: Based on the set of candidate paving surface attributes, determine the set of candidate paving mesh cells, and determine the attitude candidate set for each candidate paving mesh cell in the set of candidate paving mesh cells; Based on the acceptable capacity of the transformer area, an hourly capacity quota table is determined, and based on the hourly capacity quota table, capacity pre-allocation and local filling processing are performed to obtain a grid selection sequence for candidate grid cells in the candidate grid cell set. Based on the grid selection sequence and the attitude candidate set, the component layout scheme of the photovoltaic modules in the distributed photovoltaic scenario is determined.

8. The method according to claim 7, characterized in that, Based on the effective irradiation sequence, the acceptability of the substation area, and the attribute set of the candidate paving surface, a power generation potential index is determined, including: Based on the effective irradiance sequence, determine the annual cumulative effective irradiance and shading risk factor within the available boundary range under each of the component layout schemes; Based on the acceptable capacity of the aforementioned transformer area, determine the average acceptable capacity under each of the aforementioned component layout schemes; The component layout scheme is verified based on the candidate paving surface attribute set to obtain a verification result identifier; The power generation potential index is determined based on the annual cumulative effective irradiance, the shading risk factor, the average acceptable capacity, and the verification result identifier.

9. A distributed photovoltaic power generation potential assessment device, characterized in that, include: The acquisition unit is used to acquire multi-source scene data of the distributed photovoltaic scenario, and perform cross-modal registration processing on the multi-source scene data according to a preset fusion dimension to obtain the data cube corresponding to the distributed photovoltaic scenario. The first determining unit is configured to determine the candidate paving surface attribute set corresponding to the distributed photovoltaic scenario based on the data cube; wherein the candidate paving surface attribute set includes candidate paving surfaces and paving surface attribute information for each candidate paving surface. The second determining unit is used to determine the effective irradiation sequence of each candidate paving surface based on the irradiation information of each candidate paving surface. The third determining unit is used to determine the acceptability of the lower transformer area of ​​the distributed photovoltaic scenario based on the electricity load sequence of the distributed photovoltaic scenario, the effective irradiance sequence and the candidate laying surface; An evaluation unit is used to determine a power generation potential index based on the effective irradiance sequence, the acceptability, and the candidate paving surface attribute set; wherein the power generation potential index is used to evaluate the photovoltaic power generation potential under the distributed photovoltaic scenario.

10. A computer device, characterized in that, include: Memory, processor; The memory stores computer-executed instructions; The processor executes computer execution instructions stored in the memory, causing the processor to perform the method as described in any one of claims 1-8.

Images