A photovoltaic module arrangement scheme generation method, system and storage medium

Abstract

Embodiments of the present application provide a photovoltaic module arrangement scheme generation method and system and a storage medium. The method comprises: obtaining multi-source data of a target installation area; determining a ventilation coefficient and an irrigation coefficient of the target installation area according to building data, historical weather data and landscape greening data in the multi-source data; obtaining lighting demand data and indoor target heat gain rate of the target installation area; generating a first arrangement scheme according to the multi-source data, the lighting demand data, the indoor target heat gain rate, the ventilation coefficient and the irrigation coefficient; simulating a digital twin model of the first arrangement scheme and the multi-source data in cooperation and correcting the first arrangement scheme to obtain a target arrangement scheme. According to the multi-source data of the target installation area, a corresponding first arrangement scheme is generated, and the generation efficiency of the arrangement scheme is high. The first arrangement scheme is corrected by simulating the digital twin model to obtain the target arrangement scheme, and the perfection of the photovoltaic module arrangement scheme is improved.

Description

Technical Field

[0001] This invention relates to, but is not limited to, the field of data processing technology, and in particular to a method, system, and storage medium for generating photovoltaic module layout schemes. Background Technology

[0002] With the development of the photovoltaic industry and technology, the installation environment for photovoltaic modules is becoming increasingly diversified.

[0003] Before installation, photovoltaic modules need to be arranged in a specific layout. However, existing technologies rely on manual analysis to determine the layout, which is inefficient and requires comprehensive analysis of data from various sources, potentially leading to an imperfect layout. Summary of the Invention

[0004] The following is an overview of the subject matter described in detail herein. This overview is not intended to limit the scope of the claims.

[0005] The main objective of this invention is to provide a method, system, and storage medium for generating photovoltaic module layout schemes, which can improve the efficiency and completeness of photovoltaic module layout scheme generation.

[0006] In a first aspect, embodiments of the present invention provide a method for generating a photovoltaic module layout scheme, comprising:

[0007] Acquire multi-source data of the target installation area, including building data, historical weather data, human data, and landscape greening data;

[0008] The ventilation coefficient of the target installation area is determined based on the building data, the historical weather data, and the landscape greening data;

[0009] The irrigation coefficient for the target installation area is determined based on the historical weather data and the landscape greening data.

[0010] Obtain the lighting requirements data and indoor target heat gain rate of the target installation area;

[0011] A first layout scheme is generated based on the multi-source data, the lighting demand data, the indoor target heat gain rate, the ventilation coefficient, and the irrigation coefficient;

[0012] Construct a digital twin model of the first layout scheme and the multi-source data collaboration;

[0013] The simulation results were obtained after simulating the digital twin model.

[0014] The target layout scheme is obtained by correcting the first layout scheme based on the simulation results.

[0015] In some optional embodiments, determining the ventilation coefficient of the target installation area based on the building data, the historical weather data, and the landscape greening data includes:

[0016] A first model is constructed based on the building data and the landscape greening data;

[0017] The first ventilation volume is obtained by performing ventilation simulation on the first model based on the historical weather data.

[0018] A first coefficient is determined based on the first ventilation volume and the maximum ventilation volume, wherein the maximum ventilation volume represents the ventilation volume when there is no obstruction.

[0019] The seasonal correction factor is determined based on the historical weather data.

[0020] Determine the greening growth correction coefficient based on the aforementioned landscape greening data;

[0021] The ventilation coefficient is obtained by correcting the first coefficient based on the seasonal correction coefficient and the greening growth correction coefficient.

[0022] In some optional embodiments, determining the irrigation coefficient of the target installation area based on the historical weather data and the landscape greening data includes:

[0023] Obtain the rainfall and evaporation from the historical weather data;

[0024] Obtain the green area and water requirement for greening from the landscape greening data;

[0025] The basic water requirement is determined based on the green area and the water demand of the green area;

[0026] The amount of irrigation water is determined based on the rainfall and evaporation.

[0027] The irrigation coefficient is determined based on the irrigation water volume and the basic water requirement.

[0028] In some optional embodiments, generating the first layout scheme based on the multi-source data, the lighting requirement data, the indoor target heat gain rate, the ventilation coefficient, and the irrigation coefficient includes:

[0029] Construct a neural network fusion model combining multi-objective optimization algorithms and reinforcement learning frameworks;

[0030] The multi-source data, the lighting demand data, the indoor target heat gain rate, the ventilation coefficient, and the irrigation coefficient are input into the neural network fusion model;

[0031] The optimization objectives of the neural network fusion model are configured as maximizing photovoltaic power generation, maximizing indoor lighting satisfaction rate, maximizing indoor heat gain rate compliance rate, minimizing ventilation coefficient loss, and maximizing irrigation coefficient matching degree.

[0032] Construct the objective function corresponding to the optimization objective;

[0033] The design variables of the neural network fusion model are configured as photovoltaic module installation tilt angle, photovoltaic array spacing, photovoltaic module installation position, photovoltaic module orientation, photovoltaic module transmittance, and water collection tank coverage.

[0034] The constraints of the neural network fusion model are configured as structural load constraints, solar spacing constraints, human landscape constraints, and greening synergy constraints.

[0035] Based on the constraints, the objective function, and the optimization objective, the Pareto optimal solution set is obtained by iterating the design variables.

[0036] The first layout scheme is obtained by selecting the target solution from the Pareto optimal solution set based on the overall benefit requirements.

[0037] In some optional embodiments, obtaining the Pareto optimal solution set by iterating the design variables based on the constraints, the objective function, and the optimization objective includes:

[0038] A preset number of initial design variable combinations are randomly generated, and the initial design variable combinations satisfy the constraints.

[0039] The initial design variables are substituted into the objective function to calculate the fitness value of the optimization objective, and the fitness value represents the degree of matching with the optimization objective.

[0040] The non-dominated solution is obtained by performing a non-dominated sorting of the initial design variable combination based on the fitness value.

[0041] After performing gene recombination and random mutation on the non-dominated solution, a new combination of iterative design variables is generated, and the iteration number is incremented by one.

[0042] When the number of iterations is less than or equal to the preset number of iterations, the iterative design variable combination is configured as the initial design variable combination and then iterative calculation is performed. The iterative calculation represents the steps described above for generating the iterative design variable combination.

[0043] If the number of iterations is greater than a preset number, the non-dominated solution is configured as the Pareto optimal solution set.

[0044] In some optional embodiments, constructing the first layout scheme and the digital twin model of multi-source data collaboration includes:

[0045] Convert the building data into a BIM model;

[0046] The first layout scheme is superimposed onto the BIM model to obtain the first model;

[0047] The historical weather data is correlated with the first model to obtain the second model;

[0048] The third model is obtained by overlaying the humanistic data and the landscape greening data onto the second model.

[0049] Based on the first layout scheme, the historical weather data, the landscape greening data, the human data, and the building data, a first coupled model is generated that can calculate the hourly power generation, indoor heat gain rate, ventilation volume, daylighting rate, and water collection volume at each location. The first coupled model is then superimposed on the third model to obtain the digital twin model.

[0050] In some optional embodiments, after obtaining the simulation results by simulating the digital twin model, the process includes:

[0051] A simulation map is generated based on the simulation results. The simulation map is used to indicate the difference between the multi-dimensional simulation data and the target data at each location of the target installation area. The multi-dimensional simulation data characterizes the simulated hourly power generation, indoor heat gain rate, daylighting rate, ventilation volume, and water collection volume.

[0052] The simulated image is then subjected to grayscale conversion, filtering, and denoising processes in sequence to obtain a preprocessed image.

[0053] The preprocessed image is divided into multiple spatial image sub-regions;

[0054] Abnormal region information is obtained based on the similarity between multiple spatial image sub-regions. The abnormal region information represents the contour information and spatial location information of the target region that is significantly different from the adjacent regions.

[0055] In some optional embodiments, obtaining abnormal region information based on the similarity between the plurality of spatial image sub-regions includes:

[0056] The comparison result is obtained by comparing the first similarity between the first image sub-region and the third image sub-region and the second similarity between the second image sub-region and the third image sub-region. The third image sub-region is any of the spatial image sub-regions. The first image sub-region is located on the first side of the third image sub-region, and the second image sub-region is located on the second side of the third image sub-region. The first image sub-region, the second image sub-region, and the third image sub-region all represent the spatial image sub-region.

[0057] If the comparison result indicates that the first similarity is greater than the second similarity, the third image sub-region is adjusted by unit distance and scaled by unit scale sequentially towards the location of the second image sub-region, and the contour information and spatial information of the third image sub-region with the largest difference between the first similarity and the second similarity are configured as the abnormal region information;

[0058] If the comparison result indicates that the first similarity is less than the second similarity, the third image sub-region is adjusted by unit distance and scaled by unit scale sequentially towards the location of the first image sub-region, and the contour information and spatial information of the third image sub-region with the largest difference between the first similarity and the second similarity are configured as the abnormal region information;

[0059] If the comparison result indicates that the first similarity is equal to the second similarity, the first image sub-region and the second image sub-region are simultaneously moved by a unit distance and scaled by a unit until the first similarity is not equal to the second similarity.

[0060] In a second aspect, embodiments of the present invention provide a controller, including: a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the photovoltaic module layout scheme generation method described in the first aspect.

[0061] Thirdly, embodiments of the present invention provide a photovoltaic module layout scheme generation system, including the controller mentioned in the second aspect above.

[0062] Fourthly, a computer storage medium stores computer-executable instructions, which are used to execute the photovoltaic module layout scheme generation method described in the first aspect.

[0063] The beneficial effects of this invention include: acquiring multi-source data of the target installation area, including building data, historical weather data, human data, and landscape greening data; determining the ventilation coefficient of the target installation area based on the building data, historical weather data, and landscape greening data; determining the irrigation coefficient of the target installation area based on the historical weather data and landscape greening data; acquiring the light requirement data and indoor target heat gain rate of the target installation area; generating a first layout scheme based on the multi-source data, the light requirement data, the indoor target heat gain rate, the ventilation coefficient, and the irrigation coefficient; constructing a digital twin model of the first layout scheme and the multi-source data; simulating the digital twin model to obtain simulation results; and correcting the first layout scheme based on the simulation results to obtain the target layout scheme. The invention automatically generates the corresponding first layout scheme based on the multi-source data of the target installation area, resulting in high generation efficiency. By combining the first layout scheme and the multi-source data to generate a corresponding digital twin model and performing simulation, the first layout scheme is corrected to obtain the target layout scheme, thus improving the completeness of the photovoltaic module layout scheme.

[0064] Other features and advantages of the invention will be set forth in the description which follows, and will be apparent in part from the description, or may be learned by practicing the invention. The objects and other advantages of the invention may be realized and obtained by means of the structures particularly pointed out in the description, claims and drawings. Attached Figure Description

[0065] Figure 1 This is a flowchart illustrating the steps of a method for generating a photovoltaic module layout scheme according to an embodiment of the present invention.

[0066] Figure 2 This is a schematic diagram of a controller provided in one embodiment of the present invention.

[0067] Reference numerals: Controller 1000, Processor 1100, Memory 1200. Detailed Implementation

[0068] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the invention.

[0069] It should be noted that although functional modules are divided in the device schematic diagram and a logical order is shown in the flowchart, in some cases, the steps shown or described may be performed in a different order than the module division in the device or the order in the flowchart. The terms "first," "second," etc., in the specification, claims, or the aforementioned drawings are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence.

[0070] This application provides a method, system, and storage medium for generating photovoltaic module layout schemes, which will be described in detail in the following embodiments.

[0071] like Figure 1 As shown, this embodiment of the invention provides a method for generating a photovoltaic module layout scheme, including steps S100, S200, S300, S400, S500, S600, S700, and S800:

[0072] Step S100: Obtain multi-source data of the target installation area, including building data, historical weather data, human data, and landscape greening data.

[0073] Specifically, the multi-source data in this application can be data that has been acquired in advance and stored on a backend server or in a corresponding storage space, or it can be acquired by accessing a corresponding data storage system or platform, or it can be acquired by detecting the target installation area. The specific acquisition method is not limited here.

[0074] Building data encompasses geometry, structural performance, and thermal properties. Geometry is obtained through 3D laser scanning to capture the 3D contours, slopes, and dimensions of surrounding obstacles of the installation area (roof, facade, etc.), which are then supplemented with internal structural details using the BIM model. Structural performance is assessed by consulting design drawings to obtain load limits; for older buildings, on-site testing of concrete strength and steel weld quality is required. Thermal properties are determined through material sampling and testing or by consulting tables to identify parameters such as the thermal conductivity and solar radiation absorption rate of the external envelope.

[0075] Historical weather data is collected by gathering hourly temperature and humidity, total solar radiation, rainfall and typhoon data over the past 30 years, establishing a climate parameter database, and combining it with a GIS system to create corresponding maps.

[0076] Humanistic data includes population activities, landscape culture, and management standards. By acquiring relevant humanistic data, we can ensure that the installation of photovoltaic modules conforms to the corresponding humanistic characteristics.

[0077] The landscape greening data covers plant attributes, soil topography, and irrigation systems. Plant attributes include species, plant height, canopy width, daily water consumption, and growth cycle; soil topography is obtained through sampling and testing of permeability and pH value, and by drone mapping to obtain slope and aspect; the irrigation system is examined by checking existing methods, pipeline routes, and sprinkler status, and annual water consumption and peak demand are calculated.

[0078] Step S200: Determine the ventilation coefficient of the target installation area based on the building data, the historical weather data, and the landscape greening data.

[0079] Specifically, the three-dimensional outline of the target installation area, the location and size of the ventilation openings, and the height and distance of surrounding obstacles are obtained from the building data. The coordinate accuracy is ensured by using three-dimensional laser scanning and BIM model. Hourly wind speed and direction for the past 5 years are selected from historical weather data. The frequency of the dominant seasonal wind direction, maximum wind speed, and turbulence intensity are statistically analyzed, and abnormal data are eliminated. Plant types, crown width, height, planting density, and relative positions to the target area are obtained from the landscape greening data to determine the drag coefficient of different plants.

[0080] A wind field simulation model was constructed, simplifying the main building structure while retaining ventilation openings. Green areas were designated as porous media with corresponding drag coefficients, and the target area was set as the monitoring surface. Boundary conditions were set: the inlet used a logarithmic wind speed profile to input historical wind speed and direction, the outlet was set to standard atmospheric pressure, and the ground and building surfaces were assigned appropriate roughness based on their materials. Then, ventilation volumes were simulated in different scenarios. In an interference-free scenario, an ideal wind field without buildings or greenery was simulated to calculate the theoretical maximum ventilation volume. In a real-foundation scenario, the effects of building obstruction and greenery drag were simulated to calculate the actual foundation ventilation volume. Based on the actual foundation ventilation volume and the theoretical maximum ventilation volume, the ventilation coefficient was calculated.

[0081] In some optional embodiments, determining the ventilation coefficient of the target installation area based on the building data, the historical weather data, and the landscape greening data includes:

[0082] S210. Construct a first model based on the building data and the landscape greening data;

[0083] Specifically, by combining building data (3D outline of the target area, ventilation opening size, location of surrounding obstacles, etc.) and landscape greening data (plant type, resistance coefficient, planting distribution, etc.), a geometric model is constructed in CFD software. The building is simplified, ventilation openings are retained, the green area is set as a porous medium with a corresponding resistance coefficient, and the target installation area is set as the monitoring surface.

[0084] S220. The first ventilation volume is obtained by performing ventilation simulation on the first model based on the historical weather data.

[0085] Specifically, using historical weather data (hourly wind speed and prevailing wind direction over the past 5 years) as boundary conditions (logarithmic wind speed profile at the inlet and standard atmospheric pressure at the outlet), the first model is used to simulate ventilation through the RNG k-ε turbulence model, and the actual basic ventilation volume under the combined effect of building obstruction and greening resistance is output, which is the first ventilation volume.

[0086] S230. Determine a first coefficient based on the first ventilation volume and the maximum ventilation volume, wherein the maximum ventilation volume represents the ventilation volume when there is no obstruction.

[0087] Specifically, an ideal wind field without buildings or greenery is simulated to obtain the maximum ventilation volume that represents the maximum ventilation potential; the first coefficient is the ratio of the first ventilation volume to the maximum ventilation volume, reflecting the static impact of buildings and greenery on ventilation.

[0088] S240. Determine the seasonal correction factor based on the historical weather data;

[0089] Specifically, the coefficient is determined based on the ratio of the average wind speed of each season to the annual average wind speed in historical weather data. For example, the coefficient is greater than 1 when the summer wind speed is higher than the annual average, and less than 1 when the winter wind speed is lower than the annual average.

[0090] S250. Determine the greening growth correction coefficient based on the aforementioned landscape greening data;

[0091] Specifically, the coefficient is adjusted based on the seasonal growth status of plants in the landscape greening data. For example, the resistance is lower when deciduous trees are leafless in winter, so the coefficient is taken as 1.2 to 1.3; the resistance is higher when the trees are in full leaf in summer, so the coefficient is taken as 1.0.

[0092] S260. The ventilation coefficient is obtained by correcting the first coefficient according to the seasonal correction coefficient and the greening growth correction coefficient.

[0093] Specifically, the first coefficient is multiplied by the seasonal correction coefficient and the greening growth correction coefficient to obtain the ventilation coefficient that reflects the dynamic ventilation potential (the value ranges from 0 to 1, with a larger value indicating stronger ventilation capacity).

[0094] Step S300: Determine the irrigation coefficient of the target installation area based on the historical weather data and the landscape greening data.

[0095] Specifically, historical weather data is used to select monthly rainfall, monthly evaporation, and the distribution of rainy days over the past five years. Landscape greening data is used to obtain the total green area (distinguishing between tree areas, shrub areas, and lawn areas), daily water consumption of various plants, critical water requirement periods for plants (e.g., water demand increases by 30% during summer high temperatures), and the efficiency of existing irrigation facilities. The average monthly greening water requirement is calculated, and the total basic water requirement is obtained. Then, combined with seasonal evaporation differences from historical weather data, a seasonal correction coefficient is introduced (the seasonal correction coefficient equals the ratio between the seasonal average evaporation and the annual average evaporation). Simultaneously, irrigation facility efficiency is considered (the actual water supply must be greater than the plant absorption, which needs to be divided by irrigation efficiency) to finally obtain the average monthly greening water requirement. Next, the irrigation coefficient is calculated; the irrigation coefficient is the ratio of the average monthly greening water requirement to the average monthly available rainfall. If the ratio is ≤1, it indicates that rainfall can meet the greening irrigation needs; if the ratio is >1, it indicates a rainfall deficit, requiring additional water supply.

[0096] In some optional embodiments, determining the irrigation coefficient of the target installation area based on the historical weather data and the landscape greening data includes:

[0097] S310. Obtain the rainfall and evaporation from the historical weather data;

[0098] S320. Obtain the greening area and greening water requirement from the landscape greening data;

[0099] S330. Determine the basic water requirement based on the greening area and the greening water requirement;

[0100] S340. Determine the irrigation water volume based on the rainfall and evaporation.

[0101] S350. Determine the irrigation coefficient based on the irrigation water volume and the basic water requirement.

[0102] Specifically, the monthly rainfall and evaporation of the target area over the past five years are obtained from historical weather data; the total area of ​​green areas (distinguishing between types such as trees, shrubs, and lawns) and the daily water consumption of various types of green plants are obtained from landscape greening data.

[0103] The basic water requirement is the total water consumption required each month to maintain the healthy growth of the green area. It needs to be calculated in combination with the green area and the water requirements of the plants. First, calculate the water requirement of each type of greening area, and then summarize the water requirements of each type to obtain the basic water requirement of the target area each month.

[0104] Irrigation water volume is based on historical weather data and represents the effective rainfall available for greening irrigation. It needs to be corrected for rainfall and evaporation: first, calculate the average monthly rainwater collection potential; then, deduct water loss due to evaporation to obtain the actual amount of water available for irrigation. The irrigation coefficient is the ratio of basic water requirement to irrigation water volume. If the coefficient is ≤1, it indicates that rainwater can meet the greening irrigation needs; if the coefficient is >1, it indicates a rainwater shortage, with a larger coefficient indicating a larger shortage. For example, if the basic water requirement is 150 m³ and the irrigation water volume is 100 m³, then the irrigation coefficient is 1.5, meaning an additional 50% of irrigation water is needed.

[0105] Step S400: Obtain the lighting requirements data and indoor target heat gain rate of the target installation area.

[0106] Specifically, the corresponding lighting requirements are determined based on the function of each building in the target installation area; for example, the minimum lighting coefficient for offices (side lighting) is ≥3%, for bedrooms (side lighting) it is ≥2%, and for garage passageways it is ≥0.5%.

[0107] The target indoor heat gain rate is determined based on building energy efficiency design standards and local climate zones (such as frigid regions and hot-summer-cold-winter regions) to determine the target indoor heat gain rate for different seasons. For example, in hot-summer-cold-winter regions, the target indoor heat gain rate is ≤40% in summer (reducing solar radiation heat gain) and ≥60% in winter (increasing solar radiation heat gain). The target value is determined after calculating the baseline heat gain rate using building energy consumption simulation software (such as EnergyPlus).

[0108] Step S500: Generate a first layout scheme based on the multi-source data, the lighting demand data, the indoor target heat gain rate, the ventilation coefficient, and the irrigation coefficient.

[0109] Specifically, a multi-objective optimization algorithm is used to generate the first layout scheme. Environmental data, ventilation coefficient, irrigation coefficient, lighting requirements data, and indoor target heat gain rate are used as input parameters. Iterative calculations are performed to output a Pareto optimal solution set, and the scheme with the best overall performance is selected as the first layout scheme.

[0110] In some optional embodiments, generating the first layout scheme based on the multi-source data, the lighting requirement data, the indoor target heat gain rate, the ventilation coefficient, and the irrigation coefficient includes:

[0111] S510. Construct a neural network fusion model of multi-objective optimization algorithm and reinforcement learning framework;

[0112] Specifically, a fusion model is built that integrates multi-objective optimization algorithms (such as the improved NSGA-III) with reinforcement learning frameworks (such as DQN). The reinforcement learning framework is used to quickly locate the potential optimal design variable range. The multi-objective optimization algorithm then performs a refined search within the range selected by reinforcement learning, balancing exploration efficiency with solution accuracy. Simultaneously, a fully connected neural network is used to construct the mapping relationship between the objective function and the design variables. The model is pre-trained using historical photovoltaic (PV) configuration case data (performance parameters such as power generation and daylight efficiency) to improve the optimization convergence speed.

[0113] S520. Input the multi-source data, the lighting demand data, the indoor target heat gain rate, the ventilation coefficient, and the irrigation coefficient into the neural network fusion model;

[0114] Specifically, multi-source data (load limits in building data, irradiance / wind speed in historical weather data, landscape requirements in human data, water requirements of plants in landscape greening data, etc.), daylighting demand data (standard values ​​of daylighting coefficients for each space, sunshine duration), indoor target heat gain rate (heat gain rate range by season), ventilation coefficient (ventilation potential value by region), and irrigation coefficient (rainwater supply and demand matching degree) are standardized in the format of spatial ID-time dimension-attribute parameter and then input into the neural network fusion model as the basic input matrix for optimization solution.

[0115] S530. Configure the optimization objective of the neural network fusion model as maximizing photovoltaic power generation, maximizing indoor lighting satisfaction rate, maximizing indoor heat gain rate compliance rate, minimizing ventilation coefficient loss, and maximizing irrigation coefficient matching degree.

[0116] S540. Construct the objective function corresponding to the optimization objective;

[0117] Specifically, the objective function for maximizing photovoltaic power generation is:

[0118]

[0119] in, The total area of ​​the photovoltaic modules. For hourly irradiance, For the power generation efficiency of photovoltaic modules. For hourly ambient temperature, The installation tilt angle of the photovoltaic modules. This is the shading coefficient associated with the spacing of the photovoltaic array.

[0120] The objective function for maximizing indoor lighting satisfaction is:

[0121] in, This represents the actual daylight value of room i. Let m be the standard daylight value for room i, m be the total number of rooms, and sing() be the sign function (1 if the condition is met, 0 if not).

[0122] The objective function for maximizing the indoor heat gain rate compliance rate is:

[0123]

[0124] in, The indoor target minimum heat gain rate for season S. The indoor target maximum heat gain for season S.

[0125] The objective function for minimizing ventilation coefficient loss is:

[0126]

[0127] in, The ventilation coefficient after photovoltaic module installation. This refers to the ventilation coefficient before the photovoltaic modules are installed.

[0128] The objective function that maximizes the matching degree of the irrigation coefficient is:

[0129]

[0130] in, For the water collection capacity of photovoltaic modules, Water requirements for greening.

[0131] S550. Configure the design variables of the neural network fusion model as photovoltaic module installation tilt angle, photovoltaic array spacing, photovoltaic module installation position, photovoltaic module orientation, photovoltaic module transmittance, and water collection tank coverage.

[0132] Specifically, the installation tilt angle of photovoltaic modules, the spacing between photovoltaic arrays, the installation position of photovoltaic modules, the orientation of photovoltaic modules, the light transmittance of photovoltaic modules, and the coverage of the water collection tank are set as adjustable parameters for model iterative optimization.

[0133] S560. Configure the constraints of the neural network fusion model as structural load constraints, solar spacing constraints, human landscape constraints, and greening coordination constraints.

[0134] Specifically, the structural load constraint is: the total weight of photovoltaic modules and supports ≤ the building load limit;

[0135] The sunlight spacing constraint is: the length of the shadow cast by the photovoltaic module at noon on the winter solstice is ≤ one-third of the spacing between surrounding buildings;

[0136] The constraints for the cultural landscape are: the light transmittance of the photovoltaic modules on the facade must be ≥60% (around historical buildings), and the color of the photovoltaic modules must match the main color scheme of the building;

[0137] The greening coordination constraint is: the height of the support frame ≥ the height of the plant + 0.5m, to avoid obstructing plant growth.

[0138] S570. Based on the constraints, the objective function, and the optimization objective, the Pareto optimal solution set is obtained by iterating the design variables.

[0139] Specifically, the model initializes a population of design variables (size 100, for example only) based on constraints, explores variable combinations through reinforcement learning (prioritizing variable intervals with good performance in historical interactions), calculates the objective function value of each group of variables using a multi-objective optimization algorithm, and performs non-dominated sorting (screening non-dominated solutions where no single objective is inferior to other solutions); after 1000 iterations (the specific number can be set according to requirements and is not limited here), it outputs a Pareto optimal solution set (containing 20-30 solutions).

[0140] S580. Based on the comprehensive benefit requirements, the first layout scheme is obtained by selecting the target solution from the Pareto optimal solution set.

[0141] Specifically, based on the comprehensive benefit requirements (e.g., if power generation is given priority, the solution with the top 20% power generation is selected; if ecological synergy is given priority, the solution with both excellent ventilation and irrigation matching is selected), one optimal solution is selected as the first layout solution.

[0142] In some optional embodiments, obtaining the Pareto optimal solution set by iterating the design variables based on the constraints, the objective function, and the optimization objective includes:

[0143] S571. Randomly generate a preset number of initial design variable combinations, wherein the initial design variable combinations satisfy the constraint conditions;

[0144] Specifically, based on a preset number (100-200 sets, balancing computational efficiency and solution diversity), combinations of design variables (covering all adjustable parameters such as photovoltaic module tilt angle, array spacing, and installation location) are randomly generated. During generation, it is necessary to verify in real time whether constraints are met: for example, the total weight of the modules and supports must be ≤ the building load limit (structural load constraint), the shadow length of the modules on the winter solstice must be ≤ one-third of the distance between surrounding buildings (sunlight spacing constraint), and the light transmittance of the facade modules must meet the requirements of the human landscape. Combinations that do not meet the constraints are directly eliminated, and regeneration is performed until a preset number of valid initial design variable combinations are obtained, forming the initial population.

[0145] S572. Substitute the initial design variables into the objective function to calculate the fitness value of the optimization objective, wherein the fitness value represents the degree of matching with the optimization objective;

[0146] Specifically, each initial design variable combination is substituted into the objective function corresponding to the five optimization objectives, and the fitness value of each objective is calculated. The fitness value is used to quantify the degree of matching between the variable combination and the optimization objective. For example, in the photovoltaic power generation objective, the fitness value is equal to the ratio between the actual calculated power generation and the theoretical maximum power generation (the theoretical maximum power generation is calculated based on the conditions of no shading and optimal tilt angle). The closer the ratio is to 1, the higher the fitness value. In the ventilation coefficient loss objective, the fitness value = 1 - actual ventilation loss rate. The calculation logic of the fitness value of the light satisfaction rate, heat gain rate compliance rate, and irrigation matching degree objective is similar, and finally a five-dimensional fitness value vector (such as [0.92, 0.88, 0.95, 0.91, 0.86]) is formed for each variable combination.

[0147] S573. After performing a non-dominated sorting of the initial design variable combination based on the fitness values, a non-dominated solution is obtained.

[0148] Specifically, a non-dominated sorting method (such as the sorting logic in the NSGA-II algorithm) is used to classify the combinations of design variables in the initial population: if all fitness values ​​of a certain variable combination are not lower than another combination, and at least one fitness value is higher than that combination, then the former dominates the latter, and the latter is eliminated; all variable combinations not dominated by other combinations are retained, which are the non-dominated solutions (first frontier solutions). At the same time, the crowding distance (quantifying the degree of dispersion of the solution in the target space) is calculated for the non-dominated solutions to ensure the diversity of solutions. The larger the crowding distance, the more unique the solution is in the target space, which can avoid solution set clustering.

[0149] S574. Gene recombination and random mutation are performed on the non-dominated solution to generate a new combination of iterative design variables, and the iteration number is incremented by one.

[0150] Specifically, gene recombination uses a two-point crossover method. Two sets of non-dominated solutions (parents) are randomly selected, and two crossover points are chosen in the design variable sequence. The variable segments between the crossover points are exchanged (e.g., the "tilt angle 35°, spacing 2.5H" of parent 1 is crossed with the "tilt angle 40°, spacing 2.0H" of parent 2 to generate the "tilt angle 35°, spacing 2.0H" of the offspring). After recombination, it is necessary to check whether the offspring meets the constraints. If not, crossover is performed again.

[0151] Random mutation involves randomly adjusting a design variable in the recombined offspring variables according to a preset mutation probability (5%–10%) (e.g., mutating "transmittance 40%" to "transmittance 45%", or "installation position R1" to "installation position R2"). Constraint satisfaction must also be verified to ensure the offspring are valid. Repeating the above steps generates a new iterative combination of design variables with the same initial population size, and increments the iteration count by 1 (the initial iteration count is 0).

[0152] S575. When the number of iterations is less than or equal to a preset number, the iterative design variable combination is configured as the initial design variable combination and then iterative calculation is performed. The iterative calculation represents the above-mentioned step of generating the iterative design variable combination.

[0153] Specifically, compare the current number of iterations with the preset number (1000 to 2000 times, which needs to be set according to computing resources and accuracy requirements, and is not limited here); if the number of iterations is less than or equal to the preset number, the newly generated combination of iterative design variables is used as the "initial combination of design variables", and the process returns to steps S572-S574 to repeat the iterative calculation, continuously optimizing the quality and diversity of non-dominated solutions.

[0154] S576. If the number of iterations is greater than a preset number, the non-dominated solution is configured as the Pareto optimal solution set.

[0155] Specifically, if the number of iterations is greater than the preset number, the iteration is stopped, and the final non-dominated solutions (after multiple rounds of screening, 20-30 sets are retained) are formally configured as the Pareto optimal solution set. This solution set covers the optimal trade-offs among multiple objectives (such as a solution with the highest power generation but a moderate light-meeting rate, or a solution with excellent matching degree between light and irrigation but slightly lower power generation).

[0156] Step S600: Construct a digital twin model of the first layout scheme and the multi-source data collaboration.

[0157] Specifically, the multi-source data is fused with the first layout scheme to obtain a fused BIM model. The fused BIM model is then imported into the first layout scheme, and buildings, photovoltaic modules, green areas, and human activity paths are recreated according to a preset scale. Multi-physics simulation result visualization layers (such as wind speed vector arrows, daylight coefficient heat maps, and rainwater harvesting path animations) are overlaid to obtain a corresponding digital twin model. The digital twin model supports scene scaling, panning, and rotation. Clicking on any component allows viewing its real-time parameters such as power generation and temperature, as well as associated multi-source data (such as the water demand of the green area below the component). It provides historical weather replay (such as recreating the scheme's operational status on extreme high-temperature days) and parameter adjustment previews (dragging sliders to modify component tilt angles, with the platform displaying real-time trends in power generation and ventilation coefficients). Performance charts are generated, displaying real-time comparisons between current and target values ​​for power generation, daylight satisfaction rate, heat gain rate, ventilation coefficient, and irrigation matching degree. Simulation reports are automatically generated (including key parameter cloud maps, time series curves, and constraint satisfaction descriptions).

[0158] In some optional embodiments, constructing the first layout scheme and the digital twin model of multi-source data collaboration includes:

[0159] S610. Convert the building data into a BIM model;

[0160] Specifically, building data is extracted from multi-source data and a model is built using BIM software. The main structure, such as the roof, facade, and floor slabs, is restored according to the actual building layout. The location and parameters of key components such as load-bearing beams, ventilation openings, doors, and windows are marked. At the same time, the structural performance parameters in the building data are associated with the corresponding BIM components.

[0161] S620. The first layout scheme is superimposed onto the BIM model to obtain the first model;

[0162] Specifically, the core parameters (installation area, quantity, tilt angle, spacing, light transmittance, and water collection tank coverage of photovoltaic modules) are extracted from the first layout scheme. A photovoltaic module family (including module size, weight, electrical parameters, and other attributes) is created in the BIM software. According to the location coordinates marked in the scheme, the module family is accurately superimposed onto the corresponding area of ​​the BIM model (such as the southeast area of ​​the roof or the south side of the facade). During the superposition process, spatial collision detection is performed to ensure that there is no positional conflict between the modules and building components, and that the total weight of the modules and the support load does not exceed the building load limit. At the same time, the performance parameters of the modules (such as the power of a single module and the temperature coefficient) are bound to the module family in the BIM model to form the first model that includes the building structure and the arrangement of photovoltaic modules.

[0163] S630. Associate the historical weather data with the first model to obtain the second model;

[0164] Specifically, historical weather data is filtered from multi-source data and associated with the first model according to the logic of timestamp and spatial location: in the BIM model, meteorological parameters corresponding to the location are bound to different areas (such as the hourly wind speed at a height of 15m for the roof area and the solar irradiance at a height of 8m for the facade area). Through the API interface, historical weather data is embedded into the components of the first model in the form of attribute parameters. At the same time, data association rules are set (such as automatically calling the component efficiency degradation coefficient at the corresponding temperature during the high temperature period in summer) to form a second model that integrates building, photovoltaic and meteorological data.

[0165] S640. The human data and the landscape greening data are superimposed on the second model to obtain the third model;

[0166] Specifically, landscape requirements (such as building facade component colors matching the main color scheme), activity paths (such as roof maintenance access width ≥ 1.2m), and management regulations (such as component spacing ≥ 0.5m from fire lanes) are extracted from the humanistic data. Constraint areas are marked in the BIM model using annotation layers (such as marking maintenance access with red lines to prevent component coverage), and humanistic constraint parameters (such as component color RGB value range) are associated with the corresponding component families. Plant types, heights, crown widths, planting locations, and soil parameters are extracted from the landscape greening data. Green plant families (such as tree families and shrub families) are created in the BIM model and overlaid onto the green areas of the model according to the actual planting distribution. Simultaneously, parameters such as plant water requirements and resistance coefficients are embedded into the plant families, forming a third model encompassing architecture, photovoltaics, meteorology, humanistic elements, and greening.

[0167] S650. Based on the first layout scheme, the historical weather data, the landscape greening data, the human data, and the building data, generate a first coupled model that can calculate the hourly power generation, indoor heat gain rate, ventilation volume, daylighting rate, and water collection volume at each location, and then superimpose the first coupled model onto the third model to obtain the digital twin model.

[0168] Specifically, based on the first layout scheme, historical weather data, landscape greening data, human data, and building data, the first coupling model is developed using software such as EnergyPlus, Fluent, and Radiance to realize hourly calculation of key performance.

[0169] Hourly power generation calculation: Combining the module tilt angle, light transmittance (from the first layout scheme) with hourly irradiance and temperature (from historical weather data), the PVsyst algorithm is used to calculate the hourly power generation of each module.

[0170] Indoor heat gain rate calculation: Based on the thermal parameters of the building envelope (from building data), component shading rate (from the first layout scheme) and hourly solar radiation (from historical weather data), the hourly indoor heat gain rate is calculated using EnergyPlus and matched with the target indoor heat gain rate.

[0171] Ventilation volume calculation: Combining the location of building ventilation openings (from building data), component array spacing (from the first layout scheme), plant resistance coefficient (from landscape greening data), and hourly wind speed (from historical weather data), Fluent's RNG k-ε model is used to calculate the hourly indoor and outdoor ventilation volume.

[0172] Daylight efficiency calculation: Based on the component transmittance (from the first layout scheme), window position (from building data), and solar altitude angle (from historical weather data), the indoor hourly daylight efficiency is calculated using the Radiance ray tracing algorithm to match the daylighting demand data.

[0173] Water collection calculation: Combining the component's water collection area, water collection trough coverage (from the first layout scheme), and hourly rainfall (from historical weather data), the hourly water collection volume of each water collection trough is calculated, and correlated with the greening water demand (from landscape greening data). Finally, through a data interface, the performance calculation function of the first coupled model is linked with the geometric space of the third model, allowing users to view the hourly power generation, heat gain rate, ventilation volume, daylighting rate, and water collection volume of any location on the model, forming a digital twin model that combines geometric visualization and performance simulation functions.

[0174] Step S700: Simulate the digital twin model to obtain simulation results.

[0175] Specifically, the digital twin model is simulated, and the model is driven by multi-source data, focusing on the core performance objectives of the first layout scheme. The simulation results, which combine quantification and scenario-based analysis, are output through multi-dimensional calculations.

[0176] In terms of power generation, hourly irradiance and ambient temperature are used as inputs. A solar incidence correction coefficient is calculated based on the tilt angle of the modules. Module efficiency is corrected according to the temperature coefficient, and the hourly power generation of a single module, zone, and the entire system is output, accumulating to obtain the total daily / weekly / seasonal power generation. In terms of daylighting and heat gain, the impact of module shading on indoor illuminance is simulated using a ray tracing algorithm. The hourly daylighting rate and daylighting satisfaction rate of each room are calculated. Simultaneously, based on the thermal calculation module, building envelope parameters and module transmittance are input, and the hourly heat gain rate is output to verify whether it meets the target indoor heat gain rate range. In terms of ventilation, hourly wind speed and direction are used as boundary conditions. Combined with building outline, module spacing, and plant resistance coefficient, the wind field distribution is simulated, and the hourly indoor and outdoor ventilation volume is calculated, outputting the actual ventilation coefficient and ventilation loss rate. In terms of irrigation, the hourly rainwater collection volume is calculated based on hourly rainfall, module water collection area, and water collection trough coverage, and evaporation loss is deducted. Combined with plant water requirements, the irrigation matching degree is obtained, and the irrigation satisfaction rate and rainwater deficit are statistically analyzed.

[0177] Finally, the simulation results are output, including quantitative reports and visualizations. The quantitative reports cover core data such as power generation (generation, efficiency, and loss rate of each zone), daylighting and heat gain (daylighting rate and heat gain rate of each room), ventilation (ventilation volume, ventilation coefficient, and loss rate), and irrigation (water collection volume, matching degree, and deficit). The visualizations overlay the spatial distribution of performance parameters (such as power generation heat map and daylighting cloud map), time series curves (hourly changes in power generation and heat gain rate), and dynamic scene simulations (wind field streamline animation and rainwater collection path) into the BIM scene, intuitively presenting the performance of the solution under different scenarios.

[0178] In some optional embodiments, after obtaining the simulation results by simulating the digital twin model, the process includes:

[0179] S710. Generate a simulation map based on the simulation results. The simulation map is used to indicate the difference between the multi-dimensional simulation data and the target data at each location of the target installation area. The multi-dimensional simulation data characterizes the simulated hourly power generation, indoor heat gain rate, daylighting rate, ventilation volume, and water collection volume.

[0180] Specifically, based on the multi-dimensional simulation data (hourly power generation, indoor heat gain rate, daylighting rate, ventilation volume, and water collection rate) output by the digital twin model and the preset target data (such as the target value of power generation, the standard range of indoor heat gain rate, and the standard value of daylighting rate), a simulation map is generated according to the correspondence between spatial location and performance difference.

[0181] First, determine the spatial coordinate system of the map (consistent with the BIM coordinate system of the digital twin model), and divide the target installation area into 1m×1m grid cells, with each cell corresponding to a spatial location; for each grid cell, calculate the difference between the multi-dimensional simulation data and the target data (e.g., if a cell simulates a power generation of 200W and the target power generation is 250W, the difference is -50W; if a cell simulates a daylighting rate of 2.8% and the target daylighting rate is 3%, the difference is -0.2%).

[0182] The difference is visualized using a color coding rule: for example, red indicates that the difference exceeds the standard (such as power generation difference < -50W, light transmittance difference < -0.3%), yellow indicates that the difference is close to the critical value (such as power generation difference -50W to -20W, light transmittance difference -0.3% to -0.1%), and green indicates that the difference meets the standard (such as power generation difference ≥ -20W, light transmittance difference ≥ -0.1%). Finally, a simulation map covering the entire area and containing multi-dimensional performance difference information is generated.

[0183] S720. After performing grayscale processing, filtering processing and noise reduction processing on the simulated map in sequence, a preprocessed image is obtained.

[0184] Specifically, the color simulation spectrum is converted into a grayscale image, and the performance difference is quantified by grayscale value (e.g., grayscale value 0 corresponds to the maximum negative difference, and grayscale value 255 corresponds to the maximum positive difference, establishing a mapping relationship between the difference and grayscale value), eliminating the interference of color information on subsequent processing, and simplifying the data dimensions.

[0185] The filtering process uses a Gaussian filtering algorithm to smooth the grayscale image, removing subtle noise generated during the image generation process due to grid division and data calculation (such as grayscale value jumps caused by calculation errors in adjacent grid cells), making the image grayscale changes smoother and preserving the overall performance difference distribution trend.

[0186] The denoising process further eliminates isolated noise points (such as individual grid cells with abnormal gray values) remaining after filtering by using a median filtering algorithm, ensuring that the changes in gray values ​​in the image are only caused by differences in actual performance, and finally obtaining a preprocessed image with uniform gray distribution and low noise interference.

[0187] S730. Divide the preprocessed image into multiple spatial image sub-regions;

[0188] Specifically, a region growing algorithm is used to spatially segment the preprocessed image to obtain multiple spatial image sub-regions: first, grid cells with similar gray values ​​in the image are selected as initial seed points (e.g., cells with gray values ​​of 180~200 are used as seed points for areas with compliant power generation, and cells with gray values ​​of 50~70 are used as seed points for areas with abnormal power generation); with the seed points as the center, adjacent grid cells with a gray value difference ≤5 (a threshold set according to the gray value range of the image) are gradually merged to form continuous spatial regions.

[0189] Repeat the above process until all grid cells in the image are assigned to their corresponding regions. Finally, the preprocessed image is divided into multiple spatial image sub-regions with clear boundaries and high similarity of internal gray values ​​(such as sub-regions with qualified power generation and sub-regions with abnormal light transmittance).

[0190] S740. Obtain abnormal region information based on the similarity between multiple spatial image sub-regions. The abnormal region information represents the contour information and spatial location information of the target region that is significantly different from the adjacent regions.

[0191] Specifically, the similarity of gray values ​​between each spatial image sub-region and its adjacent sub-regions is calculated (using the structural similarity index SSIM, with a value range of 0 to 1, and the closer the value is to 1, the higher the similarity), and abnormal regions are identified and information is extracted.

[0192] In some optional embodiments, obtaining abnormal region information based on the similarity between multiple spatial image sub-regions includes:

[0193] S741. A comparison result is obtained by comparing the first similarity between the first image sub-region and the third image sub-region and the second similarity between the second image sub-region and the third image sub-region. The third image sub-region is any of the spatial image sub-regions. The first image sub-region is located on the first side of the third image sub-region, and the second image sub-region is located on the second side of the third image sub-region. The first image sub-region, the second image sub-region, and the third image sub-region all represent the spatial image sub-region.

[0194] Specifically, any sub-region is selected from the segmented spatial image sub-regions as the third image sub-region (the core region to be determined as abnormal). Simultaneously, its two adjacent sub-regions are identified: the first image sub-region located on the first side of the third image sub-region (e.g., the left side), and the second image sub-region located on the second side (e.g., the right side). The structural similarity index (SSIM) is used to calculate the first similarity between the third image sub-region and the first image sub-region, and the second similarity between the third image sub-region and the second image sub-region. The closer the SSIM value is to 1, the more similar the grayscale distributions (corresponding to performance difference distributions) of the two sub-regions are; the closer the value is to 0, the more significant the difference.

[0195] S742. If the comparison result indicates that the first similarity is greater than the second similarity, the third image sub-region is adjusted by unit distance and scaled by unit scale sequentially towards the location of the second image sub-region, and the contour information and spatial information of the third image sub-region with the largest difference between the first similarity and the second similarity are configured as the abnormal region information.

[0196] Specifically, the first similarity is compared with the second similarity, and the third image sub-region is dynamically adjusted based on the comparison result.

[0197] If the comparison results show that the first similarity is greater than the second similarity (indicating that the third sub-region is more similar to the first sub-region on the left and more different from the second sub-region on the right): The third image sub-region is then adjusted sequentially towards the side (right) where the second image sub-region is located, using unit distance adjustments (e.g., moving 1 grid cell to the right horizontally each time, corresponding to 0.5m in actual space) and unit scale scaling (e.g., shrinking by 5% along the axis perpendicular to the direction of movement each time, maintaining the integrity of the sub-region). After each adjustment, the similarity between the current third sub-region and the first and second sub-regions is recalculated. During the dynamic adjustment of the third image sub-region, the difference between the first and second similarities is continuously recorded after each adjustment. When the difference reaches its maximum value (indicating that the third sub-region is extremely similar to one side of the sub-region and the difference from the other side reaches its peak, meeting the characteristic of "significant differences between the abnormal region and adjacent regions"), the adjustment is stopped. The contour information (such as the scaled polygon boundary) and spatial location information (such as the coordinate range based on the digital twin model coordinate system) of the third image sub-region at this time are extracted and configured as abnormal region information. This abnormal region information accurately reflects the range of the region with the most significant performance difference from the surrounding area, which can be directly mapped to the actual location of the target installation area, facilitating subsequent rapid analysis of photovoltaic layout issues.

[0198] S743. If the comparison result indicates that the first similarity is less than the second similarity, the third image sub-region is adjusted by unit distance and scaled by unit scale sequentially towards the location of the first image sub-region, and the contour information and spatial information of the third image sub-region with the largest difference between the first similarity and the second similarity are configured as the abnormal region information.

[0199] Specifically, if the comparison result shows that the first similarity is less than the second similarity (indicating that the third sub-region is more similar to the second sub-region on the right and more different from the first sub-region on the left): the third image sub-region is sequentially adjusted by unit distance and scaled by unit scale towards the side (left) where the first image sub-region is located. Similarly, the two sets of similarities are recalculated each time the adjustment is performed. When the difference between the first and second similarities is the largest, the contour information and spatial information of the third image sub-region are configured as the abnormal region information.

[0200] S744. If the comparison result indicates that the first similarity is equal to the second similarity, simultaneously move the first image sub-region and the second image sub-region by a unit distance and scale until the first similarity is not equal to the second similarity.

[0201] Specifically, if the comparison results show that the first similarity is equal to the second similarity (indicating that the third sub-region is balanced with the two sub-regions and it is not possible to judge an anomaly at this time): simultaneously perform unit distance adjustment (e.g., move both towards the third sub-region by 1 grid cell) and unit scale scaling (e.g., shrink both by 5%) on the first image sub-region and the second image sub-region. After each adjustment, recalculate the similarity between the third sub-region and the two, until the two sets of similarities are no longer equal, and then perform subsequent adjustments according to the above two cases.

[0202] In some optional embodiments, after obtaining abnormal region information based on the similarity between multiple spatial image sub-regions, the method further includes:

[0203] S745. Obtain the target data corresponding to the spatial location indicated by the abnormal area information, wherein the target data is simulated data;

[0204] Specifically, based on the identified anomaly location information, the simulated data corresponding to that spatial location, i.e., the target data, is extracted from the simulated map.

[0205] S746. Determine the target anomaly type of the target area based on the target data and preset analysis rules;

[0206] Specifically, the preset analysis rules are pre-defined rules for the correspondence between parameter thresholds and anomaly types. By analyzing the characteristics of the target data through the preset analysis rules, the corresponding anomaly type can be determined: the extracted target data is compared one by one with the parameter thresholds in the preset analysis rules to determine the parameters that exceed the normal range and their corresponding anomaly types; if multiple parameters in the target data exceed the standard, the main anomaly type is determined according to the priority of impact, and finally the target anomaly type of the target area is obtained.

[0207] S747. If the target anomaly type does not belong to the anomaly type table, the target area is manually reviewed and analyzed to obtain the reviewed anomaly type.

[0208] Specifically, the anomaly type table is a standardized set of anomaly types formed from previous practices and historical data accumulation. The target anomaly type is matched against the anomaly type table. If the target anomaly type is not in the anomaly type table, it indicates that the anomaly may be a new type of anomaly caused by a special installation environment or installation location, and its accuracy needs to be verified through review, analysis, and processing.

[0209] S748. If the reviewed anomaly type is equal to the target anomaly type, add the target anomaly type to the anomaly type table;

[0210] Specifically, if the re-examined anomaly type is consistent with the previously determined target anomaly type, it indicates that the target anomaly type was accurately determined. The target anomaly type is then added to the anomaly type table, and its corresponding analysis rules are improved.

[0211] S749. If the reviewed anomaly type is not equal to the target anomaly type, replace the target anomaly type with the reviewed anomaly type and add the reviewed anomaly type to the anomaly type table.

[0212] Specifically, if the anomaly type being reviewed is inconsistent with the target anomaly type, it indicates that there is a model prediction bias in the previous target data. The target anomaly type will be replaced with the anomaly type being reviewed, and the anomaly type being reviewed will be added to the anomaly type table.

[0213] S800. Based on the simulation results, the first layout scheme is corrected to obtain the target layout scheme.

[0214] Specifically, based on the information (outline and spatial location) of abnormal areas in the simulation results, the corresponding physical areas (such as the northwest area of ​​the roof and the second floor on the south side of the facade) are located in the digital twin model, and the multi-dimensional performance differences of these areas in the simulation results are matched. For different performance issues, parameter adjustment strategies are formulated based on constraints in multi-source data (such as building load and human landscape requirements). The corrected parameters (new tilt angle, spacing, light transmittance, etc.) are updated to the first layout scheme, generating a corrected scheme and importing it into the digital twin model for a second simulation. If the simulation results meet the layout requirements, the corrected scheme can be used as the target layout scheme for the installation of photovoltaic modules.

[0215] In some optional embodiments, after correcting the first layout scheme based on the simulation results to obtain the target layout scheme, the method further includes:

[0216] S810. Determine the first priority of each layout sub-region based on the simulation data corresponding to each layout sub-region of the target layout scheme.

[0217] Specifically, taking the layout sub-areas (such as residential sub-areas, commercial sub-areas, municipal road sub-areas, etc.) divided by the target layout scheme as units, the simulation data corresponding to each layout sub-area is extracted, priority evaluation indicators and weights are set, and the first priority score of each layout sub-area is calculated. The higher the score, the higher the first priority.

[0218] S820. Determine the correlation between each arranged sub-region through network analysis;

[0219] Specifically, network analysis is used to identify the interrelationships among the various sub-regions in terms of construction resources, process connections, and environmental impact, and an association matrix is ​​constructed to quantify the association strength.

[0220] S830. After correcting the first priority based on the correlation, the second priority is obtained so that the photovoltaic modules in each layout sub-area are installed in sequence according to the second priority.

[0221] Specifically, the correlation weight is used as a correction coefficient to adjust the score of the first priority, resulting in a second priority. This ensures that the installation sequence of photovoltaic modules takes into account both the characteristics of the sub-regions and the overall correlation. Photovoltaic modules are installed according to the corrected second priority to guarantee the optimization of resources, efficiency, and economic benefits.

[0222] The beneficial effects of this invention include: acquiring multi-source data of the target installation area, including building data, historical weather data, human data, and landscape greening data; determining the ventilation coefficient of the target installation area based on the building data, historical weather data, and landscape greening data; determining the irrigation coefficient of the target installation area based on the historical weather data and landscape greening data; acquiring the light requirement data and indoor target heat gain rate of the target installation area; generating a first layout scheme based on the multi-source data, the light requirement data, the indoor target heat gain rate, the ventilation coefficient, and the irrigation coefficient; constructing a digital twin model of the first layout scheme and the multi-source data; simulating the digital twin model to obtain simulation results; and correcting the first layout scheme based on the simulation results to obtain the target layout scheme. The invention automatically generates the corresponding first layout scheme based on the multi-source data of the target installation area, resulting in high generation efficiency. By combining the first layout scheme and the multi-source data to generate a corresponding digital twin model and performing simulation, the first layout scheme is corrected to obtain the target layout scheme, thus improving the completeness of the photovoltaic module layout scheme.

[0223] like Figure 2 As shown, Figure 2 A structural block diagram of a controller 1000 according to an embodiment of this application is shown. The components of the controller 1000 include, but are not limited to, a memory 1200 and a processor 1100. The processor 1100 is connected to the memory 1200 via a bus, and the memory 1200 is used to store data.

[0224] The controller 1000 also includes an access device that enables the controller 1000 to communicate via one or more networks. Examples of such networks include a Public Switched Telephone Network (PSTN), a Local Area Network (LAN), a Wide Area Network (WAN), a Personal Area Network (PAN), or a combination of communication networks such as the Internet. The access device may include one or more of any type of wired or wireless network interface (e.g., a Network Interface Card (NIC)), such as an IEEE 802.11 Wireless Local Area Network (WLAN) wireless interface, a Global System for Microwave Access (GSM) interface, or a Wi-Fi interface. Interfaces include MAX, Ethernet, USB, cellular, Bluetooth, NFC, and more.

[0225] The controller 1000 can be any type of stationary or mobile electronic device, including mobile computers or mobile electronic devices (e.g., tablet computers, personal digital assistants, laptop computers, notebook computers, netbooks, etc.), mobile phones (e.g., smartphones), wearable electronic devices (e.g., smartwatches, smart glasses, etc.) or other types of mobile devices, or stationary electronic devices such as desktop computers or PCs. The controller 1000 can also be a mobile or stationary server.

[0226] The processor 1100 is used to execute computer-executable instructions for generating a photovoltaic module layout scheme.

[0227] The above is an illustrative scheme of a controller according to this embodiment. It should be noted that the technical solution of this controller belongs to the same concept as the technical solution of the photovoltaic module layout scheme generation method described above. For details not described in detail in the technical solution of the controller, please refer to the description of the technical solution of the photovoltaic module layout scheme generation method described above.

[0228] According to an embodiment of this application, a photovoltaic module layout scheme generation system is also provided. The photovoltaic module layout scheme generation system includes a controller 1000, which automatically generates the construction scheme. It should be noted that the technical solution of this photovoltaic module layout scheme generation system belongs to the same concept as the technical solution of the aforementioned photovoltaic module layout scheme generation method. Details not described in detail in the technical solution of the computing device can be found in the description of the technical solution of the aforementioned photovoltaic module layout scheme generation method.

[0229] This application embodiment also provides a storage medium, which is a computer-readable storage medium, storing a computer program that, when executed by a processor, implements the above-described photovoltaic module layout scheme generation method.

[0230] Memory, as a non-transitory computer-readable storage medium, can be used to store non-transitory software programs and non-transitory computer-executable programs. Furthermore, memory may include high-speed random access memory, and may also include non-transitory memory, such as at least one disk storage device, flash memory device, or other non-transitory solid-state storage device. In some embodiments, memory may include memory remotely located relative to the processor, and these remote memories can be connected to the processor via a network. Examples of such networks include, but are not limited to, the Internet, intranets, local area networks, mobile communication networks, and combinations thereof. The device embodiments described above are merely illustrative, and the units described as separate components may or may not be physically separate, and may be located in one place or distributed across multiple network units. Some or all of the modules can be selected to achieve the purpose of this embodiment according to actual needs.

[0231] It will be understood by those skilled in the art that all or some of the steps and systems in the methods disclosed above can be implemented as software, firmware, hardware, and suitable combinations thereof. Some or all of the physical components can be implemented as processors, such as central processing units, digital signal processors, or microprocessors executing software, or as hardware, or as integrated circuits, such as application-specific integrated circuits. Such software can be distributed on a computer-readable medium, which can include computer storage media (or non-transitory media) and communication media (or transient media). As is known to those skilled in the art, the term computer storage media includes volatile and non-volatile, removable and non-removable media implemented in any method or technology for storing information (such as computer-readable instructions, data structures, program modules, or other data). Computer storage media includes, but is not limited to, RAM, ROM, EEPROM, flash memory or other memory technologies, CD-ROM, digital versatile disc (DVD) or other optical disc storage, magnetic cartridges, magnetic tape, disk storage or other magnetic storage devices, or any other medium that can be used to store desired information and is accessible to a computer. Furthermore, as is known to those skilled in the art, communication media typically include computer-readable instructions, data structures, program modules, or other data in modulated data signals such as carrier waves or other transmission mechanisms, and may include any information delivery medium.

[0232] The above provides a detailed description of the preferred embodiments of this application. However, this application is not limited to the above-described embodiments. Those skilled in the art can make various equivalent modifications or substitutions without departing from the spirit of this application. All such equivalent modifications or substitutions are included within the scope defined by the claims of this application.

Claims

1. A method for generating a photovoltaic module layout scheme, characterized in that, include: Acquire multi-source data of the target installation area, including building data, historical weather data, human data, and landscape greening data; The ventilation coefficient of the target installation area is determined based on the building data, historical weather data, and landscape greening data. Specifically, this includes: constructing a first model based on the building data and landscape greening data; performing ventilation simulation on the first model based on the historical weather data to obtain a first ventilation volume; determining a first coefficient based on the first ventilation volume and a maximum ventilation volume, where the maximum ventilation volume represents the ventilation volume without obstruction; determining a seasonal correction coefficient based on the historical weather data; determining a greening growth correction coefficient based on the landscape greening data; and correcting the first coefficient based on the seasonal correction coefficient and the greening growth correction coefficient to obtain the ventilation coefficient. The irrigation coefficient for the target installation area is determined based on the historical weather data and the landscape greening data. Obtain the lighting requirements data and indoor target heat gain rate of the target installation area; A first layout scheme is generated based on the multi-source data, the lighting demand data, the indoor target heat gain rate, the ventilation coefficient, and the irrigation coefficient; Construct a digital twin model of the first layout scheme and the multi-source data collaboration; The simulation results were obtained after simulating the digital twin model. The target layout scheme is obtained by correcting the first layout scheme based on the simulation results.

2. The method for generating a photovoltaic module layout scheme according to claim 1, characterized in that, The step of determining the irrigation coefficient of the target installation area based on the historical weather data and the landscape greening data includes: Obtain the rainfall and evaporation from the historical weather data; Obtain the green area and water requirement for greening from the landscape greening data; The basic water requirement is determined based on the green area and the water demand of the green area; The amount of irrigation water is determined based on the rainfall and evaporation. The irrigation coefficient is determined based on the irrigation water volume and the basic water requirement.

3. The method for generating a photovoltaic module layout scheme according to claim 1, characterized in that, The step of generating a first layout scheme based on the multi-source data, the lighting demand data, the indoor target heat gain rate, the ventilation coefficient, and the irrigation coefficient includes: Construct a neural network fusion model combining multi-objective optimization algorithms and reinforcement learning frameworks; The multi-source data, the lighting demand data, the indoor target heat gain rate, the ventilation coefficient, and the irrigation coefficient are input into the neural network fusion model; The optimization objectives of the neural network fusion model are configured as maximizing photovoltaic power generation, maximizing indoor lighting satisfaction rate, maximizing indoor heat gain rate compliance rate, minimizing ventilation coefficient loss, and maximizing irrigation coefficient matching degree. Construct the objective function corresponding to the optimization objective; The design variables of the neural network fusion model are configured as photovoltaic module installation tilt angle, photovoltaic array spacing, photovoltaic module installation position, photovoltaic module orientation, photovoltaic module transmittance, and water collection tank coverage. The constraints of the neural network fusion model are configured as structural load constraints, solar spacing constraints, human landscape constraints, and greening synergy constraints. Based on the constraints, the objective function, and the optimization objective, the Pareto optimal solution set is obtained by iterating the design variables. The first layout scheme is obtained by selecting the target solution from the Pareto optimal solution set based on the overall benefit requirements.

4. The method for generating a photovoltaic module layout scheme according to claim 3, characterized in that, The process of obtaining the Pareto optimal solution set by iterating the design variables based on the constraints, the objective function, and the optimization objective includes: A preset number of initial design variable combinations are randomly generated, and the initial design variable combinations satisfy the constraints. The initial design variables are substituted into the objective function to calculate the fitness value of the optimization objective, and the fitness value represents the degree of matching with the optimization objective. The non-dominated solution is obtained by performing a non-dominated sorting of the initial design variable combination based on the fitness value. After performing gene recombination and random mutation on the non-dominated solution, a new combination of iterative design variables is generated, and the iteration number is incremented by one. When the number of iterations is less than or equal to the preset number of iterations, the iterative design variable combination is configured as the initial design variable combination and then iterative calculation is performed. The iterative calculation represents the above-mentioned step of generating a new iterative design variable combination. If the number of iterations is greater than a preset number, the non-dominated solution is configured as the Pareto optimal solution set.

5. The method for generating a photovoltaic module layout scheme according to claim 3, characterized in that, The construction of the first layout scheme and the digital twin model of multi-source data collaboration includes: Convert the building data into a BIM model; The first layout scheme is superimposed onto the BIM model to obtain the first model; The historical weather data is correlated with the first model to obtain the second model; The third model is obtained by overlaying the humanistic data and the landscape greening data onto the second model. Based on the first layout scheme, the historical weather data, the landscape greening data, the human data, and the building data, a first coupled model is generated that can calculate the hourly power generation, indoor heat gain rate, ventilation volume, daylighting rate, and water collection volume at each location. The first coupled model is then superimposed on the third model to obtain the digital twin model.

6. The method for generating a photovoltaic module layout scheme according to claim 1, characterized in that, After obtaining the simulation results by simulating the digital twin model, the process includes: A simulation map is generated based on the simulation results. The simulation map is used to indicate the difference between the multi-dimensional simulation data and the target data at each location of the target installation area. The multi-dimensional simulation data characterizes the simulated hourly power generation, indoor heat gain rate, daylighting rate, ventilation volume, and water collection volume. The simulated image is then subjected to grayscale conversion, filtering, and denoising processes in sequence to obtain a preprocessed image. The preprocessed image is divided into multiple spatial image sub-regions; Abnormal region information is obtained based on the similarity between multiple spatial image sub-regions. The abnormal region information represents the contour information and spatial location information of the target region that is significantly different from the adjacent regions.

7. The method for generating a photovoltaic module layout scheme according to claim 6, characterized in that, The step of obtaining abnormal region information based on the similarity between multiple spatial image sub-regions includes: The comparison result is obtained by comparing the first similarity between the first image sub-region and the third image sub-region and the second similarity between the second image sub-region and the third image sub-region. The third image sub-region is any of the spatial image sub-regions. The first image sub-region is located on the first side of the third image sub-region, and the second image sub-region is located on the second side of the third image sub-region. The first image sub-region, the second image sub-region, and the third image sub-region all represent the spatial image sub-region. If the comparison result indicates that the first similarity is greater than the second similarity, the third image sub-region is adjusted by unit distance and scaled by unit scale sequentially towards the location of the second image sub-region, and the contour information and spatial information of the third image sub-region with the largest difference between the first similarity and the second similarity are configured as the abnormal region information; If the comparison result indicates that the first similarity is less than the second similarity, the third image sub-region is adjusted by unit distance and scaled by unit scale sequentially towards the location of the first image sub-region, and the contour information and spatial information of the third image sub-region with the largest difference between the first similarity and the second similarity are configured as the abnormal region information; If the comparison result indicates that the first similarity is equal to the second similarity, the first image sub-region and the second image sub-region are simultaneously moved by a unit distance and scaled by a unit until the first similarity is not equal to the second similarity.

8. A photovoltaic module layout scheme generation system, characterized in that, The system includes a controller, which comprises a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, it implements the photovoltaic module layout scheme generation method according to any one of claims 1-7.

9. A computer storage medium, characterized in that, The computer storage medium stores computer-executable instructions, which are used to execute the photovoltaic module layout scheme generation method according to any one of claims 1-7.

Images