A method for simulating and evaluating a building integrated photovoltaics design

Abstract

The present application relates to the technical field of building integrated photovoltaics, in particular to a photovoltaic building integrated design simulation evaluation method, comprising: dividing photovoltaic components in building design data into a plurality of grids, and establishing a binding relationship between the grids and load areas; randomly selecting an initial control point from the grid space, simulating the shading characteristics and load characteristics of each initial control point under all feasible configuration combinations, and establishing a control point set; performing shadow interaction correction near the control point to preliminarily demarcate the consumption characteristics of the photovoltaic building; based on the consumption characteristics of each load area, extracting the consumption volatility, potential self-sufficiency rate and power grid impact risk of each area, and using the same to construct a cost matrix to generate regional early warning coordinates; using the regional early warning coordinates for global optimization solution, and outputting a photovoltaic layout scheme. The utilization efficiency and interactivity of the photovoltaic components are improved.

Description

Technical Field

[0001] This invention relates to the field of building-integrated photovoltaics (BIPV) technology, specifically a method for BIPV design simulation and evaluation. Background Technology

[0002] Building-integrated photovoltaics (BIPV) is a technology that uses photovoltaic modules installed on the exterior of buildings to generate electricity. This technology not only meets the requirements of architectural aesthetics, lighting, and safety, but also avoids the land occupation of photovoltaic modules. However, current technology analyses often emphasize the simulation of single power generation effects or meteorological effects, neglecting the interactive arrangement of modules, thus affecting the overall utilization efficiency of photovoltaic power generation.

[0003] For example, Chinese Patent Publication No. CN116070335A discloses an integrated photovoltaic building design method, which relates to the field of photovoltaic building integration technology. It solves the technical problem that the existing technology does not consider the layout and replacement costs of photovoltaic modules, resulting in increased design costs for photovoltaic buildings. This invention simulates and calculates the effective solar radiation of each grid, and predicts the power generation by combining climate data and power prediction models of the area where the photovoltaic building is located. It calculates the power generation rate of return based on the module cost, and judges whether it is appropriate to set up photovoltaic modules at the grid based on the power generation rate of return. This invention determines whether to set up photovoltaic cost based on the rate of return, which can reduce the layout cost of photovoltaic modules. After judging whether each grid on the outer surface of the photovoltaic building is qualified, this invention combines qualified and adjacent grids to generate a module setting area, and then further expands the module setting area according to the external rule contour.

[0004] For example, Chinese Patent Publication No. CN119989468A discloses the field of solar power generation technology, and in particular relates to a method for deploying photovoltaic modules. In this invention, a three-dimensional model is constructed based on the environment where the photovoltaic modules are to be deployed and the specific parameters of the photovoltaic modules. Considering that excessive reflectivity of photovoltaic modules can cause light pollution and affect the space utilization of surrounding buildings, this invention preliminarily determines whether the deployment of photovoltaic modules is qualified based on the reflective area of ​​the photovoltaic modules on surrounding buildings in a single detection cycle. If the deployment is initially determined to be unqualified, the detection cycle is divided into several sub-cycles, and further determination is made based on the fluctuation of the reflective area in each sub-cycle, or the reasons for the unqualified deployment are analyzed based on the number of photovoltaic modules installed.

[0005] In existing technologies, the grid distribution is planned based on the maximum solar irradiance to determine the simulation status of photovoltaic modules; or the reflectance of the illumination angle is used to determine whether the photovoltaic module layout is qualified. Existing technologies tend to focus on the photovoltaic module layout mapping estimation problem, which is a relative reflection of the independent operation of the modules. This easily ignores the load competition and mutual shading problems under the operation of photovoltaic modules, thus causing a decrease in the accuracy of photovoltaic module power prediction. Summary of the Invention

[0006] To solve the above-mentioned technical problems, the technical solution adopted by the present invention is: a photovoltaic building integrated design simulation and evaluation method, including: S1, acquiring building design data, dividing the photovoltaic modules in the building design data into several grids, and establishing the binding relationship between the grids and the load area.

[0007] S2, randomly select initial control points from the grid space, simulate the shading and load characteristics of each initial control point under all feasible configuration combinations, and establish a control point set.

[0008] S3, perform shadow interaction correction near the control point, determine the hourly power generation under the joint power supply of multiple control points, and initially delineate the absorption characteristics of photovoltaic buildings.

[0009] S4, based on the absorption characteristics of each load region, extracts the absorption volatility, potential self-sufficiency rate and grid impact risk of each region, and uses this to construct a cost matrix to generate regional early warning coordinates.

[0010] S5 uses regional early warning coordinates to perform global optimization, determines the optimal configuration combination for each location, performs sensitivity analysis based on the optimal configuration combination, and outputs a photovoltaic layout scheme.

[0011] The beneficial effects of this invention are as follows: First, this invention extracts installable polygonal areas from building design data and uses axis-aligned bounding boxes for initial division; it divides the area equally according to a grid granularity that perfectly matches the physical dimensions of commercial photovoltaic modules, establishes an electrical binding relationship between the grid and the building load area, ensures that the simulated photovoltaic modules are aligned with the current building granularity, and improves the positional accuracy of the modules in multiple simulations.

[0012] Second, this invention generates all feasible configuration combinations by component type, installation method, and layout method, and filters them by tilt angle constraint and structural bearing capacity constraint. After calculating the annual average shading coefficient and average load of each grid, the grid is divided into multiple feature groups, and each feature group is assigned at least one control point. The spatial distribution uniformity of control points is ensured by minimum distance constraint. While reducing the overall computation, numerical values ​​are extracted from all feature groups by hierarchical sampling to improve the data accuracy of subsequent analysis and processing.

[0013] Third, this invention defines the shadow influence domain of each control point, determines the shading relationship between control points based on spatial distance and relative azimuth, and constructs a shadow interaction matrix; it corrects the hourly power generation of the rear row control points and superimposes the shading effects of all the front row control points; it allocates power generation based on electrical distance weights and calculates the absorption characteristics under multi-control point joint power supply. It quantifies the dynamic self-shading effect between components, simulates the physical law of power priority flow in low-voltage power distribution systems, reflects the dynamic matching relationship between photovoltaic output and building load, and provides a reliable data foundation for subsequent multi-dimensional evaluation.

[0014] Fourth, this invention extracts three indicators from the absorption characteristic sequence: absorption volatility, potential self-sufficiency rate, and grid impact risk; constructs a multi-dimensional cost matrix containing the above three dimensions, aligns them with the early warning conditions, uses a genetic algorithm to quantify the configuration combination of photovoltaic modules, and uses the grid change affected by multiple parameters as the flat plate standard to obtain the most stable photovoltaic layout scheme, thereby reducing the uncertainty risk in the project implementation process, optimizing the processing method of the overall scheme combination, and improving the utilization efficiency of photovoltaic modules. Attached Figure Description

[0015] The present invention will be further described below with reference to the accompanying drawings and embodiments.

[0016] Figure 1 This is a flowchart illustrating a method for simulating and evaluating building-integrated photovoltaics (BIPV) design.

[0017] Figure 2 This is a flowchart illustrating step S2 of a photovoltaic building-integrated design simulation and evaluation method.

[0018] Figure 3 This is a flowchart illustrating step S3 of a photovoltaic building-integrated design simulation and evaluation method.

[0019] Figure 4 This is a flowchart illustrating step S4 of a photovoltaic building-integrated design simulation and evaluation method.

[0020] Figure 5 This is a schematic diagram showing the characteristic group distribution of the average shading coefficient and average load per unit area in a photovoltaic building-integrated design simulation and evaluation method.

[0021] Figure 6 This is a comparison curve of hourly power generation at control points before and after shadow interaction correction in a photovoltaic building integrated design simulation and evaluation method.

[0022] Figure 7 This is a convergence curve of the optimal individual comprehensive cost of a genetic algorithm in a photovoltaic building integrated design simulation and evaluation method;

[0023] Figure 8This is a sensitivity analysis curve of the grid configuration change ratio under different influencing factors in a photovoltaic building integrated design simulation and evaluation method. Detailed Implementation

[0024] The embodiments of the present invention are described in detail below. The embodiments described below are exemplary and are only used to explain the present invention, and should not be construed as limiting the present invention. Where specific techniques or conditions are not specified in the embodiments, they shall be performed in accordance with the techniques or conditions described in the literature in the art or in accordance with the product manual.

[0025] See Figure 1 A method for design simulation and evaluation of building-integrated photovoltaics (BIPV) includes: S1, acquiring building design data, dividing the photovoltaic modules in the building design data into several grids, establishing the binding relationship between the grids and the load area, with each grid corresponding to one photovoltaic module.

[0026] S2, randomly select initial control points from the grid space, simulate the shading and load characteristics of each initial control point under all feasible configuration combinations, and establish a control point set.

[0027] S3, perform shadow interaction correction near the control point, determine the hourly power generation under the joint power supply of multiple control points, and initially delineate the absorption characteristics of photovoltaic buildings.

[0028] S4, based on the absorption characteristics of each load region, extracts the absorption volatility, potential self-sufficiency rate and grid impact risk of each region, and uses this to construct a cost matrix to generate regional early warning coordinates.

[0029] S5 uses regional early warning coordinates to perform global optimization, determines the optimal configuration combination for each location, performs sensitivity analysis based on the optimal configuration combination, and outputs a photovoltaic layout scheme.

[0030] Photovoltaic modules are defined as devices that directly utilize photovoltaic energy, such as photovoltaic solar panels. Based on the combination of module type (monocrystalline silicon / double glass / thin film / BIPV, etc.), installation method (flat / fixed tilt / adjustable tilt / curtain wall / external mounting, etc.), and arrangement method (horizontal / vertical / adjustable spacing, etc.), the photovoltaic modules on each grid are recorded, along with the corresponding building type, to quantify the photovoltaic utilization efficiency at each location.

[0031] Before acquiring building design data, determine the layout of photovoltaic modules at each location in each simulation. Take the photovoltaic modules and corresponding areas deployed each time as the main body of analysis. Divide the photovoltaic building into multiple grids according to the surface that can be illuminated as the analysis unit. Based on the photovoltaic modules deployed in each grid, the output data is statistically analyzed to quantify its photovoltaic utilization efficiency.

[0032] One implementation of step S1 includes: S11, obtaining multiple installable polygonal regions from architectural design data, and initially dividing each installable polygonal region using an axis-aligned bounding box.

[0033] S12, for the axis-aligned regions, divide each region into an output grid according to the grid granularity corresponding to the photovoltaic module.

[0034] Furthermore, after finding the boundary coordinates of the polygon formed by axis alignment, a coordinate system corresponding to the photovoltaic building is generated. Horizontal and vertical grid lines are generated based on the selected grid granularity of this local coordinate system to complete the division of the region.

[0035] Furthermore, the selected grid granularity is based on the photovoltaic module being analyzed. For example, when analyzing a solar panel, if the solar panel is a 60-cell monocrystalline silicon module, an industry approximation of 1.6m × 1m should be used to set the grid granularity. In other words, for different photovoltaic module requirements, the grid granularity used for the current analysis is retrieved from the database according to the part to be analyzed.

[0036] Furthermore, in the axis-aligned mesh, most of the internal mesh is completely within the mountable polygon, but the edge mesh intersects with the polygon boundary, and some areas are outside the mountable area. It is necessary to remove the edge-damped mesh to obtain a relatively mountable mesh.

[0037] Specifically, it is required that the edge meshes be retained at least 80% of the mesh granularity to ensure the integrity of the region after the current meshing.

[0038] In the current step, the originally continuous and irregular building surface is transformed into a finite number of standardized discrete units, making it possible to perform numerical simulations of photovoltaic power generation, load matching, cost calculation, etc.

[0039] Furthermore, the load area represents the area bound to the corresponding grid, which can be a specific area such as a data center, server room, or 24-hour convenience store. These areas directly point to the electricity consumption under different building uses.

[0040] In one embodiment of the present invention, a feasible configuration combination is a combination of component type + installation method + layout method; based on the obtained grid space, the grid space is synchronized with globally distributed control points in an interval control manner.

[0041] Specifically, such as Figure 2 As shown, one implementation of step S2 includes: S21, setting up multiple configuration combinations of photovoltaic modules according to the combination of module type, installation method and arrangement method.

[0042] S22, For each configuration combination, determine the feasible configuration combination of the corresponding grid according to the installation tilt angle constraint and structural bearing constraint of the corresponding grid.

[0043] S23, in response to the average occlusion factor and average load at the location of the feasible configuration combination, filters the output control points.

[0044] Specifically, the aforementioned installation tilt angle constraint indicates whether the current grid area can adopt the corresponding installation method and layout. If installation is possible, it is considered to meet the installation tilt angle constraint. Similarly, the structural bearing capacity constraint determines whether the corresponding photovoltaic module can be set at the corresponding location. If it can be set, the feasible configuration combination at the corresponding location is determined from all possible photovoltaic configuration combinations.

[0045] In the current processing, the average occlusion coefficient processing flow includes: First, taking the current mesh space as the main body, extracting all objects that may cause occlusion, and establishing a three-dimensional occlusion model.

[0046] The extracted shading sources include the building's own shading sources and surrounding shading sources. The building's own shading sources include ventilation ducts, solar water heaters, parapet walls, chimneys, elevator machine rooms, air conditioning outdoor units, and roof skylights. Surrounding shading sources include adjacent buildings, tall trees, mountains, and utility poles. The three-dimensional coordinates, height, and shape of these shading sources are recorded. This data is the preliminary data before the photovoltaic modules are deployed and represents the basis for all photovoltaic module simulations.

[0047] Next, the NREL solar position algorithm is used to calculate the hourly solar altitude angle and azimuth angle throughout the year, obtain the solar altitude angle and azimuth angle of the three-dimensional shading model at all times, and calculate the radiance of the tilted surface in the unshading case.

[0048] Then, starting from the center point of each grid, a ray is established between the grid center and the sun based on the solar altitude angle and azimuth angle to determine the occlusion state of each grid with the 3D occlusion model. If the ray intersects, it means that it intersects with any occlusion source, and the corresponding grid is occluded. The occlusion state is set to 1. Otherwise, it is considered that it is not occluded, and the occlusion state is set to 0.

[0049] The average shading coefficient of the output is determined based on the values ​​of the radiance of the inclined surface at each time point under both shading and unshading conditions.

[0050] In this process, based on feasible configuration combinations of corresponding grid positions, all surrounding objects that may cause shading are identified, the average shading coefficient of the corresponding grid is calculated, and the proportion of solar radiation that each grid can receive throughout the year is quantified to the total solar radiation received. The quantification method is as follows: ;in, The average shading coefficient represents the proportion of unshaded radiation to total radiation at all times of daylight throughout the year for grid i. This represents the total number of effective sunshine hours throughout the year, and is usually selected when the solar altitude angle is greater than 5 degrees. This represents the amount of solar radiation on the tilted surface of grid i under unobstructed conditions at time t; This represents the occlusion state of grid i at time t.

[0051] Furthermore, the solar radiation on the tilted surface is composed of a combination of direct load and diffuse radiation, as shown below: ;in, This represents the direct radiation on the horizontal surface at time t, measured using meteorological source data. The scattered radiation at time t is also obtained through meteorological measurements, or it can be approximated by multiplying the difference between the total radiation at the horizontal surface and the direct radiation by 0.9, thereby quantifying the relative state of the scattered radiation. This represents the total radiation on the horizontal surface at time t. Indicates the angle of inclination of the inclined surface. Indicates ground reflectivity; This indicates the ratio of direct radiation from an inclined surface to a horizontal surface.

[0052] As for the average load, it represents the average load of the corresponding grid-bound load area. That is, after labeling each photovoltaic module with its building electrical circuit and the corresponding house type or room of each grid, the annual average power load per unit area is calculated, and this value is bound to each photovoltaic module to obtain the average load corresponding to each photovoltaic module; that is, quantifying how many watts of power are needed per square meter of building area at any given time.

[0053] Furthermore, in response to the average shading coefficient and average load, when screening the output control points, multiple feature groups are selected using the average shading coefficient, average load, and configurable installation combinations for each grid. Control points are randomly selected from each feature group, and the number of selected control points is based on the total number of grids, such as selecting 5% of the total number of grids as the initial control portion; each feature group includes at least one initial control point.

[0054] Furthermore, when dividing each feature group, based on the specific load areas such as data centers, server rooms, and 24-hour convenience stores bound to the grid, the average shading coefficient and average load are divided into combinations of high / medium / low load + high / medium / low shading. An average shading coefficient greater than 0.9 is considered low shading, representing almost no shading throughout the year and optimal lighting resources; an average shading coefficient between 0.7 and 0.9 is considered medium shading, representing partial shading and good lighting resources; an average shading coefficient less than 0.7 is considered low shading, representing significant shading from tall buildings, dense trees, etc., and poor lighting resources.

[0055] Furthermore, the unit area load of all functional buildings corresponding to the load areas is used as endpoints, including the annual average peak and annual average load values. The areas corresponding to high load, medium load and low load are divided from large to small, so that the photovoltaic modules currently being analyzed can be associated with the actual electricity consumption, thereby determining the self-consumption situation of the corresponding area. The resulting area can cover as many electricity consumption scenarios as possible, avoiding the problem of insufficient scenario simulation.

[0056] Using the Euclidean distance between initial control points, the initial control points are arranged in a spacing ratio and quantized uniformly. After the initial control points meet the spacing requirements and spatial uniformity, the corresponding initial control points are regarded as the output control points.

[0057] First, based on the Euclidean distance between all control points, initial control points less than three times the grid width are removed until the distance between all initial control points is greater than three times the grid width. This is considered to meet the spacing requirements and avoids excessive clustering of initial control points, which would lead to over-coupling of the overall simulation results. Then, the ratio of the standard deviation to the average distance between all initial control points is calculated. This value represents the relative fluctuation of the distance between initial control points. The smaller the value, the more uniform the overall distribution. In the current scenario, when this ratio is set to be less than or equal to 0.3, it is considered to meet the spatial uniformity requirement, and the selected control points are output.

[0058] In this embodiment, by ensuring that each possible combination of features has a representative control point, the situation of having no data to rely on during subsequent interpolation is avoided; at the same time, the distance between control points can further improve the regional utilization and spatial uniformity of photovoltaic building simulation, and avoid the impact of extreme feature groups on the accuracy of integrated simulation; at the same time, reducing the number of control points can improve the efficiency of subsequent processing while reducing the computational rate, ultimately achieving the effect of improving the accuracy of photovoltaic control.

[0059] To further illustrate the specific processing procedures for control point selection, shadow interaction correction, regional early warning, and global optimization described in this invention, the following simulation examples are provided. The data in these examples are calculated according to the aforementioned processing steps in this embodiment and are used to characterize the correspondence between various parameters and the corresponding technical effects. They do not constitute a limitation on building scale, component parameters, or judgment thresholds.

[0060] The example building has an installable surface area of ​​1920 m², and an initial grid size of 1.6 m × 1.0 m was used for mesh generation. After screening based on edge integrity, installation tilt angle, and structural load-bearing constraints, 960 feasible meshes were obtained. 48 initial control points were extracted, representing 5% of the total mesh size. Then, based on a minimum spacing constraint of three times the mesh width and a uniformity constraint of a distance variation coefficient not exceeding 0.3, 36 control points were ultimately retained. Example data for each feature group are shown in Table 1.

[0061] Table 1. Statistical table of control point feature grouping and screening results;

[0062]

[0063] From Table 1 and Figure 5 It can be seen that each control point covers a combination of low, medium, and high shading areas, as well as low, medium, and high load areas, and the number of control points is allocated according to the number of feasible grids for each feature group. This hierarchical selection method allows control points to simultaneously preserve differences in illumination resources and load demands, reducing representativeness bias caused by random sampling based solely on spatial location.

[0064] In one embodiment of the present invention, the photovoltaic module deployment at the control point is simulated, and the data is superimposed according to the shading situation after the module deployment. The power generation loss caused by self-shading of the photovoltaic modules at each installation location is determined, and the change in building load caused by the loss is quantified, thereby assisting in the selection of the optimal configuration combination.

[0065] During this process, the power generation of photovoltaic modules after being bound to the corresponding area will be judged in real time, and the actual power generation will be recorded in real time to determine whether it can be used for self-consumption or transmitted to adjacent areas. The hourly load allocation under the joint power supply of multiple control points will be determined, and the load change before and after installation will be quantified.

[0066] Specifically, such as Figure 3 As shown, one implementation of step S3 includes: S31, starting from each control point, extending backward along the direction perpendicular to the long side of the component, defining the shadow influence domain of each control point, and the extension length is the maximum shadow length under the current configuration.

[0067] The maximum shadow length is generally determined by the solar altitude angle at 9:00 AM on the local winter solstice, which is used as the current quantified standard value to determine the extension length after the component is installed.

[0068] S32, for any two control points, determine whether the adjacent control points are located within the shadow influence domain of the current control point, and determine the hourly power generation after shadow interaction correction.

[0069] During shadow interaction, select any two control points as a data pair and calculate the spatial distance and relative azimuth between the two points respectively.

[0070] When the spatial distance between two points is less than the maximum shadow length and the relative azimuth angle is less than the preset angle, it is determined that the adjacent control point is within the shadow influence domain of the current control point; the preset angle is set to 15 degrees. If the difference in azimuth angle between the two points exceeds 15 degrees, their shadows will not affect each other.

[0071] Subsequently, shadow interaction matrix quantization was performed on the two control points affected by shadowing to determine the hourly power generation after shadow interaction correction.

[0072] Specifically, the shadow interaction matrix is ​​used to record the occlusion situation under the corresponding configuration, where the value of each element is represented as follows: ;in, This indicates the occlusion of control point k on control point j when control point k is configured with j, and represents the occlusion of the current point by adjacent points when the shadow length is at its maximum. This represents the spatial distance between control point i and its adjacent control point k; This indicates the maximum shadow length when control point k uses configuration j. This indicates the solar altitude angle at 9:00 AM on the local winter solstice.

[0073] Then, the real-time occlusion status received at the current point is introduced to quantify the relative constraints of power generation at each location.

[0074] ;in, This indicates the hourly power generation of control point i using configuration j, representing the power generation simulated at time t. This value can be obtained through simulation using photovoltaic simulation software such as PVsyst, by inputting typical local meteorological year data. This represents the corrected hourly power generation at control point i after superimposing all shading at time t. The larger this value, the smaller the shadowing effect, and vice versa. This indicates the number of adjacent control points that are occluded with the current control point; This indicates the real-time occlusion status of the current control point i by adjacent control points at time t. Its calculation method is the same as that of each element in the shadow interaction matrix, which is to introduce the solar altitude angle at the corresponding time for judgment and processing.

[0075] It should be noted that for areas without obstructions, the simulated power generation is directly output.

[0076] Taking a control point in feature group G4 as an example, the shadow interaction matrix is ​​calculated according to the sun position from 8:00 to 17:00 on a typical day and the spatial relationship between adjacent control points. The real-time shading status of adjacent control points is then superimposed on the original power generation to obtain the hourly correction results shown in Table 2.

[0077] Table 2. Hourly power generation data before and after shadow interaction correction;

[0078]

[0079] From Table 2 and Figure 6It can be seen that the control point is significantly affected by shading interaction during the low solar altitude angle period in the morning, and the power generation at 9:00 is corrected from 1.18kW to 0.88kW. As the solar altitude angle increases, the shading loss rate decreases to about 7% from 12:00 to 13:00, and then increases again when the solar altitude angle decreases in the afternoon. Therefore, the hourly correction results can reflect the non-constant shading loss caused by the combined effect of the relative position of the components and the position of the sun, avoiding hourly power generation deviation caused by using a fixed reduction factor.

[0080] S33, the hourly power generation of each control point is merged according to the bound load area, and the merging weight is determined by the electrical distance between the control point and the load area, and the absorption characteristics of the load area under multi-point joint operation are recorded.

[0081] The photovoltaic (PV) absorption characteristic refers to the combined value of multi-point combined power supply after aggregation, representing the electricity savings in the corresponding area. The power generated by each grid is distributed to all load areas inversely proportional to the electrical distance; the closer the electrical distance, the greater the weight and the more power is allocated. In other words, after completing the hourly power generation correction calculation, for each load area, all power generation is weighted and summed based on the normalized reciprocal of its electrical distance. Then, the weighted sum is divided by the total power generation to obtain the PV absorption ratio sequence of each load area's total power generation at each time point. This ratio sequence is the absorption characteristic of that area. The higher the value of this ratio sequence, the greater the degree of PV replacement of grid power supply in that area at that time, and the greater the electricity savings; the lower the value, the more the area relies on the grid at that time, and PV plays no role.

[0082] It should be noted that the essence of power consumption is the real-time flow and distribution of electricity in the low-voltage power distribution system. This process is entirely determined by Ohm's law and is unrelated to the size of the power generation and load. Regardless of the amount of photovoltaic power generation, the electricity will flow to various load areas through the distribution network. Only after the electricity demand of all load areas is met will the excess electricity be fed into the grid. In this scenario, as long as there is a fluctuation in the net output of photovoltaic power generation and the total building load, it will be affected by photovoltaic power generation, which will lead to various forms of risks.

[0083] In the current embodiment, the electrical distance is the physical length of the line between the photovoltaic module and the load area. The longer the line, the greater the power generation loss. Using the electrical distance as a weight is essentially a correction for the absorption of power generation from a distance, making the aggregated power generation closer to the actual usable power in the area, and synchronizing all calculated power values ​​and the original load of the area with the output absorption characteristics.

[0084] In one embodiment of the present invention, since photovoltaic power itself has zero inertia and no primary frequency regulation capability, when regional power dispatch fluctuates, photovoltaic output cannot provide inertia support, which can easily lead to a sudden increase in the equivalent load of the power grid, causing anomalies such as a sudden drop in frequency, threatening the safe operation of the power grid.

[0085] Specifically, the data processing focuses on three relative risk dimensions corresponding to the photovoltaic (PV) absorption characteristics: absorption volatility, potential self-sufficiency rate, and grid impact risk. Absorption volatility reflects the stability of the PV absorption ratio in a region, representing the coefficient of variation of that ratio. Potential self-sufficiency rate reflects the average annual electricity consumption of a region that is replaced by PV, i.e., the proportion of PV power consumption in the region's total electricity consumption. Grid impact risk reflects the volatility of net power exchange between buildings and the grid, representing the standard deviation of the difference between PV output and the original load in the corresponding region, directly reflecting the potential impact of PV absorption on grid frequency. When the grid impact risk is greater than 0 in any region, it indicates that the volatility of PV output is greater than the volatility of load, which will impact the grid; the larger the difference, the higher the impact risk. These three factors are normalized and combined to form a cost matrix, quantifying the comprehensive risk level of each region under multi-point joint power supply, and output as early warning coordinates for regional hierarchical management.

[0086] like Figure 4 As shown, one implementation of step S4 includes: S41, based on the matrix dimension of the cost matrix, each matrix dimension is regarded as a judgment condition, and the warning judgment conditions of each region are determined by comparing them region by region.

[0087] The early warning criteria include extremely low absorption rate warning, high fluctuation warning, high power grid impact warning, and high overall cost warning.

[0088] The extremely low consumption warning is when the potential self-sufficiency rate corresponds to an excessively low characteristic value. A ratio threshold of 0.2 is set for the current region. When the proportion of photovoltaic power to the total electricity consumption in the region is less than this ratio threshold, it is considered that the photovoltaic power generation utilization rate is too low, resulting in the project being completely uneconomical.

[0089] A high volatility warning indicates that the value of absorption volatility is too large, representing an excessively large coefficient of variation in absorption in the region. Setting 0.8 as a threshold for the current region, when the absorption volatility value is greater than 0.8, it indicates that the absorption volume fluctuates very drastically, sometimes appearing and sometimes disappearing, and sometimes large and sometimes small, which will lead to extremely unstable building power consumption curves. It is necessary to use a small-capacity energy storage system to smooth out the fluctuations, or to use adjustable tilt brackets to improve the matching degree between power generation and load.

[0090] A high grid impact warning occurs when the grid impact risk is too high, and the standard deviation of the net photovoltaic output exceeds a specific value. In this case, a threshold of 0.7 is set for the current area. When the difference in the normalized standard deviation is greater than this value, it indicates that the fluctuation of photovoltaic output is much greater than the fluctuation of building load, which will have a significant impact on grid frequency and voltage stability, causing problems such as scheduling difficulties. It is necessary to mark the corresponding area and install energy storage systems to smooth the fluctuations, or configure reactive power compensation devices to stabilize the voltage.

[0091] The high overall cost warning is a weighted sum of three costs. Based on the order of absorbing volatility, potential self-sufficiency rate and grid impact risk, the weights are set to 0.3, 0.4 and 0.3 respectively, and the weights are adjusted in real time according to the building design requirements. Then, a threshold of 0.9 is set. When the overall cost exceeds this value, it means that the cost of the grid configuration combination is too high and performs poorly among all the options. The configuration combination and region need to be marked to determine the current configuration scheme to be implemented.

[0092] It should be noted that since the smaller the characteristic value corresponding to the potential self-sufficiency rate, the higher the cost, the characteristic value of the potential self-sufficiency rate will be subtracted from 1 to complete the weighted processing of the comprehensive cost.

[0093] S42, perform regional mapping on the early warning judgment conditions to determine the regional early warning coordinates under different combinations of conditions.

[0094] In this embodiment, by identifying early warning conditions for different load areas, the overall photovoltaic consumption situation of the region is transformed into a matrix form with multiple cost combinations, and the global early warning coordinates are set according to the early warning conditions recorded for each region.

[0095] The hourly power generation of the corrected control points was grouped into five load areas according to the inverse weight of electrical distance, and the absorption volatility, potential self-sufficiency rate, grid impact risk, and comprehensive cost were calculated for each area, resulting in the regional early warning judgment results shown in Table 3. The comprehensive cost was calculated as 0.3 × absorption volatility + 0.4 × (1 - potential self-sufficiency rate) + 0.3 × grid impact risk.

[0096] Table 3 Load Area Absorption Indicators and Early Warning Judgment Results;

[0097]

[0098] Table 3 shows that region C has a high potential self-sufficiency rate and low volatility, with a comprehensive cost of 0.329. Region D has volatility absorption and grid impact risk exceeding 0.8 and 0.7 respectively, generating high volatility warnings and high grid impact warnings accordingly. Region E simultaneously meets the criteria of extremely low absorption, high volatility, high grid impact, and high comprehensive cost. By mapping different warning conditions to corresponding regions, the genetic algorithm can impose higher costs on high-risk regions during global optimization.

[0099] In one embodiment of the present invention, the cost matrix corresponding to the early warning coordinates of the receiving area is used to optimize the global configuration combination through a genetic algorithm, and then the robustness is verified through sensitivity analysis, so as to finally complete the design of the photovoltaic module.

[0100] Specifically, when performing global optimization based on parameter constraints, one implementation of step S5 includes: S51, setting individuals for each grid using each feasible configuration combination, and randomly generating an initial population; each individual represents an integer sequence, where each element in the sequence represents the configuration selection for the corresponding grid, and the configuration combination is represented by a specific numerical code.

[0101] S52 synchronizes the cost matrix to all grids, sets the fitness using the reciprocal of the comprehensive cost, and obtains the optimal configuration combination after genetic analysis. Each individual is composed of the comprehensive costs of multiple grids. When setting the fitness of a corresponding individual, the reciprocal of the cumulative comprehensive cost of the individual's corresponding elements is used to ensure that the lower the cost, the higher the fitness.

[0102] Furthermore, the genetic algorithm sets the population size to 50-200, the crossover probability to 0.8, and allows 80% of the parents to undergo gene recombination to balance exploration and utilization capabilities; the mutation probability is set to 0.1, with each gene locus having a 10% probability of mutation, to avoid rapid convergence causing a decrease in overall accuracy; at the same time, the best 10% of individuals are retained in each generation, making the maximum number of iterations 200, and the convergence threshold is set to 1e-5, thus completing the basic parameter settings of the genetic algorithm.

[0103] In genetic analysis, firstly, three individuals are randomly selected each time, and the one with the highest fitness is retained.

[0104] Secondly, two intersection points are randomly selected, and gene segments between the two individuals are exchanged.

[0105] Next, a gene locus is randomly selected, and its value is changed to another random value other than the current combination.

[0106] Then, the best 10% of individuals from each generation are directly copied to the next generation to ensure that the optimal solution is not lost; this process is repeated until the convergence threshold is met, at which point the configuration combination set for the corresponding individual is output.

[0107] In the example computation with the above 960 feasible grids, the population size was set to 100, the crossover probability to 0.8, and the mutation probability to 0.1, and the top 10% of individuals with the highest fitness in each generation were retained. The overall cost of the optimal individual in the genetic algorithm varies with the number of iterations as shown in Table 4. Figure 7 As shown.

[0108] Table 4. Statistical table of iterative results of the optimal individual comprehensive cost in the genetic algorithm;

[0109]

[0110] From Table 4 and Figure 7 It can be seen that the overall cost decreases rapidly in the first 80 generations, and the change is less than 0.001 after the 120th generation, and remains stable at about 0.430. This indicates that the relevant operation can gradually eliminate high-cost configuration combinations and form a stable configuration result.

[0111] It should be noted that the data portion of the genetic optimization includes not only the grid corresponding to the regional early warning coordinates, but also grids at other locations, in order to complete the combined analysis under multiple feature combinations.

[0112] Furthermore, during the sensitivity analysis, the dimensions of the analysis include, but are not limited to, influencing factors such as annual solar radiation, component conversion efficiency, and building load. These are entered as the average value under normal conditions, the adjusted value of the average value ± 10%, and the adjusted value of the average value ± 20%, to determine the proportion of changes in grid configuration after the genetic algorithm optimization.

[0113] In the current scenario, these influencing factors will directly affect the simulated power generation, load saving ratio, and other factors, directly impacting the cost assessment of each grid. Therefore, it is necessary to identify the grid with the smallest percentage change due to parameter variations to obtain a photovoltaic layout scheme that is relatively insensitive to parameter changes.

[0114] Specifically, when performing sensitivity analysis based on the optimal configuration combination, the influencing factors of the optimal configuration combination are obtained. When the value range of the influencing factors changes, the proportion of change of each grid under the optimal configuration combination is determined, and the combination with the smallest proportion of grid change is regarded as the output photovoltaic layout scheme.

[0115] Using the optimal configuration combination obtained by the genetic algorithm as the benchmark, the annual solar radiation, component conversion efficiency and building load were varied by single factors within ±10% and ±20% of the benchmark values, respectively. The cost calculation and configuration combination selection were then re-executed to obtain the grid configuration variation ratio shown in Table 5.

[0116] Table 5. Statistics on the percentage change in grid configuration under different influencing factors;

[0117]

[0118] From Table 5 and Figure 8It can be seen that within the normal disturbance range of ±10%, the proportion of grid configuration change corresponding to the three types of factors does not exceed 9.8%; when the disturbance amplitude increases to ±20%, the change proportion still does not exceed 18.6%. Among them, the proportion of configuration change caused by building load disturbance is relatively low, indicating that the output layout scheme has good stability against load prediction deviation; the impact of annual solar radiation variation is relatively large, so in practical applications, priority should be given to improving the representativeness and temporal resolution of meteorological data.

[0119] Although embodiments of the present invention have been shown and described above, it is understood that the above embodiments are exemplary and should not be construed as limiting the present invention. Those skilled in the art can make changes, modifications, substitutions and variations to the above embodiments within the scope of the present invention, which are still covered within the protection scope of the present invention.

Claims

1. A method for simulating and evaluating building-integrated photovoltaics (BIPV) design, characterized in that, include: S1, acquire building design data, divide the photovoltaic modules in the building design data into several grids, and establish the binding relationship between the grids and the load area; S2, randomly select initial control points from the grid space, simulate the shading and load characteristics of each initial control point under all feasible configuration combinations, and establish a control point set; S3, perform shadow interaction correction near the control point, determine the hourly power generation under the joint power supply of multiple control points, and initially delineate the absorption characteristics of photovoltaic buildings; S4. Based on the absorption characteristics of each load region, the absorption volatility, potential self-sufficiency rate and grid impact risk of each region are extracted, and a cost matrix is ​​constructed to generate regional early warning coordinates. S5 uses regional early warning coordinates to perform global optimization, determines the optimal configuration combination for each location, performs sensitivity analysis based on the optimal configuration combination, and outputs a photovoltaic layout scheme.

2. The photovoltaic building integrated design simulation and evaluation method according to claim 1, characterized in that, The methods for dividing the photovoltaic module into several grids in step S1 include: S11. Obtain multiple installable polygonal regions from the architectural design data. For each installable polygonal region, use an axis-aligned bounding box for initial division. S12, for the axis-aligned regions, divide each region into an output grid according to the grid granularity corresponding to the photovoltaic module.

3. The photovoltaic building integrated design simulation and evaluation method according to claim 1, characterized in that, The methods for establishing the control point set in step S2 include: S21, set up multiple configuration combinations of photovoltaic modules according to the combination of module type, installation method and layout method; S22, For each configuration combination, determine the feasible configuration combination of the corresponding grid according to the installation tilt angle constraint and structural bearing constraint of the corresponding grid; S23, in response to the average occlusion factor and average load at the location of the feasible configuration combination, filters the output control points.

4. The photovoltaic building integrated design simulation and evaluation method according to claim 3, characterized in that, The average occlusion coefficient can be achieved in the following ways: Using the current mesh space as the main body, extract all objects that may cause occlusion and build a 3D occlusion model; Obtain the solar elevation angle and azimuth angle of the 3D occlusion model at all times, and calculate the radiance of the tilted surface in the unoccupied case; Starting from the center point of each grid, based on the solar altitude angle and azimuth angle, establish the ray between the grid center and the sun to determine the occlusion state of each grid and the 3D occlusion model; The average shading coefficient of the output is determined based on the values ​​of the radiance of the inclined surface at each time point under both shading and unshading conditions.

5. The photovoltaic building integrated design simulation and evaluation method according to claim 3, characterized in that, When filtering the output control points, the implementation methods include: Multiple feature groups are selected by using the average shading coefficient, average load, and configurable installation combinations for each grid, with each feature group including at least one initial control point; Using the Euclidean distance between initial control points, the initial control points are arranged in a spacing ratio and quantized uniformly. After the initial control points meet the spacing requirements and spatial uniformity, the corresponding initial control points are regarded as the output control points.

6. The photovoltaic building integrated design simulation and evaluation method according to claim 1, characterized in that, The methods for implementing the elimination feature in step S3 include: S31, starting from each control point, extend backward along the direction perpendicular to the long side of the component to define the shadow influence area of ​​each control point; S32, For any two control points, determine whether the adjacent control points are located within the shadow influence domain of the current control point, and determine the hourly power generation after shadow interaction correction; S33, the hourly power generation of each control point is merged according to the bound load area, and the absorption characteristics of the load area under multi-point joint operation are recorded.

7. The photovoltaic building integrated design simulation and evaluation method according to claim 6, characterized in that, In shadow interaction, the implementation methods include: Select any two control points as a data pair, and calculate the spatial distance and relative azimuth between the two points respectively; When the spatial distance between two points is less than the maximum shadow length and the relative azimuth angle is less than the preset angle, it is determined that the adjacent control point is located within the shadow influence domain of the current control point. For two control points affected by shadowing, shadow interaction matrix quantization is performed to determine the hourly power generation after shadow interaction correction.

8. The photovoltaic building integrated design simulation and evaluation method according to claim 1, characterized in that, The implementation methods for the regional early warning coordinates in step S4 include: S41, based on the matrix dimension of the cost matrix, each matrix dimension is regarded as a judgment condition and compared region by region to determine the early warning judgment condition for each region; S42, perform regional mapping on the early warning judgment conditions to determine the regional early warning coordinates under different combinations of conditions.

9. The photovoltaic building integrated design simulation and evaluation method according to claim 1, characterized in that, The optimal configuration combination in step S5 is implemented in the following ways: S51, individuals are set up for each grid using each feasible configuration combination, and an initial population is randomly generated; S52 synchronizes the cost matrix to all grids, sets the fitness based on the inverse of the comprehensive cost, and obtains the optimal configuration combination after genetic analysis.

10. The photovoltaic building integrated design simulation and evaluation method according to claim 1, characterized in that, When performing sensitivity analysis based on the optimal configuration combination, the implementation methods include: To determine the influencing factors of the optimal configuration combination, when the value range of the influencing factors changes, the proportion of change of each grid under the optimal configuration combination is determined, and the combination with the smallest proportion of grid change is regarded as the output photovoltaic layout scheme.

Images