A distributed photovoltaic planning method and system

Abstract

The application provides a distributed photovoltaic planning method and system, comprising: acquiring basic data of a target area, determining a target power grid node corresponding to each candidate area available for layout; generating a photovoltaic output prediction sequence of each candidate area and determining a corresponding power generation potential index; determining a corresponding consumption capacity index of each candidate area; determining a node price prediction feature corresponding to each target power grid node and predicting a corresponding node spot price sequence; determining a corresponding income adaptation index of each candidate area; constructing a planning model with the power generation potential index, the consumption capacity index and the income adaptation index as optimization targets to obtain a distributed photovoltaic planning scheme containing a recommended installed capacity of each candidate area. The method can identify candidate areas with high low-price risk or insufficient power grid carrying capacity in the planning stage, thereby optimizing the recommended installed capacity of each candidate area and improving the consumption capacity and economic income of the distributed photovoltaic planning scheme.

Description

Technical Field

[0001] This application relates to the field of power and new energy planning technology, and in particular to a distributed photovoltaic planning method and system. Background Technology

[0002] With the accelerated green and low-carbon energy transition and the full-scale advancement of the electricity spot market, the penetration rate of distributed photovoltaic (PV) power in distribution networks is increasing daily. Scientific spatial layout and capacity planning for PV have become crucial for ensuring system safety and improving project profitability. Current distributed PV planning often focuses on the abundance of solar resources, tending to concentrate deployment in areas with high solar irradiance.

[0003] Current technologies primarily employ traditional optimization models aimed at maximizing photovoltaic (PV) power generation, supplemented by static historical average electricity prices, for capacity allocation and site selection. However, this static and one-dimensional approach fails to accurately reflect the underlying physical operation of the power grid and the dynamic market clearing mechanism under conditions of high-proportion renewable energy integration. In actual operation, neglecting issues such as local energy storage regulation, nodal reverse power flow risks, transmission congestion, and concentrated oversupply during midday periods can easily lead to extremely low or even negative electricity prices during peak PV generation periods. This means that planning strategies based on traditional methods not only fail to deliver theoretically high economic returns but also exacerbate local reverse power flow by deviating from the grid's actual carrying capacity, potentially causing severe curtailment of solar power.

[0004] In summary, existing technologies are insufficient to accurately depict dynamic price risks in the spot market while simultaneously addressing the underlying physical security constraints of the power grid. This results in distributed photovoltaic planning schemes facing difficulties in grid integration and overall economic benefits falling far short of expectations. Summary of the Invention

[0005] In view of this, the purpose of this application is to provide a distributed photovoltaic planning method and system. By combining the photovoltaic output forecast of candidate areas, grid absorption capacity and spot price forecast of nodes to construct a planning model, candidate areas with high low price risk or insufficient grid carrying capacity can be identified in the planning stage, thereby optimizing the recommended installed capacity of each candidate area and improving the absorption capacity and economic benefits of distributed photovoltaic planning schemes.

[0006] In a first aspect, the present invention provides a distributed photovoltaic planning method, comprising: Acquire basic data of the target area, and based on the basic data, select at least one candidate area that can be deployed, and determine the target power grid node corresponding to each candidate area.

[0007] Based on the photovoltaic resource data in the basic data, a photovoltaic output prediction sequence for each candidate region is generated, and the power generation potential index corresponding to each candidate region is determined.

[0008] Based on the grid node operation data, grid structure parameters, and photovoltaic output prediction sequence in the basic data, the absorption capacity index corresponding to each candidate region is determined.

[0009] Based on the system supply and demand data, new energy power output forecast data, local load forecast data, local regulation resource capacity data, critical line power flow forecast data, line capacity data, and historical node spot price data in the basic data, as well as the photovoltaic power output forecast sequence, the node price forecast characteristics corresponding to each target grid node are determined, and the node spot price sequence of each target grid node is predicted based on the node price forecast characteristics.

[0010] Based on the matching relationship between the spot price series of nodes and the photovoltaic power output forecast series, the corresponding return matching index for each candidate region is determined.

[0011] A planning model is constructed with power generation potential, grid connection capacity and revenue matching as optimization objectives, and solved under preset constraints to obtain a distributed photovoltaic planning scheme that includes the recommended installed capacity for each candidate region.

[0012] In an optional implementation, the basic data includes buildable resource data, planning constraint data, and geospatial data.

[0013] The steps of acquiring basic data of the target area, filtering at least one candidate area for layout based on the basic data, and determining the target power grid node corresponding to each candidate area include: Restricted areas within the target region are identified based on planning constraint data.

[0014] The effective operating area is obtained by removing restrictive areas from the target area range corresponding to the geospatial data.

[0015] The effective working area is spatially discretized based on a preset size grid to obtain multiple basic grids.

[0016] The actual available buildable area within each basic grid is determined based on buildable resource data.

[0017] The basic grid with an actual usable construction area greater than the preset area threshold is identified as the candidate area.

[0018] Based on the geographical location of the candidate region, the location of the power grid access point, and the distribution network topology, the target power grid node corresponding to each candidate region is determined.

[0019] In an optional implementation, the photovoltaic resource data includes solar irradiance data, temperature data, and photovoltaic module parameters.

[0020] The steps of generating a photovoltaic output prediction sequence for each candidate region based on photovoltaic resource data in the basic data, and determining the corresponding power generation potential index for each candidate region, include: Extract solar irradiance data, temperature data, and photovoltaic module parameters for each candidate region from the photovoltaic resource data.

[0021] Based on solar irradiance data, temperature data, and photovoltaic module parameters, the photovoltaic output per unit installed capacity of each candidate region is calculated within the predicted time scale.

[0022] A photovoltaic output prediction sequence is generated by arranging the unit installed capacity photovoltaic output based on the time granularity of the prediction time scale.

[0023] The expected power generation of each candidate region within the prediction timescale is calculated based on the photovoltaic power output prediction sequence.

[0024] The expected power generation is normalized to obtain the power generation potential index for each candidate region.

[0025] In an optional implementation, the power grid node operation data includes node load data, line power flow data, main transformer load rate data, and bus voltage data.

[0026] The steps for determining the grid absorption capacity index for each candidate region based on grid node operation data, grid structure parameters, and photovoltaic output prediction sequences in the basic data include: The local load characteristic parameters of each target power grid node are determined based on the node load data.

[0027] Based on power grid structure parameters, line power flow data, main transformer load rate data, and bus voltage data, the power grid strength parameters of each target power grid node are determined.

[0028] Based on the photovoltaic output prediction sequence and node load data, the node reverse power flow risk parameters for each target grid node are determined.

[0029] The local load characteristic parameters, grid strength parameters, and node reverse power flow risk parameters are weighted to obtain the absorption assessment value for each candidate region.

[0030] The absorption capacity assessment values ​​are normalized to obtain the absorption capacity index corresponding to each candidate area.

[0031] In an optional implementation, the step of determining the node reverse power flow risk parameters for each target grid node based on the photovoltaic output prediction sequence and node load data includes: The distributed photovoltaic (PV) predicted output of each target grid node in each prediction period is determined based on the PV output prediction sequence.

[0032] The local forecast load for each target grid node in each forecast period is determined based on node load data.

[0033] Calculate the difference between the predicted output of distributed photovoltaic power and the predicted local load, and determine the reverse power flow prediction value for each target grid node in each prediction period for the difference that is greater than zero.

[0034] The reverse flow risk coefficient is determined based on the maximum value in the reverse flow prediction and the transformer capacity corresponding to the target grid node.

[0035] The reverse flow risk parameters of nodes are determined based on the reverse flow risk coefficient.

[0036] In optional implementations, the node price prediction characteristics include system supply and demand tension characteristics, new energy output characteristics, node reverse power flow risk characteristics, transmission congestion characteristics, and price lag characteristics.

[0037] Based on the system supply and demand data, renewable energy forecast output data, local load forecast data, local regulation resource capacity data, critical line forecast power flow data, line capacity data, and historical spot price data for each node in the basic data, as well as the photovoltaic output forecast sequence, the steps to determine the node price forecast characteristics corresponding to each target grid node include: Based on system supply and demand data, the characteristics of system supply and demand tension are determined.

[0038] Based on the predicted power output data of new energy sources and the predicted power output sequence of photovoltaic power, the characteristics of new energy power output are determined.

[0039] Based on photovoltaic power output forecast sequences, local load forecast data, and local regulation resource capacity data, the characteristics of node reverse power flow risk are determined.

[0040] Based on the predicted power flow data and line capacity data of the critical path, the characteristics of transmission congestion are determined.

[0041] Based on historical spot price data, price lag characteristics are determined.

[0042] In an optional implementation, the step of predicting the spot price sequence of each target grid node based on node price prediction characteristics includes: The node price prediction features corresponding to each target power grid node are input into the pre-trained node spot price prediction model.

[0043] The nodal spot price prediction model outputs the nodal spot price for each target power grid node in each prediction period within the prediction time scale.

[0044] Arrange the spot prices of each forecast period according to the time granularity of the forecast time scale to obtain the spot price sequence of each target power grid node.

[0045] In an optional implementation, the step of determining the revenue matching index corresponding to each candidate region based on the matching relationship between the spot price series of nodes and the photovoltaic power output prediction series includes: The photovoltaic power output forecast sequence is used to determine the photovoltaic power output forecast for each candidate region in each forecast period.

[0046] The node spot price for each candidate region corresponding to the target power grid node in each prediction period is determined based on the node spot price series.

[0047] The photovoltaic (PV) power output for each forecast period is used to weight the spot prices at the corresponding nodes for that forecast period to obtain the PV weighted average spot price.

[0048] The probability of low-price periods and / or negative electricity price periods is determined based on the spot price series of nodes.

[0049] Based on the weighted average spot price of photovoltaic power and the probability of low-price periods and / or negative electricity price periods, the corresponding revenue matching index for each candidate region is determined.

[0050] In an optional implementation, the steps of constructing a planning model with power generation potential, grid connection capacity, and revenue matching as optimization objectives, and solving it under preset constraints to obtain a distributed photovoltaic planning scheme containing the suggested installed capacity for each candidate region, include: The recommended installed capacity for each candidate region is used as the decision variable.

[0051] Based on the preset power generation weight coefficient, absorption weight coefficient, and revenue weight coefficient, the power generation potential index, absorption capacity index, and revenue matching index corresponding to each candidate region are weighted and processed to obtain the planning matching score corresponding to each candidate region.

[0052] Based on the recommended installed capacity and planning adaptation score of each candidate region, the single-point planning benefits corresponding to each candidate region are determined.

[0053] A planning model is constructed with the objective of maximizing the sum of planning benefits for each single point corresponding to all candidate areas within the target area.

[0054] The planning model is solved under the pre-defined constraints to obtain the suggested installed capacity for each candidate region.

[0055] Based on the suggested installed capacity for each candidate region, a distributed photovoltaic planning scheme is generated.

[0056] Secondly, the present invention provides a distributed photovoltaic planning system, comprising: The candidate region determination module is used to obtain basic data of the target region, and based on the basic data, filter out at least one candidate region that can be laid out, and determine the target power grid node corresponding to each candidate region.

[0057] The power generation potential determination module is used to generate a photovoltaic output prediction sequence for each candidate region based on the photovoltaic resource data in the basic data, and to determine the power generation potential index corresponding to each candidate region.

[0058] The grid absorption capacity determination module is used to determine the grid absorption capacity index corresponding to each candidate region based on the grid node operation data, grid structure parameters and photovoltaic output prediction sequence in the basic data.

[0059] The node price forecasting module is used to determine the node price forecasting characteristics corresponding to each target grid node based on the system supply and demand data, new energy power output forecasting data, local load forecasting data, local regulation resource capacity data, critical line power flow forecasting data, line capacity data, historical node spot price data, and photovoltaic power output forecasting sequence in the basic data, and to predict the node spot price sequence of each target grid node based on the node price forecasting characteristics.

[0060] The revenue matching determination module is used to determine the revenue matching index corresponding to each candidate region based on the matching relationship between the spot price series of nodes and the photovoltaic output prediction series.

[0061] The planning scheme generation module is used to construct a planning model with power generation potential indicators, absorption capacity indicators and revenue matching indicators as optimization objectives, and solve it under the condition of satisfying preset constraints to obtain a distributed photovoltaic planning scheme containing the suggested installed capacity of each candidate area.

[0062] This application provides a distributed photovoltaic (PV) planning method and system. By acquiring basic data of the target area and determining candidate areas and their corresponding target grid nodes, the distributed PV planning object can be further linked from a simple geographical area to a specific grid access node. By generating a PV output prediction sequence and determining the power generation potential index, the PV power generation capacity of each candidate area can be accurately evaluated. By combining grid node operation data, grid structure parameters, and the PV output prediction sequence to determine the absorption capacity index, the PV acceptance capacity of each candidate area under the condition of safe grid operation can be reflected. By predicting the node spot price sequence based on data such as system supply and demand, new energy output, local load, regulation resources, key line power flow, line capacity, and historical node spot prices, and combining the PV output prediction sequence to determine the revenue matching index, the degree of matching between PV output time periods and node spot price changes can be evaluated. Thus, the planning model can comprehensively consider power generation potential, grid absorption capacity, and spot price revenue matching, thereby obtaining a distributed PV planning scheme with more reasonable recommended installed capacity, stronger absorption capacity, and higher economic benefits for each candidate area.

[0063] Other features and advantages of this application will be set forth in the following description and will be apparent in part from the description or may be learned by practicing the application. The objectives and other advantages of this application are realized and obtained through the structures particularly pointed out in the description, claims and drawings.

[0064] To make the above-mentioned objectives, features and advantages of this application more apparent and understandable, preferred embodiments are described below in detail with reference to the accompanying drawings. Attached Figure Description

[0065] To more clearly illustrate the technical solutions in the specific embodiments of this application or the prior art, the drawings used in the description of the specific embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of this application. For those skilled in the art, other drawings can be obtained from these drawings without creative effort.

[0066] Figure 1 This is a flowchart of a distributed photovoltaic planning method provided in an embodiment of this application; Figure 2 A flowchart of the target power grid node determination method provided in the embodiments of this application; Figure 3 A flowchart illustrating the method for determining power generation potential indicators provided in this application embodiment; Figure 4 Flowchart of the method for determining absorption capacity indicators provided in the embodiments of this application; Figure 5Flowchart of the method for determining the revenue adaptation index provided in the embodiments of this application; Figure 6 A flowchart illustrating the method for determining distributed photovoltaic planning schemes provided in this application embodiment; Figure 7 This is a schematic diagram of a distributed photovoltaic planning system provided in an embodiment of this application.

[0067] Icons: 1-Candidate area determination module; 2-Power generation potential determination module; 3-Absorption capacity determination module; 4-Nodal price prediction module; 5-Revenue matching determination module; 6-Planning scheme generation module. Detailed Implementation

[0068] To make the objectives, technical solutions, and advantages of the embodiments of this application clearer, the technical solutions of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0069] To facilitate a better understanding of this application by those skilled in the art, a brief introduction to the application scenarios and design concepts of this application is provided.

[0070] As the scale of distributed photovoltaic (PV) grid integration continues to expand, distributed PV planning is no longer simply a matter of determining feasible construction areas and installed capacity. It also requires comprehensive consideration of the PV power generation capacity of candidate areas, the absorption capacity of corresponding grid nodes, and changes in spot prices at the nodes under the electricity market environment. Especially in scenarios with a high proportion of distributed PV grid integration, the target grid nodes corresponding to different candidate areas differ in terms of load levels, grid structure, line carrying capacity, reverse power flow risk, and spot price levels. Even if a candidate area has good solar resources, it may still be difficult to achieve ideal absorption effects and economic benefits due to insufficient local load, limited transmission capacity, or low prices during peak PV generation periods.

[0071] Existing distributed photovoltaic (PV) planning methods typically focus on factors such as solar irradiance, available construction area, or historical average returns, tending to concentrate PV capacity in areas with high solar irradiance. While these methods can increase theoretical power generation to some extent, they do not adequately consider the actual operating conditions of grid nodes and future spot price changes, making it difficult to accurately reflect the grid node's absorption capacity and price-return suitability during PV output periods. For example, when distributed PV output in a candidate area surges during midday, while the corresponding target grid node experiences low local load, insufficient regulation resources, or transmission congestion on critical lines, the target grid node is prone to reverse power flow risks, limited local absorption, and low or negative electricity prices. This results in a higher risk of curtailment and lower-than-expected returns for the installed capacity determined using traditional methods.

[0072] Based on this, this application provides a distributed photovoltaic (PV) planning method and system. This application does not simply determine the distributed PV planning scheme based on PV resource conditions, but first obtains basic data of the target area, and then filters candidate areas for layout based on the basic data, further determining the target grid node corresponding to each candidate area; then, it evaluates the candidate areas from the perspectives of PV power generation, grid absorption, and market price. Specifically, it generates a PV output prediction sequence and determines power generation potential indicators through PV resource data, enabling the planning model to reflect the power generation capacity of the candidate areas; it determines absorption capacity indicators through grid node operation data, grid structure parameters, and the PV output prediction sequence, enabling the planning model to reflect the ability of the target grid node corresponding to the candidate area to accept distributed PV power under safe operating conditions; and it determines node price prediction characteristics and predicts node spot price sequences through system supply and demand data, new energy predicted output data, local load prediction data, local regulation resource capacity data, critical line predicted power flow data, line capacity data, historical node spot price data, and the PV output prediction sequence, enabling the planning model to reflect future spot price changes of the target grid node and their matching relationship with PV output periods.

[0073] Furthermore, this application determines the revenue matching index based on the matching relationship between the nodal spot price series and the photovoltaic output forecast series. It then uses the power generation potential index, grid connection capacity index, and revenue matching index together as the optimization objectives of the planning model. Under preset constraints, the planning model is solved to obtain a distributed photovoltaic planning scheme that includes the suggested installed capacity for each candidate region. Therefore, this application can comprehensively identify candidate regions with high power generation potential, strong grid connection capacity, and a good match between photovoltaic output periods and nodal spot prices during the planning stage. This avoids concentrating photovoltaic capacity in areas with insufficient grid capacity, high reverse power flow risk, or high low-price risk, thereby improving the grid connection reliability and overall economic benefits of the distributed photovoltaic planning scheme.

[0074] To facilitate understanding of this embodiment, the embodiments of this application will be described in detail below.

[0075] This application provides a distributed photovoltaic planning method, referring to... Figure 1 The distributed photovoltaic planning method provided in this application includes: Step S101: Obtain basic data of the target area, and filter out at least one candidate area that can be deployed based on the basic data, and determine the target power grid node corresponding to each candidate area.

[0076] Here, the target area refers to the region requiring distributed photovoltaic (PV) planning, such as provincial, municipal, county, industrial park, substation power supply area, feeder power supply area, or transformer substation power supply area. Basic data describes the PV resources, grid operating status, grid structure, spot prices, available resources, and planning constraints of the target area. Candidate areas are spatial regions within the target area that meet the conditions for distributed PV construction. Target grid nodes are the substation nodes, feeder nodes, transformer substation nodes, access points, or market nodes corresponding to the candidate areas after grid connection.

[0077] Step S102: Generate a photovoltaic output prediction sequence for each candidate region based on the photovoltaic resource data in the basic data, and determine the power generation potential index corresponding to each candidate region.

[0078] Here, photovoltaic (PV) resource data is used to characterize the solar resources and PV power generation conditions of candidate regions. The PV output prediction sequence represents the predicted PV output of candidate regions arranged according to a preset time granularity within the prediction timescale. The power generation potential index characterizes the PV power generation capacity of candidate regions. By generating the PV output prediction sequence and determining the power generation potential index, the distributed PV planning method can identify candidate regions with good power generation capacity.

[0079] Step S103: Based on the grid node operation data, grid structure parameters and photovoltaic output prediction sequence in the basic data, determine the absorption capacity index corresponding to each candidate region.

[0080] Here, grid node operation data is used to characterize the load, power flow, voltage, and main transformer operating status of the target grid node. Grid structure parameters are used to characterize the grid connection relationship, line capacity, transformer capacity, and distribution network topology of the target grid node. The absorption capacity index is used to characterize the ability of the target grid node corresponding to the candidate area to accept distributed photovoltaic power under the condition of satisfying grid safety operation. By determining the absorption capacity index, the distributed photovoltaic planning method can avoid planning based solely on solar resources while ignoring grid absorption capacity.

[0081] Step S104: Based on the system supply and demand data, new energy power output forecast data, local load forecast data, local regulation resource capacity data, critical line power flow forecast data, line capacity data, and historical node spot price data in the basic data, as well as the photovoltaic power output forecast sequence, determine the node price forecast characteristics corresponding to each target grid node, and predict the node spot price sequence of each target grid node based on the node price forecast characteristics.

[0082] Here, nodal price forecast features are used to characterize the factors influencing the future spot prices of target grid nodes. Nodal spot price sequences represent the spot prices of target grid nodes for each forecast period within the forecast timescale. By forecasting nodal spot price sequences, distributed photovoltaic planning methods can consider future spot market price changes during the planning phase, rather than relying solely on historical average electricity prices.

[0083] Step S105: Based on the matching relationship between the spot price series of nodes and the photovoltaic power output prediction series, determine the revenue matching index corresponding to each candidate region.

[0084] Here, the revenue fit index is used to characterize the degree of matching between the photovoltaic (PV) output periods of the candidate region and the spot price periods of the target grid nodes. A higher revenue fit index occurs when PV output is concentrated during periods with higher spot prices at higher nodes; a lower index occurs when PV output is concentrated during periods with low prices or negative electricity prices. By determining the revenue fit index, distributed PV planning methods can identify candidate regions with good market revenue fit.

[0085] Step S106: Construct a planning model with power generation potential index, absorption capacity index and income matching index as optimization objectives, and solve it under the condition of satisfying preset constraints to obtain a distributed photovoltaic planning scheme that includes the recommended installed capacity of each candidate area.

[0086] Here, the planning model is used to allocate distributed photovoltaic (PV) capacity among multiple candidate regions. Pre-defined constraints are used to limit the planning scheme to meet requirements such as regional planning capacity, installable capacity of candidate regions, grid security operation, and node access capability. The recommended installed capacity is the distributed PV installed capacity output by the planning model for the candidate regions. By constructing and solving the planning model, the distributed PV planning method can obtain distributed PV planning schemes that take into account power generation potential, grid absorption capacity, and spot price profitability.

[0087] In an optional implementation, the basic data includes buildable resource data, planning constraint data, and geospatial data.

[0088] Reference Figure 2 Step S101 includes the following steps S201-S206.

[0089] Step S201: Determine the restricted areas within the target area based on the planning constraint data.

[0090] Here, planning constraint data is used to characterize the spatial scope and related construction restrictions within the target area that are unsuitable or not permitted for the construction of distributed photovoltaic (PV) systems. Planning constraint data includes ecological protection zones, permanent basic farmland zones, construction control zones, building structural load-bearing limits, areas with severe shading, areas already planned for other uses, areas where the safety distance from grid facilities does not meet requirements, and other restrictions affecting distributed PV construction. After acquiring the planning constraint data, data from different sources are uniformly converted to the same coordinate system and cropped according to the geographic spatial boundaries of the target area to obtain a planning constraint layer within the target area. Overlay analysis is performed on the planning constraint layers, and spatial units that meet any constraint condition are marked as restricted areas.

[0091] In one implementation, the administrative boundary layer or power supply area boundary layer of the target area is read first, followed by the ecological protection area layer, land use control layer, building attribute layer, and power grid safety distance control layer. Coordinate unification, boundary clipping, and spatial overlay processing are performed on each layer to obtain a set of restricted areas. Each restricted area in the set includes its boundary, restriction type, and reason for restriction. For example, if the structural load-bearing capacity of a rooftop area is lower than the load-bearing threshold for distributed photovoltaic installation, that rooftop area is designated as a restricted area; if an open space is located within an ecological protection area, that open space is designated as a restricted area; if the distance between an area and a substation, overhead line, or other power facility does not meet the safety distance requirements, that area is designated as a restricted area.

[0092] Step S202: Remove restrictive areas from the target area range corresponding to the geospatial data to obtain the effective operating area.

[0093] Here, geospatial data is used to describe the spatial boundaries, building locations, road locations, open space areas, rooftop areas, grid connection points, substation locations, feeder routes, and transformer substation areas of the target area. The planning system uses the target area as the initial working area and subtracts restrictive areas from it to obtain the remaining spatial area that allows for distributed photovoltaic layout analysis. This remaining spatial area is the effective working area.

[0094] In one implementation, a spatial difference operation is performed on the target area and the set of restricted areas to generate a valid work area. If a restricted area only covers a portion of a building's roof, only the portion of the roof covered by the restricted area is removed, and the remaining portion of the roof that meets the construction requirements is retained. If a restricted area divides a vacant lot into multiple discontinuous spatial blocks, each discontinuous spatial block is retained as a different sub-region of the valid work area.

[0095] Step S203: Spatial discretization of the effective working area is performed based on a preset size grid to obtain multiple basic grids.

[0096] Here, spatial discretization is used to transform continuous, effective working areas into discrete spatial units suitable for computation and optimization. The size of the preset grid can be determined based on the planning scale, data accuracy, and computational load. For example, in provincial or municipal planning, the preset grid size can be 500m × 500m or 1km × 1km; in county or industrial park planning, the preset grid size can be 100m × 100m or 200m × 200m; in district or building complex planning, the preset grid size can be 10m × 10m, 20m × 20m, or based on a single rooftop.

[0097] In one implementation, a regular grid is generated according to a preset size, using the minimum bounding rectangle of the target area as the grid generation range. The regular grid is then spatially intersected with the effective working area, retaining grids that overlap with the effective working area, and these retained grids are designated as the base grids. For base grids that cross the boundary of the effective working area, only the portion of the base grid overlapping with the effective working area is included in the subsequent usable area calculation range. If multiple discontinuous sub-regions exist within the effective working area, each sub-region is individually gridded, and the generated base grids are uniformly numbered. The base grid number may include the target area number, spatial row and column number, and grid sequence number, facilitating subsequent association with grid nodes, photovoltaic resource data, and construction area data.

[0098] Step S204: Determine the actual available buildable area within each basic grid based on buildable resource data.

[0099] Here, buildable resource data is used to characterize rooftops, open spaces, parking sheds, factory rooftops, public building rooftops, and other usable spaces within the basic grid that can be used for distributed photovoltaic (PV) installations. The planning system identifies buildable resources within each basic grid and subtracts unusable portions to obtain the actual usable construction area. Unusable portions include shaded areas, equipment-occupied areas, maintenance access areas, roof edge safety distance areas, areas where the slope does not meet requirements, and areas where the load-bearing capacity does not meet requirements.

[0100] In one implementation, the spatial area of ​​various buildable resources within the basic grid is first calculated, and then a construction utilization coefficient is configured according to the resource type. For rooftop resources, the construction utilization coefficient is related to roof type, roof orientation, roof slope, shading conditions, and maintenance access requirements; for open land resources, the construction utilization coefficient is related to land shape, terrain slope, surrounding shading, access distance, and reserved access requirements. The planning system can calculate the actual available buildable area of ​​the i-th basic grid according to the following formula. :

[0101] in, This represents the area of ​​the j-th type of constructible resource within the i-th basic grid. This represents the construction and utilization coefficient of the j-th type of constructible resource within the i-th basic grid. This represents the occlusion correction coefficient for the j-th type of buildable resource within the i-th base grid. This represents the carrying capacity correction factor for the j-th type of buildable resource within the i-th basic grid. The shading correction factor reflects the impact of building shading, tree shading, and other facility shading on the buildable area. The carrying capacity correction factor reflects the supporting capacity of the roof structure or ground foundation for distributed photovoltaic construction. If a resource type does not require shading correction or carrying capacity correction, the corresponding correction factor is set to 1.

[0102] In one implementation, rooftop resources within a base grid are identified based on remote sensing imagery, building outline data, and rooftop attribute data. The shading area is estimated based on building height, distance to surrounding buildings, and solar altitude angle. The shading area, maintenance access area, and equipment occupancy area are subtracted from the rooftop resource area to obtain the actual usable construction area of ​​the rooftop. Open space resources within the base grid are identified based on land use data. Road setback areas, boundary setback areas, and non-construction areas are subtracted from the open space resource area to obtain the actual usable construction area of ​​the open space. The actual usable construction area of ​​the rooftop and the actual usable construction area of ​​the open space are summed to obtain the actual usable construction area within the base grid.

[0103] Step S205: The basic grid with an actual usable construction area greater than the preset area threshold is determined as the candidate area.

[0104] Here, a preset area threshold is used to exclude basic grids that are too small, have low construction value, or are difficult to connect to the grid. The preset area threshold can be determined based on the minimum economic installed capacity, the installed capacity density per unit area, and construction management requirements. For example, the minimum economic installed capacity can be set first. And calculate the preset area threshold based on the installed capacity density ρ per unit area. :

[0105] in, This represents the minimum economic installed capacity for distributed photovoltaic (PV) projects. This represents the installed capacity density per unit area. It includes the actual available construction area. Greater than The basic grid was used to determine the candidate area, which will be the actual available construction area. Less than or equal to The basic grid is eliminated. The installed density per unit area can be determined based on the type of photovoltaic modules, installation method, module spacing, and inverter configuration requirements.

[0106] In one implementation, after determining the candidate areas, the maximum installable capacity of the candidate areas is estimated based on the actual available construction area. The maximum installable capacity of the candidate areas can be expressed by the following formula:

[0107] in, This represents the maximum installable capacity of the i-th candidate region. This represents the actual available construction area of ​​the i-th candidate region. This represents the installed capacity density per unit area. The maximum installable capacity is used as a constraint on the maximum installable capacity of candidate areas in subsequent planning models.

[0108] Step S206: Based on the geographical location of the candidate region, the location of the power grid access point, and the distribution network topology, determine the target power grid node corresponding to each candidate region.

[0109] Here, the target grid node refers to the substation node, feeder node, distribution area node, access point, or market node corresponding to the candidate area after it is connected to the grid. The purpose of determining the target grid node is to establish the correlation between the candidate area and grid operation data, node spot price data, and access capacity data, so that subsequent absorption capacity indicators, node price prediction characteristics, and revenue matching indicators can be calculated according to the same target grid node.

[0110] In one implementation, the spatial location of the candidate region is determined based on its geometric center point or the centroid of constructable resources, and then the spatial distance between the candidate region and each grid access point is calculated. Among the grid access points that meet the access voltage level, access capacity, and topology reachability conditions, the grid access point with the shortest access distance is selected as the candidate access point. Based on the distribution network topology, the corresponding feeder node, distribution area node, substation node, or market node is traced upwards from the candidate access point, and the traced node is determined as the target grid node.

[0111] In one implementation, a target grid node is determined based on access distance, available capacity, line capacity margin, and voltage safety margin. Multiple candidate grid nodes are generated for candidate regions, and an access evaluation value is calculated for each candidate node within the candidate region. The access evaluation value can be determined based on a weighted average of access distance, available capacity, line capacity margin, and voltage safety margin. The candidate grid node with the highest access evaluation value is then determined as the target grid node.

[0112] In an optional implementation, the photovoltaic resource data includes solar irradiance data, temperature data, and photovoltaic module parameters.

[0113] Reference Figure 3 Step S102 includes the following steps S301-S305.

[0114] Step S301: Extract solar irradiance data, temperature data, and photovoltaic module parameters corresponding to each candidate region from the photovoltaic resource data.

[0115] Here, solar irradiance data is used to characterize the solar radiation intensity received by the candidate area at different time periods. Solar irradiance data includes at least one of total horizontal irradiance, direct irradiance, diffuse irradiance, and tilted irradiance. Temperature data is used to characterize the ambient temperature of the candidate area at different time periods, and is used to correct for the power generation efficiency of the photovoltaic modules under different temperature conditions. Photovoltaic module parameters are used to characterize the power generation performance of the photovoltaic modules and inverters. Photovoltaic module parameters include at least one of the following: module rated power, module conversion efficiency, temperature coefficient, inverter efficiency, installation tilt angle, installation azimuth angle, module degradation rate, and system loss rate.

[0116] In one implementation, based on the geographical location of each candidate region, solar irradiance data and temperature data corresponding to the candidate region are extracted from meteorological databases, historical observation data, predicted meteorological data, or remote sensing data. When a candidate region is located between multiple meteorological observation points, spatial interpolation is performed on the solar irradiance data and temperature data of the multiple meteorological observation points based on the distance between the candidate region and each meteorological observation point to obtain the solar irradiance data and temperature data corresponding to the candidate region. When historical photovoltaic operation data exists for the candidate region, the historical photovoltaic operation data can also be used as verification data to correct the output calculation results corresponding to the solar irradiance data, temperature data, and photovoltaic module parameters.

[0117] Step S302: Based on solar irradiance data, temperature data, and photovoltaic module parameters, calculate the photovoltaic output per unit installed capacity for each candidate region within the predicted time scale.

[0118] Here, the forecast timescale can be the next day, the next 7 days, the next month, the next year, or a typical year within the planning period. The time granularity can be 15 minutes, 30 minutes, 1 hour, or a daily scale. Unit installed capacity photovoltaic output is used to represent the photovoltaic output that can be generated after installing one unit capacity of distributed photovoltaic power during the corresponding forecast period.

[0119] In one implementation, the effective irradiance of the module receiving surface for each forecast period is determined based on solar irradiance data and the installation tilt and azimuth angles. The module operating temperature for each forecast period is determined based on air temperature data and a module temperature model. The photovoltaic output per unit installed capacity is calculated based on the effective irradiance of the module receiving surface, the module operating temperature, the module conversion efficiency, the temperature coefficient, the inverter efficiency, and the system loss rate. The photovoltaic output per unit installed capacity can be calculated using the following formula:

[0120] in, This represents the photovoltaic output per unit installed capacity in the t-th forecast period. This represents the effective irradiance of the component's receiving surface during the t-th prediction period. Indicates irradiance under standard test conditions. This represents the component operating temperature during the t-th prediction period. This indicates the component temperature under standard test conditions. Indicates the component temperature coefficient. The inverter efficiency is represented by , and L represents the system loss rate. The system loss rate includes line losses, dust accumulation losses, shading losses, component mismatch losses, and equipment availability losses. Standard test conditions can be taken as follows: =1000W / m² =25℃.

[0121] In one implementation, PVsyst (a photovoltaic system simulation software) or an equivalent photovoltaic simulation model is invoked, using solar irradiance data, temperature data, photovoltaic module parameters, installation tilt angle, installation azimuth angle, and system loss rate as inputs to simulate the photovoltaic output per unit installed capacity for each candidate region within the predicted time scale. Alternatively, a photovoltaic output prediction model trained based on historical meteorological data and historical photovoltaic output data can be used to calculate the photovoltaic output per unit installed capacity. When using a photovoltaic output prediction model, solar irradiance data, temperature data, cloud cover data, candidate region location, and photovoltaic module parameters are input into the model, and the model outputs the photovoltaic output per unit installed capacity for each prediction period.

[0122] Step S303: Based on the time granularity of the predicted time scale, arrange the photovoltaic output per unit installed capacity to generate a photovoltaic output prediction sequence.

[0123] Here, the photovoltaic output per unit installed capacity is sorted according to the chronological order of each prediction period within the prediction timescale, forming a photovoltaic output prediction sequence corresponding to the candidate region. The photovoltaic output prediction sequence can be expressed as:

[0124] in, This represents the photovoltaic output prediction sequence for the i-th candidate region. Let T represent the photovoltaic output per unit installed capacity of the i-th candidate region in the t-th prediction period, and let T represent the total number of prediction periods within the prediction time scale.

[0125] After generating the photovoltaic (PV) power output prediction sequence, an integrity check is performed on the sequence. If missing periods exist, they are filled in using interpolation of adjacent periods, historical data from the same period, or data from similar candidate regions. If outliers exceeding the theoretical upper limit of installed capacity exist, they are corrected according to a preset output upper limit. If non-zero output occurs at night, the nighttime output is reset to zero based on sunrise / sunset times or an irradiance threshold.

[0126] Step S304: Calculate the expected power generation of each candidate region within the prediction time scale based on the photovoltaic power output prediction sequence.

[0127] Here, the expected power generation is used to characterize the total power generation that a candidate region can obtain within the predicted time scale under the condition of unit installed capacity. The expected power generation is obtained by summing or integrating the photovoltaic output per unit installed capacity corresponding to each prediction period in the photovoltaic output prediction sequence. The expected power generation can be calculated using the following formula:

[0128] in, This represents the expected power generation of the i-th candidate region within the prediction timescale. This represents the photovoltaic output per unit installed capacity of the i-th candidate region during the t-th prediction period. This represents the length of time corresponding to a single forecast period, and T represents the total number of forecast periods.

[0129] When the forecast timescale is a typical year, This represents the expected annual power generation per unit installed capacity in the i-th candidate region. The forecast timescale is either 1 day or 7 days from now. This represents the expected power generation per unit installed capacity of the i-th candidate region within the corresponding short-term forecast period. Expected power generation can also be calculated separately for different seasons, weekdays, and holidays, and then weighted and aggregated based on the number of days corresponding to each typical day to obtain the expected power generation within the planning period.

[0130] In one implementation, the photovoltaic power output forecast sequence is statistically analyzed by time period to obtain the expected power generation during the midday high-output period, the expected power generation during the low-load period, and the total expected power generation for the year. The expected power generation during the midday high-output period is used to evaluate the yield fit with the nodal spot price sequence. The expected power generation during the low-load period is used for joint analysis with the grid absorption capacity index and the nodal reverse power flow risk parameter.

[0131] Step S305: Normalize the expected power generation to obtain the power generation potential index corresponding to each candidate region.

[0132] Here, the power generation potential index is used to characterize the photovoltaic power generation capacity of a candidate region relative to other candidate regions within the target region. Normalization eliminates the influence of different dimensions on the planning model, allowing the power generation potential index to participate in subsequent planning model calculations along with the grid absorption capacity index and the revenue matching index.

[0133] In one implementation, the expected power generation is normalized using a maximum-minimum normalization method to obtain a power generation potential index. The power generation potential index can be calculated using the following formula:

[0134] in, This represents the power generation potential index corresponding to the i-th candidate region. This represents the expected power generation of the i-th candidate region. This represents the maximum expected power generation across all candidate regions. This represents the minimum expected power generation for all candidate regions. equal At that time, the power generation potential index corresponding to each candidate region is set to the same preset value, such as 0.5 or 1.

[0135] In one implementation, a quantile normalization method is used to process the expected power generation. Candidate regions are sorted in ascending order of expected power generation, and the power generation potential index is determined based on the ranking of the candidate regions. This method can reduce the impact of individual extremely high-irradiance areas or anomalous data on the normalization results. Alternatively, a standardization method can be used: first, standardized values ​​are calculated based on the mean and standard deviation of the expected power generation of all candidate regions; then, the standardized values ​​are transformed to a preset range to obtain the power generation potential index.

[0136] In an optional implementation, the power grid node operation data includes node load data, line power flow data, main transformer load rate data, and bus voltage data.

[0137] Reference Figure 4 Step S103 includes the following steps S401-S405.

[0138] Step S401: Determine the local load characteristic parameters of each target power grid node based on the node load data.

[0139] Here, nodal load data is used to characterize the changes in electricity load at target grid nodes at different time scales. Nodal load data includes historical load data, predicted load data, typical daily load curves, seasonal load curves, peak load, valley load, midday load, load during peak photovoltaic (PV) generation periods, and load growth prediction data. When determining local load characteristic parameters, the nodal load data is first organized according to the prediction time scale and time granularity to ensure consistency between the nodal load data and the PV output prediction sequence in the time dimension. The prediction time scale can be the next 1 day, the next 7 days, the next 1 month, the next 1 year, or a typical year within the planning period; the time granularity can be 15 minutes, 30 minutes, 1 hour, or a daily scale.

[0140] In one implementation, based on the load sequence of the target grid node within the predicted time scale, the average load, peak load, minimum load, midday average load, average load during peak photovoltaic (PV) generation periods, and load factor are calculated, and at least one of these parameters is used as a local load characteristic parameter. The local load characteristic parameter can be used to reflect the target grid node's ability to absorb distributed PV power output from its local electricity load. Generally, during peak PV generation periods, the higher the local load, the stronger the target grid node's ability to locally absorb distributed PV power; the lower the local load, the higher the likelihood of reverse power flow and curtailment risks at the target grid node.

[0141] In one implementation, local load characteristic parameters can be determined by the ratio of the node's average annual load to its peak load:

[0142] in, This represents the local load characteristic parameter of the i-th target grid node. This represents the average load of the i-th target grid node within the forecast timescale. This represents the peak load of the i-th target grid node within the predicted time scale. This method can reflect the stability and continuous absorption capacity of the target grid node's load level. Furthermore, local load characteristic parameters can be determined based on the ratio of average load to peak load during peak photovoltaic power generation periods to highlight the local absorption capacity during the midday peak photovoltaic output period.

[0143] Step S402: Based on the power grid structure parameters, line power flow data, main transformer load rate data, and bus voltage data, determine the power grid strength parameters for each target power grid node.

[0144] Here, grid structure parameters are used to characterize the topology and equipment capacity conditions of the distribution network where the target grid node is located. These parameters include transformer capacity, line capacity, feeder length, node connection relationships, short-circuit capacity, available capacity, and the transmission capacity of the upstream grid. Line power flow data is used to characterize the active and reactive power flow of critical lines in each forecast period. Main transformer load rate data is used to characterize the current load level and capacity margin of the transformers. Bus voltage data is used to characterize the operating level and voltage deviation of the bus voltage near the target grid node.

[0145] When determining grid strength parameters, the critical lines, associated main transformers, and associated buses corresponding to the target grid node are first identified based on the distribution network topology. Then, the line capacity margin, main transformer capacity margin, and voltage deviation margin are calculated separately. The line capacity margin can be determined based on the difference between the line capacity and the predicted line power flow. The main transformer capacity margin can be determined based on the difference between the transformer capacity and the predicted main transformer load. The voltage deviation margin can be determined based on the distance between the bus voltage and the preset upper and lower voltage limits. After normalizing the line capacity margin, main transformer capacity margin, and voltage deviation margin, a weighted average is obtained to obtain the grid strength parameters. A higher grid strength parameter indicates a stronger ability of the distribution network where the target grid node is located to accommodate new distributed photovoltaic capacity.

[0146] In one implementation, the power grid strength parameter can be determined by the following formula:

[0147] in, This represents the grid strength parameter of the i-th target grid node. This represents the line capacity margin corresponding to the i-th target power grid node. This represents the main transformer capacity margin corresponding to the i-th target grid node. This represents the voltage deviation margin corresponding to the i-th target grid node. , and These represent the weighting coefficients corresponding to line capacity margin, main transformer capacity margin, and voltage deviation margin, respectively. The weighting coefficients can be determined based on the characteristics of the distribution network operation, the planning focus, or expert experience. For example, when the main access bottleneck in the target area is insufficient line capacity, the planning system will increase the weighting coefficient. When the main access bottleneck in the target area is insufficient main transformer capacity, the planning system improves... When there is a significant risk of voltage exceeding limits in the target area, the planning system improves... .

[0148] Step S403: Based on the photovoltaic output prediction sequence and node load data, determine the node reverse power flow risk parameters for each target grid node.

[0149] Here, the node reverse power flow risk parameter is used to characterize the risk level of a target grid node feeding power back to the upper-level grid when the output of distributed photovoltaic power exceeds the local load absorption capacity. The higher the reverse power flow risk, the more likely the target grid node is to experience problems such as voltage overruns, reverse power flow, equipment overload, limited photovoltaic grid connection, or curtailment.

[0150] Step S404: The local load characteristic parameters, grid strength parameters, and node reverse power flow risk parameters are weighted to obtain the absorption assessment value corresponding to each candidate region.

[0151] Here, the absorption assessment value corresponding to the candidate region is determined based on the local load characteristic parameters, grid strength parameters, and node reverse power flow risk parameters of the target grid node corresponding to the candidate region.

[0152] In one implementation, local load characteristic parameters, grid strength parameters, and nodal reverse power flow risk parameters are normalized to the same value range before being weighted. To maintain consistency in the meaning of the indicators, nodal reverse power flow risk parameters can be converted into nodal reverse power flow safety parameters. A higher nodal reverse power flow safety parameter indicates a lower reverse power flow risk. The absorption assessment value can be determined by the following formula:

[0153] in, This represents the absorption evaluation value corresponding to the i-th candidate region. This represents the local load characteristic parameters of the target power grid node corresponding to the i-th candidate region. This represents the grid strength parameter of the target grid node corresponding to the i-th candidate region. This represents the node reverse power flow risk parameter for the target power grid node corresponding to the i-th candidate region. , and These represent the weighting coefficients corresponding to local load characteristic parameters, grid strength parameters, and nodal reverse power flow security parameters, respectively. .

[0154] In one implementation, the load absorption assessment value can be configured according to local load intensity, grid capacity, and reverse flow risk. For example, the weight corresponding to the local load characteristic parameter can be set to 0.5, the weight corresponding to the grid intensity parameter to 0.3, and the weight corresponding to the node reverse flow safety parameter to 0.2. This configuration can emphasize the local load absorption capacity while taking into account the grid structure capacity and reverse flow risk. The weights can also be adjusted according to the operational shortcomings of the target area; when the target area has insufficient main transformer or line capacity, the weight corresponding to the grid intensity parameter is increased; when the target area has a large reverse flow pressure during peak photovoltaic power generation periods, the weight corresponding to the node reverse flow safety parameter is increased.

[0155] Step S405: Normalize the absorption assessment values ​​to obtain the absorption capacity index corresponding to each candidate area.

[0156] Here, the absorption capacity index is used to characterize the ability of a candidate region to accept distributed photovoltaic power while ensuring the safe operation of the power grid. Normalization is used to convert the absorption assessment value to a uniform value range, so that the absorption capacity index, power generation potential index, and revenue matching index can be used together in the planning model calculation.

[0157] In one implementation, the absorption capacity index is determined using the max-min normalization method. The absorption capacity index can be determined by the following formula:

[0158] in, This represents the absorption capacity index corresponding to the i-th candidate region. This represents the absorption evaluation value corresponding to the i-th candidate region. This represents the maximum absorption assessment value corresponding to all candidate regions. This represents the minimum absorption assessment value corresponding to all candidate regions. If equal Set the absorption capacity index of all candidate areas to the same preset value.

[0159] In one implementation, the absorption capacity of candidate regions can be verified first based on power flow calculations. The estimated access capacity of the candidate regions is superimposed onto the target grid nodes, and power flow calculations are performed to check whether the node voltage, line current carrying capacity, and main transformer load rate meet the preset safe operating conditions. If the power flow calculation results show that the node voltage exceeds the limit, the line is overloaded, or the main transformer is overloaded, the absorption assessment value corresponding to the candidate region is reduced, or a stricter upper limit is set for the recommended installed capacity of the candidate region in the subsequent planning model. Through power flow calculation verification, the absorption capacity index can simultaneously reflect the calculation results and the safe operating boundary of the power grid.

[0160] In an optional implementation, step S403 includes the following steps S501-S505.

[0161] Step S501: Determine the distributed photovoltaic predicted output of each target grid node in each prediction period based on the photovoltaic output prediction sequence.

[0162] Here, since the photovoltaic output prediction sequence is the photovoltaic output sequence per unit installed capacity of the candidate region, when determining the distributed photovoltaic predicted output of the target grid node, the photovoltaic output per unit installed capacity of the candidate region is multiplied by the installed capacity to be evaluated of the candidate region to obtain the distributed photovoltaic predicted output of the candidate region in each prediction period. If the same target grid node corresponds to multiple candidate regions, the distributed photovoltaic predicted output of the multiple candidate regions in the same prediction period is summed to obtain the distributed photovoltaic predicted output of the target grid node in that prediction period.

[0163] In one implementation, the predicted distributed photovoltaic output of the i-th target grid node in the t-th prediction period can be determined by the following formula:

[0164] in, This represents the predicted distributed photovoltaic output of the i-th target grid node in the t-th prediction period. This represents the set of candidate regions for accessing the i-th target power grid node. This represents the installed capacity to be evaluated or the suggested installed capacity for the r-th candidate region. This represents the photovoltaic output per unit installed capacity of the r-th candidate region during the t-th prediction period.

[0165] Step S502: Determine the local forecast load of each target power grid node in each forecast period based on the node load data.

[0166] Here, local forecast load represents the load level at which the target grid node can locally absorb electrical energy during the forecast period. The node load data is organized according to the same time granularity as the photovoltaic output forecast sequence to obtain the local forecast load of the target grid node for each forecast period. If the node load data is historical load data, the planning system can generate the local forecast load based on historical load for the same period, load growth rate, typical daily load curves, and industry load change information. If the node load data already includes forecast load data, the local forecast load corresponding to the forecast time scale is directly extracted.

[0167] Step S503: Calculate the difference between the predicted output of distributed photovoltaic power and the predicted local load, and determine the reverse power flow prediction value for each target grid node in each prediction period if the difference is greater than zero.

[0168] Here, the reverse power flow prediction value is used to reflect the remaining power after the photovoltaic (PV) predicted output exceeds the local predicted load. If the distributed PV predicted output is less than or equal to the local predicted load, there is no reverse power flow caused by the distributed PV output at the target grid node, and the reverse power flow prediction value is 0. If the distributed PV predicted output is greater than the local predicted load, the difference is the power that needs to be fed back to the upstream grid or absorbed through energy storage, demand response, or other means.

[0169] In one implementation, the reverse power flow prediction value can be determined by the following formula:

[0170] in, This represents the reverse power flow prediction value for the i-th target grid node in the t-th prediction period. This represents the predicted distributed photovoltaic output of the i-th target grid node in the t-th prediction period. This represents the local forecast load of the i-th target grid node in the t-th forecast period.

[0171] In one implementation, the mitigation effect of local regulation resources on reverse power flow risk can be further considered. Local regulation resources include local energy storage charging capacity, adjustable load capacity, demand response capacity, and local controllable power supply regulation capacity. After considering local regulation resources, the reverse power flow forecast can be determined by the following formula:

[0172] in, This represents the local regulation resource capacity of the i-th target grid node in the t-th forecast period. This method can reflect the supporting role of energy storage charging, load regulation, and demand response in the local consumption of distributed photovoltaic power, and avoid overestimating the risk of reverse power flow.

[0173] Step S504: Determine the reverse flow risk coefficient based on the maximum value in the reverse flow prediction and the transformer capacity corresponding to the target grid node.

[0174] Here, the reverse flow risk coefficient measures the ratio of the maximum reverse power flow to the transformer capacity at the target grid node within the predicted timescale. The higher the reverse flow risk coefficient, the greater the pressure on the target grid node to send power back to the upstream grid, and the lower the safety margin for continued integration of distributed photovoltaic power.

[0175] The risk factor for backflow can be determined by the following formula:

[0176] in, This represents the reverse flow risk coefficient of the i-th target power grid node. This represents the reverse power flow prediction value for the i-th target grid node in the t-th prediction period, where T represents the total number of prediction periods within the prediction timescale. This represents the transformer capacity corresponding to the i-th target power grid node.

[0177] The presence of a high reverse power flow risk at a target grid node can also be determined based on a preset reverse power flow risk threshold. For example, when the reverse power flow risk coefficient is greater than the preset reverse power flow risk threshold, the target grid node is marked as a node with a high reverse power flow risk; when the reverse power flow risk coefficient is less than or equal to the preset reverse power flow risk threshold, the target grid node is marked as a node with a controllable reverse power flow risk. The preset reverse power flow risk threshold can be determined based on access specifications, distribution network operation requirements, or planning experience.

[0178] Step S505: Determine the node reverse power flow risk parameters based on the reverse flow risk coefficient.

[0179] Here, the node reverse flow risk parameter can be obtained directly from the reverse flow risk coefficient, or through normalization or piecewise mapping. If the planning model requires a higher risk value indicating higher risk, the reverse flow risk coefficient is normalized and used as the node reverse flow risk parameter. If the planning model requires a higher safety value indicating greater suitability for absorption, the reverse flow safety parameter is obtained by subtracting the normalized reverse flow risk coefficient from 1, and this reverse flow safety parameter is used in the absorption assessment value.

[0180] In one specific implementation, the node reverse power flow risk parameter can be determined by the following formula:

[0181] in, This represents the node reverse power flow risk parameter for the i-th target power grid node. This represents the reverse flow risk coefficient of the i-th target power grid node. This represents the maximum reverse flow risk coefficient corresponding to all target power grid nodes. This represents the minimum reverse flow risk coefficient corresponding to all target power grid nodes. If equal The node reverse power flow risk parameters corresponding to each target power grid node are set to the same preset value.

[0182] In optional implementations, the node price prediction characteristics include system supply and demand tension characteristics, new energy output characteristics, node reverse power flow risk characteristics, transmission congestion characteristics, and price lag characteristics.

[0183] Here, nodal price forecasting features are used to describe the main influencing factors that cause price increases, decreases, or price differentiation at target grid nodes within the forecast timescale. Before forecasting the nodal spot price series, the system supply and demand status, renewable energy output level, local absorption pressure, transmission channel constraints, and historical price change patterns corresponding to the target grid nodes are converted into feature data that can be input into the forecasting model or market clearing model. Based on this, the nodal spot price series is no longer simply extrapolated from historical nodal spot price data, but simultaneously reflects the impact of power system supply and demand, concentrated renewable energy output, local grid constraints, and cyclical price changes on the target grid node prices.

[0184] In step S104, based on the system supply and demand data, new energy power output forecast data, local load forecast data, local regulation resource capacity data, critical line power flow forecast data, line capacity data, and historical node spot price data in the basic data, as well as the photovoltaic power output forecast sequence, the node price forecast characteristics corresponding to each target grid node are determined, including the following steps S601-S605.

[0185] Step S601: Based on the system supply and demand data, determine the characteristics of the system supply and demand tension.

[0186] Here, system supply and demand data are used to describe the balance between the overall power system supply capacity and electricity demand within the forecast timescale. System supply and demand data includes forecasted system load, available generating capacity, unit maintenance capacity, external power import plans, energy storage charging and discharging plans, demand response capacity, and reserve capacity. First, the system supply and demand data are aligned according to the time granularity of the forecast timescale. Then, the system load level and available supply level for each forecast period are calculated. Finally, the relationship between the system load level and available supply level forms the system supply and demand tension characteristics.

[0187] In one implementation, the system's supply and demand tension characteristics can be determined by the following formula:

[0188] in, This represents the system's supply and demand tension characteristics during the t-th forecast period. This represents the system forecast load for the t-th forecast period. This represents the available power generation capacity for the t-th forecast period. This represents the incoming power plan for the t-th forecast period. This represents the demand response capacity for the t-th forecast period. The larger the value, the tighter the supply and demand in the system, and the higher the probability of the spot price at the node rising. The smaller the value, the more abundant the system supply, and the higher the probability of low prices during periods of high demand for new energy.

[0189] In one implementation, the system supply and demand tension characteristics can also be determined based on the remaining load. The remaining load can be determined by the following formula:

[0190] in, This represents the remaining load for the t-th forecast period. This represents the system forecast load for the t-th forecast period. This represents the predicted wind power output of the system during the t-th prediction period. This represents the system's predicted photovoltaic output for the t-th forecast period. Based on the size, rate of change, and historical percentile level of the remaining load, the system's supply and demand tension characteristics are formed. When the remaining load is high, the demand for conventional power regulation is high, and the nodal spot price tends to rise; when the remaining load is low, the pressure to absorb new energy sources is greater, and the nodal spot price tends to fall.

[0191] Step S602: Determine the characteristics of new energy output based on the predicted output data of new energy sources and the predicted output sequence of photovoltaic power.

[0192] Here, the renewable energy output forecast data is used to describe the output level of renewable energy resources such as wind power, centralized photovoltaic (PV), and distributed PV within the target area or related power grid within the forecast time scale. The PV output forecast sequence is used to describe the PV output per unit installed capacity in the candidate area for each forecast period. The total renewable energy output within the forecast period is calculated based on the renewable energy output forecast data, and the impact of renewable energy output on market prices is determined by combining it with the load forecast data for the same period.

[0193] In one implementation, the characteristics of new energy output can be represented by photovoltaic penetration rate, wind and solar power output ratio, or new energy power generation identification. Photovoltaic penetration rate can be determined by the following formula:

[0194] in, This represents the photovoltaic penetration rate of the i-th target grid node in the t-th prediction period. This represents the predicted distributed photovoltaic output of the i-th target grid node in the t-th prediction period. This represents the local forecast load of the i-th target grid node in the t-th forecast period. The larger the value, the higher the proportion of photovoltaic power in the load during the same period, and the higher the risk of low or negative electricity prices at midday.

[0195] In one implementation scheme, the proportion of renewable energy output is determined based on the predicted renewable energy output data and the predicted system load. The proportion of renewable energy output can be determined using the following formula:

[0196] in, This represents the proportion of new energy power output in the t-th forecast period. This represents the predicted wind power output of the system during the t-th prediction period. This represents the system's predicted photovoltaic output for the t-th prediction period. This represents the system's predicted load for the t-th forecast period. The proportion of renewable energy output, photovoltaic penetration rate, and peak photovoltaic generation periods can be combined as characteristics of renewable energy output. Based on this, renewable energy output characteristics can reflect the downward pressure on nodal spot prices caused by a high proportion of renewable energy integration.

[0197] Step S603: Based on the photovoltaic output forecast sequence, local load forecast data, and local regulation resource capacity data, determine the node reverse power flow risk characteristics.

[0198] Here, the node reverse power flow risk characteristic is used to characterize the reverse power flow pressure after the distributed photovoltaic output near the target grid node exceeds the local load and the absorption capacity of local regulation resources.

[0199] In one implementation, the photovoltaic output prediction sequences of candidate regions connected to the same target grid node are first summarized based on the correspondence between candidate regions and target grid nodes to obtain the distributed photovoltaic predicted output of the target grid node in each prediction period. Then, the local predicted load of the target grid node in each prediction period is extracted from the local load prediction data, and the energy storage charging capacity, adjustable load capacity, demand response capacity, and local controllable power supply regulation capacity are extracted from the local regulation resource capacity data. The node reverse power flow risk characteristics can be determined by the following formula:

[0200] in, This represents the node reverse power flow risk characteristics of the i-th target power grid node in the t-th prediction period. This represents the predicted distributed photovoltaic output of the i-th target grid node in the t-th prediction period. This represents the local forecast load of the i-th target grid node in the t-th forecast period. This represents the local regulation resource capacity of the i-th target power grid node in the t-th prediction period. The larger the value, the greater the back-transmission pressure on the target grid node during the corresponding forecast period, and the higher the risk of the target grid node experiencing low prices, curtailment, or limited access.

[0201] Step S604: Based on the predicted power flow data and line capacity data of the critical line, determine the transmission congestion characteristics.

[0202] Here, critical path forecast power flow data is used to describe the power flow status of critical paths between the target grid node and its upstream grid, adjacent nodes, or load centers within the forecast timescale. Line capacity data is used to describe the maximum power that the critical path is allowed to carry. Transmission congestion characteristics are used to characterize the likelihood of a divergence between the node spot price and the system average price when the critical path power flow approaches its line capacity.

[0203] In one implementation scheme, a set of critical lines corresponding to the target grid node is determined based on the distribution network topology and the location of the target grid node. The set of critical lines includes grid connection channels to the target grid node, transmission lines from upstream substations, interconnection lines between adjacent feeders, and transmission channels from areas with concentrated renewable energy access. The predicted power flow for each critical line in each prediction period is extracted from the predicted power flow data, and the line capacity for each critical line is extracted from the line capacity data. Transmission congestion characteristics can be determined using the following formula:

[0204] in, This represents the transmission congestion characteristics of the i-th target grid node in the t-th prediction period. This represents the set of critical paths corresponding to the i-th target power grid node. This represents the predicted power flow of the l-th critical path during the t-th prediction period. This represents the line capacity of the l-th critical path. The closer it is to 1, the closer the critical path is to its capacity limit, and the higher the likelihood of a divergence between the spot price of the target grid node and the system average price. When the value exceeds the preset congestion threshold, the planning system marks the i-th target power grid node as a congestion risk period in the t-th prediction period.

[0205] Step S605: Determine price lag characteristics based on historical spot price data.

[0206] Here, historical spot price data is used to describe the changes in spot prices at target grid nodes over historical periods. Price lag characteristics are used to characterize the historical inertia, intraday periodicity, weekly periodicity, and persistence of extreme prices in spot prices. Price lag characteristics can help spot price forecasting models identify price patterns at the same time point, as well as the persistent risks of periods of low prices and negative electricity prices.

[0207] In one implementation, historical spot price data for each node is cleaned and aligned according to the time granularity of the prediction time scale to obtain the historical price sequence of the target grid node. From the historical price sequence, the following parameters are extracted: price of the previous prediction period, price of the same period the previous day, price of the same period the previous week, historical average price, historical highest price, historical lowest price, historical price volatility, duration of low-price periods, and duration of negative electricity price periods. At least one of these parameters is used as a price lag feature. The price lag feature can be expressed as:

[0208] in, This represents the price lag characteristic of the i-th target power grid node in the t-th prediction period. This represents the historical spot price of the i-th target grid node in the previous forecast period. This represents the historical spot price of the i-th target grid node at the same time period on the previous day. This represents the historical spot price of the i-th target grid node during the same period of the previous week. This represents the average price of the i-th target grid node during the same historical period. This represents the highest price of the i-th target grid node during the same historical period. Let represent the lowest price of the i-th target grid node during the same historical period. This represents the price volatility of the i-th target grid node during the same historical period. If the time granularity is not 1 hour, the planning system adjusts the lag order corresponding to the same period of the previous day and the same period of the previous week according to the actual time granularity.

[0209] Furthermore, the system's supply and demand tension characteristics, new energy output characteristics, node reverse power flow risk characteristics, transmission congestion characteristics, and price lag characteristics are concatenated to obtain the node price prediction characteristics for each target grid node in each prediction period. The node price prediction characteristics can be expressed as:

[0210] in, This represents the node price prediction characteristics of the i-th target power grid node in the t-th prediction period. This represents the system's supply and demand tension characteristics during the t-th forecast period. This represents the proportion of new energy power output in the t-th forecast period. This represents the photovoltaic penetration rate of the i-th target grid node in the t-th prediction period. This represents the node reverse power flow risk characteristics of the i-th target power grid node in the t-th prediction period. This represents the transmission congestion characteristics of the i-th target grid node in the t-th prediction period. This represents the price lag characteristic of the i-th target power grid node in the t-th prediction period.

[0211] In one implementation, market rule features and cost-based pricing features can be introduced when constructing nodal price forecasting characteristics. Market rule features include day-ahead market rules, real-time market rules, deviation assessment rules, capacity compensation rules, and green certificate revenue rules. Cost-based pricing features include coal prices, gas prices, unit marginal costs, unit start-up and shutdown costs, energy storage charging and discharging prices, and power generation resource pricing strategies. Inputting market rule features, cost-based pricing features, and the aforementioned features into the nodal spot price forecasting model further enhances the ability of the nodal spot price series to characterize the dynamic market clearing mechanism.

[0212] In an optional implementation, step S104 predicts the spot price sequence of each target power grid node based on the node price prediction characteristics, including the following steps S701-S703.

[0213] Step S701: Input the node price prediction features corresponding to each target power grid node into the pre-trained node spot price prediction model.

[0214] Here, the nodal spot price forecasting model is used to characterize the mapping relationship between nodal price forecasting features and nodal spot prices. The nodal spot price can be the nodal marginal electricity price corresponding to the target grid node, or it can be the zonal price, access point price, or regional spot price after nodalization correction that matches the target grid node.

[0215] The node price forecast features consist of the aforementioned system supply and demand tension features, new energy output features, node reverse power flow risk features, transmission congestion features, and price lag features. Before inputting the node spot price forecast model, the node price forecast features from different sources undergo time alignment and dimensional unification processing. Time alignment processing refers to organizing the node price forecast features into time-period feature data according to the time granularity of the forecast time scale. Dimensional unification processing refers to normalizing, standardizing, or quantile transforming features with different numerical ranges, so that different features can be stably identified by the node spot price forecast model.

[0216] In one implementation, the node price prediction characteristic of the i-th target grid node in the t-th prediction period can be expressed as: .

[0217] Node spot price forecasting models can employ gradient boosting tree models, random forest models, support vector regression models, long short-term memory network models, temporal convolutional network models, or Transformer time series forecasting models. Node spot price forecasting models can also use hybrid models combining physical constraints and data-driven approaches. This involves first simulating market clearing based on system supply and demand data, predicted renewable energy output data, predicted power flow data for critical lines, line capacity data, and market rule data to obtain simulated node prices, and then using historical node spot price data to correct for deviations in the simulated node prices.

[0218] In one training method, historical sample data is acquired. This historical sample data includes historical system supply and demand data, historical renewable energy output data, historical local load data, historical local regulation resource capacity data, historical critical line power flow data, historical line capacity data, and historical nodal spot price data. Historical nodal price prediction features are constructed based on this historical sample data, and the historical nodal spot prices for the corresponding periods are used as training labels. The initial prediction model is trained using these historical nodal price prediction features as model input and the historical nodal spot prices as model output labels, resulting in the nodal spot price prediction model.

[0219] In one implementation, the training objective of the node spot price prediction model can be expressed as:

[0220] in, This represents the model parameters of the spot price prediction model, where n represents the number of training samples. This represents the predicted spot price of the q-th training sample. This represents the historical spot price corresponding to the q-th training sample. During training, historical sample data is divided into training, validation, and test sets. The spot price prediction model is saved when the validation set error meets a preset error condition. The preset error condition can be determined based on root mean square error, mean absolute error, or mean absolute percentage error.

[0221] Step S702: Using the nodal spot price prediction model, output the nodal spot price for each target power grid node in each prediction period within the prediction time scale.

[0222] Here, the nodal price prediction characteristics of the i-th target grid node in the t-th prediction period are input into the nodal spot price prediction model, and the nodal spot price of the i-th target grid node in the t-th prediction period is output. The nodal spot price prediction process can be represented as:

[0223] in, This represents the spot price of the i-th target grid node in the t-th prediction period. This represents a pre-trained node spot price prediction model. This represents the node price prediction characteristic of the i-th target power grid node in the t-th prediction period.

[0224] In one implementation, the nodal spot price forecasting model outputs a price probability distribution or price range for each forecast period. Based on the price probability distribution, the expected value of the nodal spot price, the probability of low prices, the probability of negative electricity prices, and the price volatility risk are determined. For forecast periods with low price or negative electricity price risks, the positive contribution of the corresponding forecast period to the revenue fit of the candidate region is reduced in the subsequent calculation of the revenue fit index. In this way, the nodal spot price forecasting results not only reflect the expected price level but also the downside risk of prices.

[0225] In one implementation, the simulated spot price of each target grid node in each forecast period is first obtained through market clearing simulation. Then, the simulated spot price is corrected using a deviation correction model. The inputs to the deviation correction model include the simulated spot price, price lag characteristics, transmission congestion characteristics, and node reverse power flow risk characteristics. The output of the deviation correction model is the corrected spot price. This approach can take into account both the market clearing mechanism and historical price patterns.

[0226] Step S703: Arrange the spot prices of each prediction period according to the time granularity of the prediction time scale to obtain the spot price sequence of each target power grid node.

[0227] Here, the node spot price series is used to characterize the price change process of the target grid node within the forecast time scale and serves as the input for subsequent revenue matching index calculation.

[0228] In one implementation, the spot price sequence of the i-th target grid node can be represented as: .

[0229] After generating the node spot price series, its rationality is verified. If there are missing forecast periods in the node spot price series, the missing forecast periods are filled by interpolation of adjacent periods, filling with prices of similar historical periods, or correcting with prices of adjacent nodes. If there are outliers in the node spot price series that exceed the preset price upper limit or fall below the preset price lower limit, the outlier is judged whether to be retained based on the system's supply and demand tension characteristics, new energy output characteristics, and historical price percentile levels. For reasonable extreme prices caused by extreme supply and demand conditions or congestion, the corresponding node spot price is retained; for abnormal prices caused by data anomalies, the corresponding node spot price is corrected.

[0230] In an optional implementation, refer to Figure 5 Step S105 includes the following steps S801-S805.

[0231] Step S801: Determine the predicted photovoltaic output of each candidate region for each prediction period based on the photovoltaic output prediction sequence.

[0232] Here, the photovoltaic (PV) output prediction sequence is arranged according to the prediction time scale and time granularity. The PV output prediction sequence for the i-th candidate region is read to obtain the PV output prediction for the i-th candidate region in the t-th prediction period. The PV output prediction can be based on the PV output per unit installed capacity, or it can be obtained by multiplying the PV output per unit installed capacity by the installed capacity to be evaluated. Since the yield fit index is mainly used to evaluate the matching relationship between the PV output period and the spot price at the node, using the PV output per unit installed capacity does not affect the yield fit comparison between different candidate regions.

[0233] Step S802: Determine the node spot price of the target power grid node for each candidate region in each prediction period based on the node spot price sequence.

[0234] Here, the target grid node corresponding to the i-th candidate region is first determined based on the correspondence between candidate regions and target grid nodes. Then, the node spot price with the same time granularity as the photovoltaic output prediction sequence is read from the node spot price sequence corresponding to the target grid node. Based on this, each prediction period corresponds to a photovoltaic predicted output and a node spot price, which can be used to evaluate the correspondence between photovoltaic power generation and high-price periods, low-price periods, or negative-price periods from a time dimension.

[0235] Step S803: Based on the photovoltaic power output forecast for each forecast period, the spot price of the corresponding forecast period is weighted to obtain the photovoltaic weighted average spot price.

[0236] Here, the photovoltaic weighted average spot price is used to characterize the price level at which actual photovoltaic output mainly falls. Compared to directly using the arithmetic mean of spot prices at nodes within the forecast time scale, the photovoltaic weighted average spot price can increase the weight of high-output photovoltaic periods and reduce the interference of nighttime periods with no or low photovoltaic output on revenue evaluation.

[0237] The weighted average spot price of photovoltaic power can be determined by the following formula:

[0238] in, This represents the weighted average spot price of photovoltaic power corresponding to the i-th candidate region. This represents the predicted photovoltaic output of the i-th candidate region during the t-th prediction period. This represents the spot price of the target power grid node corresponding to the i-th candidate region during the t-th prediction period. This represents the target power grid node number corresponding to the i-th candidate region. This represents the length of time corresponding to a single forecast period, and T represents the total number of forecast periods within the forecast timescale.

[0239] When the sum of all predicted photovoltaic power outputs within the predicted timescale is zero, the photovoltaic weighted average spot price is set to a preset invalid value, or the corresponding candidate region is removed from the revenue matching evaluation. If there are individual missing periods in the nodal spot price series, they are supplemented by prices of adjacent periods, historical prices of similar periods, or prices of adjacent target grid nodes before the photovoltaic weighted average spot price is calculated.

[0240] Step S804: Determine the probability of low-price periods and / or negative-price periods based on the node spot price sequence.

[0241] Here, the probability of a low-price period is used to characterize the likelihood of a target grid node operating at a low price within the predicted timescale. The probability of a negative electricity price period is used to characterize the likelihood of a target grid node operating at a negative electricity price within the predicted timescale. The preset low-price threshold can be determined based on the historical spot price percentile of the target area, the minimum acceptable grid connection price for the project, the long-term contract electricity price, or the price set by the planning department.

[0242] In one implementation, the probabilities of low-price periods and negative-price periods are calculated based on the number of predicted periods. The probabilities of low-price periods and negative-price periods can be determined by the following formula:

[0243]

[0244] in, This represents the probability of a low-price period corresponding to the i-th candidate region. This represents the probability of a negative electricity price period corresponding to the i-th candidate region. This represents the spot price of the target power grid node corresponding to the i-th candidate region during the t-th prediction period. This indicates a preset low-price threshold. This indicates an indicator function. The condition within the parentheses is true when the condition is true. Take 1; when the condition within the parentheses is not true, Take 0.

[0245] In one implementation, the probability of low-price power output and the probability of negative-price power output are calculated according to the photovoltaic power output weight. The probability of low-price power output and the probability of negative-price power output can be determined by the following formula:

[0246]

[0247] in, This represents the probability of low-price output corresponding to the i-th candidate region. This represents the probability of negative electricity price output corresponding to the i-th candidate region. The probability of low-price output and the probability of negative electricity price output reflect the proportion of photovoltaic (PV) output predictions falling into low-price or negative-price periods. Compared to the method of counting by the number of prediction periods, the method of counting by PV output weight is closer to the actual return and risk of distributed PV projects.

[0248] Step S805: Based on the photovoltaic weighted average spot price and the probability of low price periods and / or negative electricity price periods, determine the revenue matching index corresponding to each candidate region.

[0249] Here, the revenue fit index is used to evaluate the degree of matching between the photovoltaic output of the candidate region and the changes in spot prices at the target grid nodes. The higher the photovoltaic weighted average spot price, the better the revenue fit of the candidate region; the higher the probability of low-price periods and negative electricity price periods, the worse the revenue fit of the candidate region.

[0250] In one implementation, the weighted average spot price of photovoltaic power is first normalized to obtain a price-return evaluation value; then, a price risk deduction value is determined based on the probability of low-price periods and negative-price periods; finally, the price-return evaluation value and the price risk deduction value are weighted to obtain a return matching index. The return matching index can be determined by the following formula:

[0251] In the formula, This represents the revenue fit metric corresponding to the i-th candidate region. This represents the normalized weighted average spot price of photovoltaic power corresponding to the i-th candidate region. This represents the probability of a low-price period corresponding to the i-th candidate region. This represents the probability of a negative electricity price period corresponding to the i-th candidate region. , and These represent the weighting coefficients corresponding to the photovoltaic weighted average spot price, the probability of low-price periods, and the probability of negative electricity price periods, respectively. .

[0252] In another specific implementation, if the probability of low-price power output and the probability of negative-price power output are adopted, the revenue matching index can be determined by the following formula:

[0253] The yield fit metric can also be adjusted by incorporating factors such as the volatility of spot prices at nodes, the duration of low prices, the duration of negative electricity prices, and the price percentile during periods of high photovoltaic output. For example, if there are consecutive periods of low prices at the target grid nodes corresponding to a candidate region, the yield fit metric for that candidate region is lowered; conversely, if the price at the target grid nodes corresponding to high photovoltaic output periods in a candidate region is consistently higher than the regional average price, the yield fit metric for that candidate region is raised. Based on this, it is possible to avoid concentrating photovoltaic capacity in areas with high risks of low prices or negative electricity prices during peak photovoltaic power generation periods.

[0254] In an optional implementation, refer to Figure 6 Step S106 includes the following steps S901-S906.

[0255] Step S901: Use the suggested installed capacity of each candidate region as the decision variable.

[0256] Here, the recommended installed capacity is used to represent the distributed photovoltaic installed capacity recommended by the planning model for construction in the candidate areas. The recommended installed capacity for the i-th candidate area can be denoted as... The planning model, during the solution process, treats each... Adjustments will be made to improve the overall planning efficiency of the target area while meeting preset constraints. It is recommended that the range of installed capacity be jointly limited by the actual available construction area of ​​the candidate area, the installed capacity density per unit area, the openable capacity of the target grid nodes, and grid security constraints.

[0257] Step S902: Based on the preset power generation weight coefficient, absorption weight coefficient, and revenue weight coefficient, the power generation potential index, absorption capacity index, and revenue matching index corresponding to each candidate region are weighted to obtain the planning matching score corresponding to each candidate region.

[0258] Here, the power generation weight coefficient is used to characterize the degree of importance the planning model places on photovoltaic power generation capacity, the grid absorption weight coefficient is used to characterize the degree of importance the planning model places on the grid's safe absorption capacity, and the revenue weight coefficient is used to characterize the degree of importance the planning model places on the adaptability of spot price revenue. These three weight coefficients can be determined based on the planning focus of the target area, project investment preferences, grid security requirements, or energy utilization goals.

[0259] The planning fit score can be determined by the following formula:

[0260] in, This represents the planning fit score corresponding to the i-th candidate region. This represents the power generation potential index corresponding to the i-th candidate region. This represents the absorption capacity index corresponding to the i-th candidate region. This represents the revenue fit metric corresponding to the i-th candidate region. Represents the power generation weighting coefficient. This represents the absorption weighting coefficient. This represents the return weighting coefficient. .

[0261] In one implementation, a balanced weight combination can be set, for example... =0.4、 =0.3、 =0.3, used to balance power generation capacity, absorption capacity, and revenue matching. Economic benefit-based weight combinations can also be set, such as increasing... This allows distributed photovoltaic (PV) planning schemes to favor candidate regions with better nodal spot price profitability. Grid-friendly weighting combinations can also be implemented, such as increasing... This makes distributed photovoltaic planning schemes more inclined towards candidate areas with strong absorption capacity and low reverse power flow risk.

[0262] Step S903: Based on the recommended installed capacity and planning adaptation score of each candidate area, determine the single-point planning benefit corresponding to each candidate area.

[0263] Here, the single-point planning benefit is used to characterize the contribution of the proposed installed capacity in the candidate area to the overall planning objective of the target area. The single-point planning benefit can be determined by the following formula:

[0264] in, This represents the single-point planning benefit corresponding to the i-th candidate region. This represents the suggested installed capacity corresponding to the i-th candidate region. This represents the planning adaptation score corresponding to the i-th candidate region. Therefore, both the adaptation degree of the candidate region itself and the actual installed capacity allocated to the candidate region are considered.

[0265] Step S904: Construct a planning model with the objective of maximizing the sum of the planning benefits of all candidate areas within the target area.

[0266] Here, the objective function of the planning model can be expressed by the following formula:

[0267] Where F represents the overall regional benefits of the target area, and N represents the number of candidate areas within the target area.

[0268] Pre-defined constraints are used to ensure that the recommended installed capacity output by the planning model meets spatial construction conditions and grid safety operation requirements. These constraints include total regional installed capacity constraints, maximum installable capacity constraints for candidate regions, grid safety constraints, node reverse power flow constraints, and openable capacity constraints. The total regional installed capacity constraint can be expressed as: .in, This indicates the total planned capacity of the target area.

[0269] The maximum installable capacity constraint for the candidate region can be expressed as: .in, This represents the maximum installable capacity of the i-th candidate region.

[0270] Power grid safety constraints are used to limit node voltages, line current carrying capacity, and transformer load rates to meet preset safe operating conditions. Node voltage constraints can be expressed as: .in, This represents the voltage of the k-th node during the t-th prediction period. This indicates the preset lower limit of the node voltage. This indicates the preset upper limit of the node voltage.

[0271] The line current carrying capacity constraint can be expressed as: .in, This represents the predicted power flow of the l-th line in the t-th prediction period. This represents the line capacity of the l-th line.

[0272] The node reverse power flow constraint can be expressed as: .in, This represents the reverse flow risk coefficient of the i-th target power grid node. This indicates a preset reverse current risk threshold.

[0273] Openable capacity constraints can be expressed as: .in, This represents the set of candidate regions for connecting to the nth target power grid node. This represents the available capacity of the nth target power grid node.

[0274] Step S905: Solve the planning model under the preset constraints to obtain the suggested installed capacity for each candidate region.

[0275] Here, particle swarm optimization (PSO), genetic algorithms, mixed-integer programming, heuristic search algorithms, or NSGA-II (Non-dominated Sorting Genetic Algorithm II) can be used to solve the planning model. When using PSO, a set of candidate regions' suggested installed capacity is treated as a particle position. The overall regional benefit for each particle is calculated, and particles that do not meet the constraints are penalized or corrected. After multiple iterations, the combination of suggested installed capacity that has a high overall regional benefit and meets the preset constraints is output. When using mixed-integer programming, the suggested installed capacity is set as a continuous variable or a discrete capacity level variable, and the optimal capacity configuration that meets the constraints is obtained through an optimization solver.

[0276] In one implementation, an initial capacity plan is first generated based on the maximum installable capacity of the candidate region. Then, a power flow verification is performed on the initial capacity plan to determine whether node voltage, line current carrying capacity, main transformer load rate, and node reverse power flow meet preset constraints. If the initial capacity plan does not meet the preset constraints, the recommended installed capacity of the candidate region corresponding to the conflicting node is reduced, or some installed capacity is transferred to a candidate region with higher absorption capacity and benefit matching indicators. Capacity adjustment and constraint verification are repeated until the planning model converges or the preset number of iterations is reached. This solution process avoids the problem that the planning result may have high benefits on paper but be unacceptable in grid operation.

[0277] Step S906: Generate a distributed photovoltaic planning scheme based on the suggested installed capacity corresponding to each candidate region.

[0278] Here, the distributed photovoltaic (PV) planning scheme includes candidate area identifiers, candidate area geographical locations, target grid nodes, suggested installed capacity, power generation potential indicators, grid integration capacity indicators, revenue matching indicators, expected power generation, PV weighted average spot price, probability of low-price periods, probability of negative-price periods, and grid security verification results. The distributed PV planning scheme can also include scenario schemes corresponding to different weight combinations, such as balanced schemes, economic benefit schemes, and grid-friendly schemes.

[0279] In one implementation, candidate regions are sorted according to planning suitability scores or recommended installed capacity, generating a planning results table and a spatial layout map. The planning results table displays the recommended installed capacity, target grid nodes, and key evaluation indicators for each candidate region. The spatial layout map shows the recommended distribution locations of distributed photovoltaic capacity within the target region. Low-price risk warnings and grid integration risk warnings can also be output. When the probability of low-price periods, negative-price periods, or reverse power flow risk at nodes corresponding to a candidate region exceeds a preset threshold, the corresponding candidate region is marked in the distributed photovoltaic planning scheme to facilitate review or adjustment by planners.

[0280] In one specific embodiment, the target area is an industrial park, urban commercial building complex, and suburban vacant land in a coastal city, City A. City A plans to add 200MW of distributed photovoltaic (PV) capacity within the target planning period. The planning system needs to determine the recommended installed capacity for different candidate areas so that the new distributed PV capacity simultaneously meets the requirements of power generation capacity, grid connection capacity, and nodal spot price profitability.

[0281] The planning system acquires basic data for City A. This basic data includes solar irradiance data, temperature data, photovoltaic module parameters, nodal load data for each substation and feeder node, power flow data, main transformer load rate data, bus voltage data, system supply and demand data, predicted output data for new energy sources, predicted local load data, local regulation resource capacity data, predicted power flow data for critical lines, line capacity data, historical spot price data, available rooftop and open space resources, planning constraints, and geospatial data. Based on the planning constraints, the planning system eliminates unsuitable areas for distributed photovoltaic construction from the geographic area of ​​City A and grids the remaining effective working areas. The planning system calculates the actual available construction area for each basic grid and filters out candidate areas whose actual available construction area exceeds a preset area threshold. After filtering, the planning system obtains candidate areas A1, A2, A3, and A4. Based on the geographical location of the candidate areas, the location of the power grid access point, and the distribution network topology, the planning system determines the target power grid node N1 corresponding to candidate area A1, the target power grid node N2 corresponding to candidate area A2, the target power grid node N3 corresponding to candidate area A3, and the target power grid node N4 corresponding to candidate area A4.

[0282] The planning system uses typical future years as the prediction time scale and 1 hour as the time granularity to generate photovoltaic output prediction sequences for candidate regions A1 to A4. Based on the photovoltaic output prediction sequences for each candidate region, the system calculates the expected annual power generation per unit installed capacity and performs normalization to obtain a power generation potential index. Candidate region A1 is located in a suburban vacant area with high solar irradiance, and its power generation potential index is 0.92; candidate region A2 is located in an industrial park with moderate to high solar irradiance, and its power generation potential index is 0.78; candidate region A3 is located in an urban commercial building complex, affected by building shading, and its power generation potential index is 0.65; candidate region A4 is located in a port logistics area with moderate solar irradiance, and its power generation potential index is 0.70.

[0283] The planning system calculates the absorption capacity index based on the node load data, grid structure parameters, and photovoltaic output prediction sequences of the corresponding candidate areas for each target grid node. Target grid node N1 corresponds to an area with low local load, and reverse power flow is prone to occur during the midday peak photovoltaic output period, resulting in an absorption capacity index of 0.45. Target grid node N2 corresponds to an industrial park, with stable production load during the day and good margins in line capacity and main transformer capacity, resulting in an absorption capacity index of 0.86. Target grid node N3 corresponds to a commercial building complex in the urban area, with high commercial load during the day, but some line power flows are close to their upper limits, resulting in an absorption capacity index of 0.72. Target grid node N4 corresponds to a port logistics area, with high load at night and relatively moderate load at midday, resulting in an absorption capacity index of 0.61.

[0284] Based on system supply and demand data, new energy power output forecast data, local load forecast data, local regulation resource capacity data, critical line power flow forecast data, line capacity data, historical node spot price data, and photovoltaic power output forecast sequence, the planning system constructs node price forecast characteristics for target grid nodes N1 to N4, and predicts the node spot price sequence for each target grid node through a pre-trained node spot price forecast model. The forecast results show that during the midday peak photovoltaic generation period, target grid node N1 is affected by insufficient local load and limited transmission channels, with a low price period probability of 0.38 and a negative price period probability of 0.08. During the midday peak, target grid node N2 has stable industrial load and relatively high node prices, with a low price period probability of 0.10 and a negative price period probability of 0. The spot price of target grid node N3 remains at a moderately high level during weekday midday, with a low price period probability of 0.16 and a negative price period probability of 0.01. During the midday peak, target grid node N4 experiences large price fluctuations, with a low price period probability of 0.24 and a negative price period probability of 0.03.

[0285] The planning system performs time matching between the photovoltaic (PV) output forecast sequence of each candidate region and the spot price sequence of the corresponding target grid node, calculates the PV weighted average spot price, and determines the revenue fit index by combining the probability of low-price periods and negative-price periods. Although candidate region A1 has high power generation potential, its high PV output periods overlap significantly with the low-price periods of target grid node N1, resulting in a revenue fit index of 0.48. Candidate region A2's high PV output periods match the higher-price periods of target grid node N2 well, resulting in a revenue fit index of 0.88. Candidate region A3's high PV output periods largely overlap with the peak commercial load periods of target grid node N3, resulting in a revenue fit index of 0.76. Candidate region A4 has a revenue fit index of 0.58.

[0286] The planning system sets the power generation weight coefficient at 0.35, the power consumption weight coefficient at 0.35, and the revenue weight coefficient at 0.30 to construct the planning model. The system uses the suggested installed capacity for each candidate region as a decision variable and sets a total installed capacity constraint of 200MW for the region. The system also sets the maximum installable capacity based on the actual available construction area of ​​each candidate region: 100MW for candidate region A1, 80MW for candidate region A2, 60MW for candidate region A3, and 70MW for candidate region A4. The system also sets node voltage constraints, line current carrying capacity constraints, node reverse power flow constraints, and available capacity constraints.

[0287] After solving the planning model, the planning system yields a distributed photovoltaic (PV) planning scheme. The recommended distributed PV planning scheme allocates 35MW to candidate region A1, 75MW to candidate region A2, 55MW to candidate region A3, and 35MW to candidate region A4. This planning result does not concentrate the maximum capacity in candidate region A1, which has the highest power generation potential, but instead allocates a significant amount of capacity to candidate regions A2 and A3, which have higher grid integration capacity and revenue matching indicators. Power flow verification shows that the node voltage, line current carrying capacity, main transformer load factor, and node reverse power flow risk corresponding to the distributed PV planning scheme all meet the preset constraints.

[0288] In the above embodiments, candidate region A1 has high solar irradiance, but target grid node N1 suffers from insufficient local load, high reverse power flow risk, and high midday low price risk. Therefore, the planning model reduces the recommended installed capacity of candidate region A1. Although candidate regions A2 and A3 do not have the highest power generation potential, the target grid nodes corresponding to candidate regions A2 and A3 have good absorption capacity and high revenue adaptability. Therefore, the planning model increases the recommended installed capacity of candidate regions A2 and A3. Thus, the distributed photovoltaic planning scheme can achieve a balance between power generation capacity, grid absorption capacity, and spot price revenue adaptability, reducing the impact of low prices, negative electricity prices, and reverse power flow risk during peak photovoltaic generation periods on the planning results, and improving the grid-connectivity and economics of the distributed photovoltaic planning scheme.

[0289] Based on this, embodiments of this application provide a distributed photovoltaic planning system, referring to... Figure 7 The distributed photovoltaic planning system provided in this application includes: Candidate region determination module 1 is used to obtain basic data of the target region, and filter out at least one candidate region that can be laid out based on the basic data, and determine the target power grid node corresponding to each candidate region.

[0290] The power generation potential determination module 2 is used to generate a photovoltaic output prediction sequence for each candidate region based on the photovoltaic resource data in the basic data, and to determine the power generation potential index corresponding to each candidate region.

[0291] The absorption capacity determination module 3 is used to determine the absorption capacity index corresponding to each candidate region based on the grid node operation data, grid structure parameters and photovoltaic output prediction sequence in the basic data.

[0292] The node price prediction module 4 is used to determine the node price prediction characteristics corresponding to each target grid node based on the system supply and demand data, new energy power output prediction data, local load prediction data, local regulation resource capacity data, critical line power flow prediction data, line capacity data, historical node spot price data, and photovoltaic power output prediction sequence in the basic data, and predict the node spot price sequence of each target grid node based on the node price prediction characteristics.

[0293] The revenue matching determination module 5 is used to determine the revenue matching index corresponding to each candidate region based on the matching relationship between the spot price series of nodes and the photovoltaic output prediction series.

[0294] The planning scheme generation module 6 is used to construct a planning model with power generation potential indicators, absorption capacity indicators and income matching indicators as optimization objectives, and solve it under the condition of satisfying preset constraints to obtain a distributed photovoltaic planning scheme containing the recommended installed capacity of each candidate area.

[0295] The distributed photovoltaic (PV) planning system provided in this application establishes a correspondence between spatial areas within the target area that meet the layout conditions and target grid nodes through a candidate area determination module. The power generation potential determination module, grid absorption capacity determination module, node price prediction module, and revenue matching determination module quantitatively evaluate the candidate areas from aspects such as PV power generation capacity, grid safe absorption capacity, node spot price changes, and the degree of matching between PV output and price periods. The planning scheme generation module constructs and solves the planning model based on the above evaluation results. This allows the distributed PV planning system to comprehensively consider natural resource conditions, grid operation constraints, and spot market revenue characteristics during the planning stage, thereby reasonably determining the recommended installed capacity of each candidate area, and thus improving the constructability, grid connection capability, absorption reliability, and overall economic benefits of the distributed PV planning scheme.

[0296] In an optional implementation, the basic data includes buildable resource data, planning constraint data, and geospatial data. The candidate region determination module 1 is further used for: Restricted areas within the target region are identified based on planning constraint data.

[0297] The effective operating area is obtained by removing restrictive areas from the target area range corresponding to the geospatial data.

[0298] The effective working area is spatially discretized based on a preset size grid to obtain multiple basic grids.

[0299] The actual available buildable area within each basic grid is determined based on buildable resource data.

[0300] The basic grid with an actual usable construction area greater than the preset area threshold is identified as the candidate area.

[0301] Based on the geographical location of the candidate region, the location of the power grid access point, and the distribution network topology, the target power grid node corresponding to each candidate region is determined.

[0302] In an optional implementation, the photovoltaic resource data includes solar irradiance data, temperature data, and photovoltaic module parameters. The power generation potential determination module 2 is also used for: Extract solar irradiance data, temperature data, and photovoltaic module parameters for each candidate region from the photovoltaic resource data.

[0303] Based on solar irradiance data, temperature data, and photovoltaic module parameters, the photovoltaic output per unit installed capacity of each candidate region is calculated within the predicted time scale.

[0304] A photovoltaic output prediction sequence is generated by arranging the unit installed capacity photovoltaic output based on the time granularity of the prediction time scale.

[0305] The expected power generation of each candidate region within the prediction timescale is calculated based on the photovoltaic power output prediction sequence.

[0306] The expected power generation is normalized to obtain the power generation potential index for each candidate region.

[0307] In an optional implementation, the power grid node operation data includes node load data, line power flow data, main transformer load rate data, and bus voltage data. The absorption capacity determination module 3 is also used for: The local load characteristic parameters of each target power grid node are determined based on the node load data.

[0308] Based on power grid structure parameters, line power flow data, main transformer load rate data, and bus voltage data, the power grid strength parameters of each target power grid node are determined.

[0309] Based on the photovoltaic output prediction sequence and node load data, the node reverse power flow risk parameters for each target grid node are determined.

[0310] The local load characteristic parameters, grid strength parameters, and node reverse power flow risk parameters are weighted to obtain the absorption assessment value for each candidate region.

[0311] The absorption capacity assessment values ​​are normalized to obtain the absorption capacity index corresponding to each candidate area.

[0312] In an optional implementation, the absorption capacity determination module 3 is further configured to: The distributed photovoltaic (PV) predicted output of each target grid node in each prediction period is determined based on the PV output prediction sequence.

[0313] The local forecast load for each target grid node in each forecast period is determined based on node load data.

[0314] Calculate the difference between the predicted output of distributed photovoltaic power and the predicted local load, and determine the reverse power flow prediction value for each target grid node in each prediction period for the difference that is greater than zero.

[0315] The reverse flow risk coefficient is determined based on the maximum value in the reverse flow prediction and the transformer capacity corresponding to the target grid node.

[0316] The reverse flow risk parameters of nodes are determined based on the reverse flow risk coefficient.

[0317] In optional implementations, the node price prediction features include system supply and demand tension features, new energy output features, node reverse power flow risk features, transmission congestion features, and price lag features. The node price prediction module 4 is also used for: Based on system supply and demand data, the characteristics of system supply and demand tension are determined.

[0318] Based on the predicted power output data of new energy sources and the predicted power output sequence of photovoltaic power, the characteristics of new energy power output are determined.

[0319] Based on photovoltaic power output forecast sequences, local load forecast data, and local regulation resource capacity data, the characteristics of node reverse power flow risk are determined.

[0320] Based on the predicted power flow data and line capacity data of the critical path, the characteristics of transmission congestion are determined.

[0321] Based on historical spot price data, price lag characteristics are determined.

[0322] In an optional implementation, the node price prediction module 4 is further configured to: The node price prediction features corresponding to each target power grid node are input into the pre-trained node spot price prediction model.

[0323] The nodal spot price prediction model outputs the nodal spot price for each target power grid node in each prediction period within the prediction time scale.

[0324] Arrange the spot prices of each forecast period according to the time granularity of the forecast time scale to obtain the spot price sequence of each target power grid node.

[0325] In an optional implementation, the revenue adaptation determination module 5 is further configured to: The photovoltaic power output forecast sequence is used to determine the photovoltaic power output forecast for each candidate region in each forecast period.

[0326] The node spot price for each candidate region corresponding to the target power grid node in each prediction period is determined based on the node spot price series.

[0327] The photovoltaic (PV) power output for each forecast period is used to weight the spot prices at the corresponding nodes for that forecast period to obtain the PV weighted average spot price.

[0328] The probability of low-price periods and / or negative electricity price periods is determined based on the spot price series of nodes.

[0329] Based on the weighted average spot price of photovoltaic power and the probability of low-price periods and / or negative electricity price periods, the corresponding revenue matching index for each candidate region is determined.

[0330] In an optional implementation, the planning scheme generation module 6 is further configured to: The recommended installed capacity for each candidate region is used as the decision variable.

[0331] Based on the preset power generation weight coefficient, absorption weight coefficient, and revenue weight coefficient, the power generation potential index, absorption capacity index, and revenue matching index corresponding to each candidate region are weighted and processed to obtain the planning matching score corresponding to each candidate region.

[0332] Based on the recommended installed capacity and planning adaptation score of each candidate region, the single-point planning benefits corresponding to each candidate region are determined.

[0333] A planning model is constructed with the objective of maximizing the sum of planning benefits for each single point corresponding to all candidate areas within the target area.

[0334] The planning model is solved under the pre-defined constraints to obtain the suggested installed capacity for each candidate region.

[0335] Based on the suggested installed capacity for each candidate region, a distributed photovoltaic planning scheme is generated.

[0336] The computer program product provided in this application includes a computer-readable storage medium storing program code. The instructions included in the program code can be used to execute the methods described in the preceding method embodiments. For specific implementation details, please refer to the method embodiments, which will not be repeated here.

[0337] Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the specific working process of the system and apparatus described above can be referred to the corresponding process in the foregoing method embodiments, and will not be repeated here.

[0338] Furthermore, in the description of the embodiments of this application, unless otherwise expressly specified and limited, the terms "installation," "connection," and "linking" should be interpreted broadly. For example, they can refer to a fixed connection, a detachable connection, or an integral connection; they can refer to a mechanical connection or an electrical connection; they can refer to a direct connection or an indirect connection through an intermediate medium; and they can refer to the internal connection of two components. Those skilled in the art can understand the specific meaning of the above terms in this application based on the specific circumstances.

[0339] If the aforementioned functions are implemented as software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, or a portion 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 described in the various embodiments of this application. 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.

[0340] This application also provides an electronic device, which is shown in the schematic diagram of the structure of the electronic device. The electronic device includes a processor and a memory. The memory stores computer-executable instructions that can be executed by the processor. The processor executes the computer-executable instructions to implement the above-mentioned method for identifying the path to be planned.

[0341] Electronic devices also include buses and communication interfaces, in which the processor, communication interface, and memory are connected via a bus.

[0342] The memory may include high-speed random access memory (RAM) and may also include non-volatile memory, such as at least one disk drive. Communication between this system network element and at least one other network element is achieved through at least one communication interface (wired or wireless), which can use the Internet, wide area network, local area network, metropolitan area network, etc. The bus can be an ISA (Industry Standard Architecture) bus, a PCI (Peripheral Component Interconnect) bus, or an EISA (Extended Industry Standard Architecture) bus, etc. The bus can be divided into address bus, data bus, control bus, etc.

[0343] The processor may be an integrated circuit chip with signal processing capabilities. In implementation, the steps of the above method can be completed by integrated logic circuits in the processor's hardware or by software instructions. The processor can be a general-purpose processor, including a Central Processing Unit (CPU), a Network Processor (NP), etc.; it can also be a Digital Signal Processor (DSP), an Application Specific Integrated Circuit (ASIC), a Field-Programmable Gate Array (FPGA), or other programmable logic devices, discrete gate or transistor logic devices, or discrete hardware components. The general-purpose processor can be a microprocessor or any conventional processor. The steps of the method disclosed in the embodiments of this application can be directly implemented by a hardware decoding processor, or by a combination of hardware and software modules in the decoding processor. The software modules can reside in random access memory, flash memory, read-only memory, programmable read-only memory, electrically erasable programmable memory, registers, or other mature storage media in the art. This storage medium is located in memory; the processor reads information from the memory and, in conjunction with its hardware, completes the steps of the distributed photovoltaic planning method of the aforementioned embodiments.

[0344] In the description of this application, it should be noted that the terms "center," "upper," "lower," "left," "right," "vertical," "horizontal," "inner," and "outer," etc., indicate the orientation or positional relationship based on the orientation or positional relationship shown in the accompanying drawings. They are used only for the convenience of describing this application and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation. Therefore, they should not be construed as limitations on this application. Furthermore, the terms "first," "second," and "third" are used for descriptive purposes only and should not be construed as indicating or implying relative importance.

[0345] Finally, it should be noted that the above-described embodiments are merely specific implementations of this application, used to illustrate the technical solutions of this application, and not to limit them. The scope of protection of this application is not limited thereto. Although this application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that any person skilled in the art can still modify or easily conceive of changes to the technical solutions described in the foregoing embodiments, or make equivalent substitutions for some of the technical features, within the scope of the technology disclosed in this application. Such modifications, changes, or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of this application, and should all be covered within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of protection of the claims.

Claims

1. A distributed photovoltaic planning method, characterized in that, include: Acquire basic data of the target area, and based on the basic data, filter out at least one candidate area that can be deployed, and determine the target power grid node corresponding to each candidate area; Based on the photovoltaic resource data in the basic data, a photovoltaic output prediction sequence for each candidate region is generated, and the power generation potential index corresponding to each candidate region is determined. Based on the grid node operation data, grid structure parameters and photovoltaic output prediction sequence in the basic data, the absorption capacity index corresponding to each candidate region is determined. Based on the system supply and demand data, new energy power output forecast data, local load forecast data, local regulation resource capacity data, critical line power flow forecast data, line capacity data, and historical node spot price data in the basic data, as well as the photovoltaic power output forecast sequence, the node price forecast characteristics corresponding to each target grid node are determined, and the node spot price sequence of each target grid node is predicted based on the node price forecast characteristics. Based on the matching relationship between the spot price sequence of the nodes and the photovoltaic power output prediction sequence, the revenue matching index corresponding to each candidate region is determined. A planning model is constructed with the power generation potential index, the absorption capacity index, and the revenue matching index as optimization objectives, and solved under preset constraints to obtain a distributed photovoltaic planning scheme that includes the suggested installed capacity for each candidate region.

2. The distributed photovoltaic planning method according to claim 1, characterized in that, The basic data includes buildable resource data, planning constraint data, and geospatial data; The steps of acquiring basic data of the target area, filtering out at least one candidate area for layout based on the basic data, and determining the target power grid node corresponding to each candidate area include: Based on the planning constraint data, the restricted areas within the target region are determined; The restricted area is removed from the target area range corresponding to the geospatial data to obtain the effective operating area; The effective working area is spatially discretized based on a preset size grid to obtain multiple basic grids; The actual available construction area within each of the basic grids is determined based on the constructable resource data. The basic grid with an actual usable construction area greater than a preset area threshold is identified as the candidate region; Based on the geographical location of the candidate region, the location of the power grid access point, and the distribution network topology, the target power grid node corresponding to each candidate region is determined.

3. The distributed photovoltaic planning method according to claim 1, characterized in that, The photovoltaic resource data includes solar irradiance data, temperature data, and photovoltaic module parameters; The step of generating a photovoltaic output prediction sequence for each candidate region based on the photovoltaic resource data in the basic data, and determining the power generation potential index corresponding to each candidate region, includes: Extract the solar irradiance data, temperature data, and photovoltaic module parameters corresponding to each candidate region from the photovoltaic resource data; Based on the solar irradiance data, the temperature data, and the photovoltaic module parameters, the photovoltaic output per unit installed capacity of each candidate region is calculated within the predicted time scale. Based on the time granularity of the predicted time scale, the photovoltaic output per unit installed capacity is arranged to generate the photovoltaic output prediction sequence. The expected power generation of each candidate region within the prediction timescale is calculated based on the photovoltaic power output prediction sequence. The expected power generation is normalized to obtain the power generation potential index corresponding to each candidate region.

4. The distributed photovoltaic planning method according to claim 1, characterized in that, The power grid node operation data includes node load data, line power flow data, main transformer load rate data, and bus voltage data. The step of determining the grid absorption capacity index corresponding to each candidate region based on the grid node operation data, grid structure parameters, and photovoltaic output prediction sequence in the basic data includes: The local load characteristic parameters of each target power grid node are determined based on the node load data. Based on the power grid structure parameters, the line power flow data, the main transformer load rate data, and the bus voltage data, the power grid strength parameters of each target power grid node are determined. Based on the photovoltaic output prediction sequence and the node load data, determine the node reverse power flow risk parameter for each target grid node; The local load characteristic parameters, the power grid strength parameters, and the node reverse power flow risk parameters are weighted to obtain the absorption assessment value corresponding to each candidate region. The absorption assessment value is normalized to obtain the absorption capacity index corresponding to each candidate region.

5. The distributed photovoltaic planning method according to claim 4, characterized in that, The step of determining the node reverse power flow risk parameter for each target grid node based on the photovoltaic output prediction sequence and the node load data includes: The distributed photovoltaic predicted output of each target grid node in each prediction period is determined based on the photovoltaic output prediction sequence. Based on the node load data, determine the local predicted load of each target power grid node in each predicted period; Calculate the difference between the predicted output of the distributed photovoltaic system and the predicted local load, and determine the difference that is greater than zero as the reverse power flow prediction value for each target grid node in each prediction period; The reverse flow risk coefficient is determined based on the maximum value in the reverse flow prediction and the transformer capacity corresponding to the target grid node. The reverse current risk parameter of the node is determined based on the reverse current risk coefficient.

6. The distributed photovoltaic planning method according to claim 1, characterized in that, The node price prediction features include system supply and demand tension features, new energy output features, node reverse power flow risk features, transmission congestion features, and price lag features; The step of determining the node price prediction characteristics corresponding to each target grid node based on the system supply and demand data, new energy power output prediction data, local load prediction data, local regulation resource capacity data, critical line power flow prediction data, line capacity data, and historical node spot price data in the basic data, as well as the photovoltaic power output prediction sequence, includes: Based on the system's supply and demand data, the characteristics of the system's supply and demand tension are determined; Based on the predicted power output data of the new energy source and the predicted power output sequence of the photovoltaic power generation, the characteristics of the new energy source power output are determined; Based on the photovoltaic output prediction sequence, the local load prediction data, and the local regulation resource capacity data, the reverse power flow risk characteristics of the node are determined; Based on the predicted power flow data of the critical line and the line capacity data, the transmission congestion characteristics are determined; Based on the historical spot price data, the price lag characteristic is determined.

7. The distributed photovoltaic planning method according to claim 6, characterized in that, The step of predicting the node spot price sequence for each target power grid node based on the node price prediction characteristics includes: The node price prediction features corresponding to each target power grid node are input into a pre-trained node spot price prediction model; The node spot price prediction model outputs the node spot price for each target power grid node in each prediction period within the prediction time scale. Arrange the spot prices of each predicted time period according to the time granularity of the predicted time scale to obtain the spot price sequence of each target power grid node.

8. The distributed photovoltaic planning method according to claim 1, characterized in that, The step of determining the revenue matching index corresponding to each candidate region based on the matching relationship between the spot price sequence of the nodes and the photovoltaic output prediction sequence includes: The photovoltaic power output prediction sequence is used to determine the photovoltaic power output prediction for each candidate region in each prediction period. Based on the node spot price sequence, determine the node spot price of the target power grid node corresponding to each candidate region in each prediction period; The photovoltaic power output forecast for each forecast period is used to weight the spot price at the corresponding node for that forecast period to obtain the photovoltaic weighted average spot price. Determine the probability of low-price periods and / or negative electricity price periods based on the spot price sequence of the nodes; Based on the photovoltaic weighted average spot price, the probability of low price periods, and / or the probability of negative electricity price periods, the revenue matching index corresponding to each candidate region is determined.

9. The distributed photovoltaic planning method according to claim 1, characterized in that, The step of constructing a planning model with the power generation potential index, the grid absorption capacity index, and the revenue matching index as optimization objectives, and solving it under preset constraints to obtain a distributed photovoltaic planning scheme containing the suggested installed capacity for each candidate region, includes: The recommended installed capacity of each candidate region is used as the decision variable; Based on preset power generation weight coefficient, absorption weight coefficient and revenue weight coefficient, the power generation potential index, absorption capacity index and revenue adaptation index corresponding to each candidate region are weighted and processed to obtain the planning adaptation score corresponding to each candidate region. Based on the recommended installed capacity of each candidate region and the planning adaptation score, the single-point planning benefit corresponding to each candidate region is determined. The planning model is constructed with the objective of maximizing the sum of the planning benefits of all candidate regions within the target area. The planning model is solved under the preset constraints to obtain the suggested installed capacity for each candidate region. The distributed photovoltaic planning scheme is generated based on the suggested installed capacity corresponding to each candidate region.

10. A distributed photovoltaic planning system, characterized in that, include: The candidate region determination module is used to acquire basic data of the target region, and filter out at least one candidate region that can be laid out based on the basic data, and determine the target power grid node corresponding to each candidate region; The power generation potential determination module is used to generate a photovoltaic output prediction sequence for each candidate region based on the photovoltaic resource data in the basic data, and to determine the power generation potential index corresponding to each candidate region. The grid absorption capacity determination module is used to determine the grid absorption capacity index corresponding to each candidate region based on the grid node operation data, grid structure parameters and photovoltaic output prediction sequence in the basic data. The node price prediction module is used to determine the node price prediction characteristics corresponding to each target grid node based on the system supply and demand data, new energy predicted output data, local load prediction data, local regulation resource capacity data, critical line predicted power flow data, line capacity data and historical node spot price data in the basic data, as well as the photovoltaic output prediction sequence, and predict the node spot price sequence of each target grid node based on the node price prediction characteristics. The revenue matching determination module is used to determine the revenue matching index corresponding to each candidate region based on the matching relationship between the spot price sequence of the node and the photovoltaic power output prediction sequence. The planning scheme generation module is used to construct a planning model with the power generation potential index, the absorption capacity index and the revenue matching index as optimization objectives, and solve it under the condition of satisfying preset constraints to obtain a distributed photovoltaic planning scheme containing the suggested installed capacity of each candidate region.

Images