Intelligent photovoltaic sunshade method based on cloud cover recognition

Abstract

The invention relates to an intelligent photovoltaic sunshade method and system based on cloud cover recognition, and aims to improve the power generation efficiency of a photovoltaic system and achieve typhoon resistance. According to the method, cloud amount data are collected in real time through a multispectral camera and a meteorological station sensor, and feature values of cloud are extracted after preprocessing. A deep belief network (DBN) model of a double-layer RBM structure is used for power supply prediction, and model parameters are optimized through particle swarm optimization (PSO). And intelligently adjusting the angle of the sun shield or starting an anti-typhoon mode according to a comparison result of the predicted power supply quantity and a dynamically adjusted threshold value. The system further comprises an environment parameter prediction module and a user interaction interface, and the comfortable and energy-saving indoor environment of the near-zero-carbon building is achieved. The method and the system provided by the invention can effectively deal with extreme weather, optimize energy utilization and have a wide application prospect.

Description

Technical Field

[0001] The present invention relates to the technical field of near-zero carbon buildings, and particularly to an intelligent photovoltaic shading method based on cloud amount recognition, which is especially applicable to tropical climate regions with high humidity, strong ultraviolet rays and frequent typhoons, and can realize intelligent adjustment of the shading angle and typhoon resistance function. Background Art

[0002] With the intensification of global climate change and the increasingly severe energy crisis, near-zero carbon buildings, as a sustainable building model, have received extensive attention. Near-zero carbon buildings significantly reduce building carbon emissions by maximizing the use of renewable energy (such as solar energy) and optimizing energy consumption. However, in tropical climate regions (such as Hainan Province), the design and operation of near-zero carbon buildings face unique climate challenges, including high temperature, high humidity, strong ultraviolet rays, and frequent typhoon weather. These climate conditions pose higher requirements for the building's shading system and solar energy utilization efficiency.

[0003] Currently, the common intelligent shading systems on the market mainly rely on light sensors and temperature sensors to adjust the shading angle. However, these systems have the following significant defects in the application in tropical climate regions:

[0004] 1. Unable to dynamically adapt to cloud amount changes: Existing systems usually only adjust the shading angle according to the light intensity and cannot identify real-time sky cloud amount changes, resulting in low solar energy utilization efficiency. For example, in cloudy weather, the system may not be able to accurately predict the fluctuations of solar radiation intensity, leading to untimely adjustment of the shading angle and affecting indoor lighting and solar power generation efficiency.

[0005] 2. Lack of typhoon resistance ability: In regions with frequent typhoons, existing shading systems usually do not have an automatic retraction function. When a typhoon comes, the sunshade panels may be damaged by strong winds, and even pose a safety hazard to the building structure. For example, when the super typhoon "Yagi" landed in Hainan in 2024, the shading systems of a large number of buildings were damaged because they could not be retracted in time.

[0006] 3. Unable to balance lighting and power generation requirements: Existing systems usually only focus on a single requirement of indoor lighting or solar power generation and cannot achieve dynamic balance between the two. For example, during sunny periods, the system may over-shade the sunlight to meet the indoor lighting requirements but ignore the maximization of solar power generation utilization.

[0007] In view of the above defects of the existing technology, the present invention proposes an intelligent photovoltaic shading method based on cloud amount recognition, which has the following innovations:

[0008] 1. Cloud amount recognition and dynamic adjustment: Real-time collection of sky cloud amount data through multi-spectral cameras and meteorological station sensors, combined with a Deep Belief Network (DBN) model to predict solar radiation intensity, and dynamically adjust the shading angle to achieve the optimal balance between solar power generation and indoor daylighting.

[0009] 2. Anti-typhoon function: Introduction of wind speed and wind direction sensors to monitor environmental wind speed changes in real time. When the wind speed exceeds the preset threshold, the system automatically retracts the sunshade to avoid equipment damage in typhoon weather.

[0010] 3. Multi-objective optimization algorithm: Prediction of future environmental parameters through a Long Short-Term Memory (LSTM) model, combined with a multi-objective optimization algorithm to dynamically adjust the shading angle, and balance building energy consumption, indoor comfort, and solar power generation efficiency.

[0011] By combining cloud amount recognition, deep learning, and multi-objective optimization technologies, the present invention solves the key problems in the application of existing intelligent sunshade systems in tropical climate regions, and provides an efficient, reliable, and adaptable sunshade solution for nearly zero-carbon buildings. Summary of the Invention

[0012] The present invention aims to solve the deficiencies of existing intelligent sunshade systems under special climatic conditions in Hainan region, and proposes an intelligent photovoltaic sunshade method based on cloud amount recognition. This method can automatically adjust the shading angle according to environmental changes, improve the power generation efficiency of the photovoltaic system, and at the same time has the anti-typhoon ability to meet the requirements of nearly zero-carbon buildings for a comfortable and energy-saving indoor environment.

[0013] The object of the present invention can be achieved through the following steps:

[0014] Step S1: Collection of sky cloud amount data

[0015] Real-time collection of sky cloud amount data above the building through multi-spectral cameras deployed on the building top and ground meteorological station sensors;

[0016] Step S2: Preprocessing of cloud amount data

[0017] Preprocess the cloud amount data, including denoising, filtering, and outlier removal, extract the area, movement speed, thickness, and density of the cloud as feature values, and perform normalization processing;

[0018] Step S3: Prediction of power supply

[0019] Construct a Deep Belief Network (DBN) model based on a two-layer RBM structure, divide the training set and test set in a ratio of 7:3, and use 5-fold cross-validation to optimize the model parameters, and output the predicted power supply;

[0020] Step S4: Analysis of power supply-demand conflict

[0021] Dynamically adjust the threshold of power supply according to seasons and weather types. If the predicted power supply is lower than the threshold, generate a signal indicating insufficient solar radiation intensity.

[0022] Step S5: Dynamic adjustment of sunshade angle

[0023] Combine the real-time sensor data of temperature, humidity, and wind speed and the meteorological forecast data. Predict the environmental parameters for the next 1 hour through a long short-term memory network (LSTM) model, input the sunshade angle adjustment model, and output the optimal adjustment angle.

[0024] Step S6: Anti-typhoon function

[0025] When receiving typhoon warning information, start the anti-typhoon mode, and fold and recycle all the photovoltaic panels to reduce the impact of typhoons on the photovoltaic panels.

[0026] As a further technical solution of the present invention: In step S1, the multispectral camera and the meteorological station sensor synchronize data every 5 minutes, and calibrate the data accuracy through the Kalman filtering algorithm to ensure that the cloud area error ≤ 5% and the movement speed error ≤ 0.2 m / s.

[0027] As a further technical solution of the present invention: In step S2, denoising is performed by Gaussian filter smoothing, filtering is performed by median filtering algorithm to eliminate impulse noise, outliers are removed by mean and standard deviation determination, and the area, movement speed, thickness, and density of the cloud are extracted as feature values and normalized.

[0028] As a further technical solution of the present invention: In step S3, the evaluation indexes of the DBN model include mean square error (MSE ≤ 0.05) and prediction accuracy rate (≥ 95%). If the standard is not met, it will automatically switch to the lightweight GRU model for prediction.

[0029] As a further technical solution of the present invention: In step S4, the adjustment formula for the dynamic threshold is:

[0030] T threshold =T base ×(1 + α × ΔW + β × R)

[0031] Where:

[0032] T base is the basic threshold, determined according to the building's historical power demand data;

[0033] ΔW is the weather type weight coefficient (sunny day = 0, cloudy day = 0.2, rainy day = 0.5);

[0034] α is the seasonal correction coefficient, with a value ranging from 1.1 to 1.2 in summer (from March to November) and a value ranging from 0.8 to 0.9 in winter (from December to February of the following year);

[0035] β is the rainy season correction coefficient, with a value of 1.2 in the rainy season (from May to October) and a value of 1.0 in the non-rainy season;

[0036] R is the rainy season weight coefficient, with a value of 0.3 in the rainy season and 0 in the non-rainy season.

[0037] For tropical monsoon climate regions (such as Hainan Province), the seasonal correction coefficient α and the rainy season correction coefficient β are dynamically adjusted in the following way:

[0038] Based on historical meteorological data and solar radiation intensity data, a random forest model is used to predict the values of α and β for the next 7 days;

[0039] If the prediction result is the rainy season and the cloud cover rate ≥ 60%, then β is automatically adjusted to 1.3;

[0040] If the prediction result is the non-rainy season and the solar radiation intensity ≥ 800 W / m 2 , then α is automatically adjusted to 1.0.

[0041] As a further technical solution of the present invention: in step S5, the sunshade angle adjustment model adopts a multi-objective optimization algorithm, and the objective function is:

[0042] min(E 能耗 +λ×∣θ 调节 -θ 舒适 ∣)

[0043] Among them, the energy consumption E is the estimated energy consumption of the building, the comfort θ is the human comfort angle range (20° - 45°), and λ is the weight coefficient.

[0044] A user feedback module is introduced to receive the user's satisfaction score for the sunshade angle in real time through a mobile terminal and dynamically adjust the value of λ.

[0045] An intelligent photovoltaic sunshade method based on cloud amount recognition, the method includes:

[0046] Data acquisition module: integrating a multi-spectral camera, a temperature and humidity sensor, a wind speed sensor, and an electricity metering unit;

[0047] Edge computing module: deployed on the local building server, performing the data preprocessing, model prediction, and threshold calculation described in claims 1 - 7;

[0048] Cloud collaborative module: uploading historical data to the cloud and optimizing the global model parameters through federated learning;

[0049] User interface: Provide visual displays of real-time sunshade angles, power supply and demand status, and abnormal warnings.

[0050] Typhoon resistance module: Receive typhoon warning information, activate the typhoon resistance mode, and adjust the photovoltaic panels to the state with the minimum wind-receiving area.

[0051] The data acquisition module transmits real-time data to the edge computing module. The edge computing module performs data preprocessing and model prediction, and uploads the results to the cloud collaboration module for global optimization. Finally, the real-time status and warning information are displayed through the user interface.

[0052] Advantages of the present invention:

[0053] By combining cloud amount recognition, deep learning, and multi-objective optimization technologies, the present invention achieves the following beneficial effects: By real-time recognizing cloud amount changes and dynamically adjusting the sunshade angle, the solar power generation efficiency is maximized; By using a wind speed sensor to monitor the ambient wind speed changes in real time, the sunshade is automatically retracted to avoid equipment damage during typhoon weather; By using a multi-objective optimization algorithm to dynamically adjust the sunshade angle, the building energy consumption, indoor comfort, and solar power generation efficiency are balanced. Description of the drawings

[0054] To more clearly illustrate the technical solutions in the embodiments of the present invention, the following will briefly introduce the drawings required for the description of the embodiments. Obviously, the following drawings are only some embodiments of the present invention. For those of ordinary skill in the art, without creative efforts, other drawings can be obtained based on these drawings.

[0055] Figure 1 is a flowchart of an intelligent photovoltaic sunshade method based on cloud amount recognition provided in Embodiment 1 of the present invention;

[0056] Figure 2 is a schematic structural diagram of an intelligent photovoltaic sunshade method based on cloud amount recognition provided in Embodiment 2 of the present invention;

[0057] Figure 3 is an algorithm flowchart of the DBN model provided in Embodiment 1 of the present invention; Detailed implementation manners

[0058] To enable those skilled in the art to better understand the solutions of the present invention, the following will clearly and completely describe the technical solutions in the embodiments of the present invention with reference to the drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all of them. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts shall fall within the protection scope of the present invention.

[0059] Example 1

[0060] As Figure 1 shown, an intelligent photovoltaic shading method based on cloud amount recognition provided by an embodiment of the present invention specifically includes the following steps:

[0061] Step S1: Sky cloud amount data collection

[0062] Real-time collect the sky cloud amount data above the building through the multispectral camera deployed on the building top and the ground meteorological station sensor;

[0063] In some embodiments, the sky cloud amount data is obtained in real time through the multispectral camera deployed on the building top and the ground meteorological station sensor; wherein, the sky cloud amount data includes the area of clouds in the sky, the moving speed of clouds, and the thickness of clouds;

[0064] As Figure 3 shown, Step S2: Preprocess the cloud amount data, including denoising, filtering, and outlier removal, extract the area, moving speed, thickness, and density of clouds as feature values, and perform normalization processing; Step S3: Construct a deep belief network (DBN) model based on a two-layer RBM structure, divide the training set and the test set in a ratio of 7:3, and use 5-fold cross-validation to optimize the model parameters, and output the predicted power supply;

[0065] For the selected feature values, the influence of the dimension on the final result is eliminated through normalization, so that different variables are comparable. The normalization formula is:

[0066]

[0067] where x ， is the normalized value, max is the maximum value in the sample, and min is the minimum value in the sample;

[0068] Adopt a DBN model with a two-layer RBM structure, input the training set into the DBN model for training, and then input the test set into the trained DBN model for prediction and evaluation; wherein, the feature value is used as the input of the DBN model, and the label is used as the output of the DBN model; the power supply at the corresponding moment can be obtained;

[0069] Exemplarily, construct and train a deep belief network DBN model: Adopt a DBN model with a two-layer RBM structure, input the training set into the DBN model for training, and then input the test set into the trained DBN model for prediction and evaluation; wherein, the feature value is used as the input of the DBN model, and the label is used as the output of the DBN model;

[0070] Optimization of DBN model network parameters: The particle swarm optimization (PSO) algorithm is used to optimize the network parameters of the DBN model. That is, according to the optimization objective, the optimization interval is set for the network parameters to be optimized: the number of neurons L1 in the first-layer RBM ∈ [200, 1000], the number of neurons L2 in the second-layer RBM ∈ [400, 1000], the RBM sample batch size ∈ [10, 100], the number of RBM iterations ∈ [100, 1000], and the BP sample batch size ∈ [1, 30]; the internal parameters of PSO are set as: c1 = 2.6, c2 = 0.6, ωmax = 0.9, ωmin = 0.4, the population size is 50, and the number of iterations is 50 times. The following formula is used as the fitness function;

[0071]

[0072] In the formula, ture i represents the actual power supply, pred i represents the predicted power supply, and N represents the number of test location samples;

[0073] The optimized network parameters are returned to the DBN model until the optimization conditions are met. The optimization condition is 50 iterations, and the optimal DBN model is output;

[0074] The dataset of feature values to be predicted is input into the optimal DBN model, and the power supply at the corresponding time can be obtained;

[0075] The evaluation metrics of the DBN model include the mean square error (MSE ≤ 0.05) and the prediction accuracy rate (≥ 95%). If the standard is not met, it will automatically switch to the lightweight GRU model for prediction.

[0076] Step S4: Analysis of power supply and demand conflicts

[0077] Dynamically adjust the power supply threshold according to the season and weather type. If the predicted power supply is lower than the threshold, a signal indicating insufficient solar radiation intensity is generated;

[0078] In some embodiments, the power supply at the corresponding time is obtained, and the power supply at the corresponding time is compared with the power supply threshold at the corresponding time;

[0079] When the power supply at the corresponding time is greater than or equal to the power supply threshold at the corresponding time, it indicates that the electricity converted from the lighting demand of the solar panels meets the conversion standard requirements, and the current outdoor solar radiation intensity is sufficient, so a signal indicating sufficient outdoor solar radiation intensity is generated;

[0080] When the power supply at the corresponding time is less than the power supply threshold at the corresponding time, it indicates that the electricity converted from the lighting demand of the solar panels does not meet the conversion standard requirements, and the current outdoor solar radiation intensity is insufficient, so a signal indicating insufficient outdoor solar radiation intensity is generated;

[0081] When receiving a sufficient outdoor solar radiation intensity signal, obtain the power supply amount at the corresponding moment and the actual power supply amount at the corresponding moment, calculate the difference between the obtained power supply amount at the corresponding moment and the actual power supply amount to obtain a power supply difference;

[0082] Compare the obtained power supply difference with a power supply difference threshold;

[0083] If the power supply difference is greater than or equal to the power supply difference threshold, it indicates that there is no conflict between the solar radiation intensity demand of the solar panel and the actual power demand of the building, that is, the power converted by the current panel through solar energy meets the supply demand of a certain device in the building, and then generate a qualified solar radiation intensity supply signal;

[0084] If the power supply difference is less than the power supply difference threshold, it indicates that there is a conflict between the solar radiation intensity demand of the solar panel and the actual power demand of the building, that is, the power converted by the current panel through solar energy does not meet the supply demand of a certain device in the building, and then generate an unqualified solar radiation intensity supply signal;

[0085] According to the generated sufficient outdoor solar radiation intensity signal, insufficient outdoor solar radiation intensity signal, qualified daylighting supply signal and unqualified daylighting supply signal, the intelligent photovoltaic shading system can dynamically adjust the shading board angle to optimize the power generation efficiency of the solar panel and the power supply inside the building.

[0086] Step S5: Dynamic adjustment of shading angle

[0087] Combined with the real-time sensor data of temperature, humidity and wind speed and the weather forecast data, predict the environmental parameters in the next 1 hour through a long short-term memory network (LSTM) model, input it into the shading board angle adjustment model, and output the optimal adjustment angle;

[0088] In step S5, the shading board angle adjustment model adopts a multi-objective optimization algorithm, and the objective function is:

[0089] min(E 能耗 +λ×∣θ 调节 -θ 舒适 ∣)

[0090] Among them, the energy consumption E is the estimated energy consumption of the building, the comfort θ is the human comfort angle range (20° - 45°), and λ is the weight coefficient.

[0091] Introduce a user feedback module, and receive the user's satisfaction score for the shading angle in real time through a mobile terminal to dynamically adjust the λ value.

[0092] Step S6: Anti-typhoon function

[0093] When a typhoon warning message is received, the anti-typhoon mode is activated, and all photovoltaic panels are gathered, folded and recycled to reduce the impact of the typhoon on the photovoltaic panels.

[0094] Embodiment 2

[0095] As Figure 2 shown, an intelligent photovoltaic sunshade method based on cloud amount recognition provided by an embodiment of the present invention specifically includes the following modules:

[0096] Data acquisition module: integrating a multi-spectral camera, a temperature and humidity sensor, a wind speed sensor, and an electricity metering unit;

[0097] Edge computing module: deployed on a local building server, performing data preprocessing, model prediction, and threshold calculation described in claims 1-7;

[0098] Cloud collaboration module: uploading historical data to the cloud and optimizing global model parameters through federated learning;

[0099] User interface: providing visual displays of real-time sunshade angles, power supply and demand status, and abnormal alarms;

[0100] Anti-typhoon module: receiving typhoon warning information, activating the anti-typhoon mode, and adjusting the photovoltaic panels to the state with the minimum windward area.

[0101] The data acquisition module transmits real-time data to the edge computing module. The edge computing module performs data preprocessing and model prediction, and uploads the results to the cloud collaboration module for global optimization. Finally, the real-time status and alarm information are displayed through the user interface.

[0102] Embodiment 3

[0103] An embodiment of the present invention also provides a computer-readable storage medium, on which a computer program is stored. When the computer program is run by a processor, it implements an intelligent photovoltaic sunshade method based on cloud amount recognition as described in any one of the above methods.

[0104] In this embodiment, if the integrated unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on such an understanding, to implement all or part of the processes in the method of the above embodiment in this application, a computer program can be used to instruct relevant hardware to complete. The computer program can be stored in a computer-readable storage medium. When the computer program is executed by a processor, the steps of the above method embodiments can be implemented. Among them, the computer program includes computer program code, and the computer program code can be in the form of source code, object code, executable file or some intermediate form, etc. The computer-readable medium can at least include: any entity or device that can carry the computer program code to the photographing device / terminal device, recording medium, computer memory, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), electrical carrier signal, telecommunication signal, and software distribution medium. For example, a USB flash drive, a mobile hard disk, a magnetic disk, or an optical disc, etc. In some jurisdictions, according to legislation and patent practice, the computer-readable medium cannot be an electrical carrier signal and a telecommunication signal.

[0105] In the above embodiments, the descriptions of the various embodiments have their own focuses. For the parts not detailed or recorded in a certain embodiment, reference can be made to the relevant descriptions of other embodiments.

[0106] Those of ordinary skill in the art can realize that the units and algorithm steps of the examples described in combination with the embodiments disclosed in this document can be implemented by electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are executed in a hardware or software manner depends on the specific application and design constraints of the technical solution. Professional technicians can use different methods to implement the described functions for each specific application, but such implementation should not be considered to exceed the scope of this application.

[0107] In the embodiments disclosed in this application, it should be understood that the disclosed device / terminal device and method can be implemented in other ways. For example, the device / terminal device embodiments described above are only illustrative. For example, the division of the modules or units is only a logical function division. In actual implementation, there can be other division methods. For example, multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the displayed or discussed coupling or direct coupling or communication connection to each other can be through some interfaces. The indirect coupling or communication connection of the device or unit can be in an electrical, mechanical or other form.

[0108] The unit described as a separation component may or may not be physically separated. The component shown as a unit may or may not be a physical unit, that is, it may be located in one place or distributed over multiple network units. Some or all of the units can be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0109] The above formulas are all dimensionless and take their numerical values for calculation. The formulas are obtained by collecting a large amount of data for software simulation to get a formula closest to the actual situation. The preset parameters in the formulas are set by those skilled in the art according to the actual situation.

[0110] The above has described in detail an embodiment of the present invention, but the content described is only a preferred embodiment of the present invention and cannot be considered as limiting the scope of implementation of the present invention. All equivalent changes and improvements made according to the scope of the application of the present invention should still fall within the scope covered by the patent of the present invention.

Claims

1. An intelligent photovoltaic shading method based on cloud cover recognition, characterized in that: The following steps are involved: Step S1: Sky cloud data collection The cloud cover data above the building is collected in real time through multispectral cameras deployed on the top of the building and ground weather station sensors; Step S2: Cloud data preprocessing Preprocess the cloud data, including denoising, filtering and outlier removal, extract the cloud area, movement speed, thickness and density as feature values, and perform normalization; Step S3: Power supply prediction A deep belief network (DBN) model based on a two-layer RBM structure was constructed, the training set and the test set were divided into a 7:3 ratio, and 5-fold cross-validation was used to optimize the model parameters and output the predicted power supply; Step S4: Power supply and demand conflict analysis Dynamically adjust the power supply threshold according to the season and weather type. If the predicted power supply is lower than the threshold, generate a signal of insufficient solar radiation intensity; Step S5: Dynamic adjustment of shading angle Combining the real-time sensor data of temperature, humidity, and wind speed with weather forecast data, the long short-term memory network (LSTM) model is used to predict the environmental parameters for the next hour, which are then input into the visor angle adjustment model to output the optimal adjustment angle. Step S6: Typhoon resistance function When a typhoon warning message is received, the anti-typhoon mode is activated and all photovoltaic panels are gathered, folded and recycled to reduce the impact of the typhoon on the photovoltaic panels.

2. The method according to claim 1, characterized in that In step S1: The multispectral camera synchronizes data with the weather station sensor every 5 minutes, and the data accuracy is calibrated through the Kalman filter algorithm to ensure that the cloud area error is ≤5% and the movement speed error is ≤0.2m / s.

3. The method according to claim 1, characterized in that In step S2: denoising adopts Gaussian filter smoothing processing, filtering adopts median filtering algorithm to eliminate impulse noise, outliers are eliminated by judging by mean and standard deviation, and the area, movement speed, thickness and density of the cloud are extracted as characteristic values ​​and normalized.

4. The method according to claim 1, characterized in that: In step S3: The evaluation indicators of the DBN model include mean square error (MSE≤0.05) and prediction accuracy (≥95%). If the standards are not met, it will automatically switch to the lightweight GRU model for prediction.

5. The method according to claim 1, characterized in that In step S4: The dynamic threshold adjustment formula is: T threshold =T base ×(1+α×ΔW+β×R) in: T base is the basic threshold, determined based on the building's historical electricity demand data; ΔW is the weather type weight coefficient (sunny = 0, cloudy = 0.2, rainy = 0.5); α is the seasonal correction coefficient, which is 1.1-1.2 in summer (March to November) and 0.8-0.9 in winter (December to February of the following year); β is the correction coefficient for the rainy season, which is 1.2 in the rainy season (May to October) and 1.0 in the non-rainy season; R is the rainy season weight coefficient, rainy season = 0.3, non-rainy season = 0; For tropical monsoon climate areas (such as Hainan Province), the seasonal correction coefficient α and the rainy season correction coefficient β are dynamically adjusted in the following manner: Based on historical meteorological data and solar radiation intensity data, the random forest model is used to predict the α and β values ​​for the next 7 days; If the forecast result is rainy season and the cloud cover is ≥60%, β is automatically adjusted to 1.3; If the forecast result is non-rainy season and the solar radiation intensity is ≥800W / m 2 , α is automatically adjusted to 1.

0.

6. The method according to claim 1, characterized in that In step S5: the sun visor angle adjustment model adopts a multi-objective optimization algorithm, and the objective function is: min(E 能耗 +λ×∣θ 调节 -θ 舒适 ∣) Among them, energy consumption E is the estimated energy consumption of the building, comfort θ is the human comfort angle range (20°~45°), and λ is the weight coefficient.

7. The method according to claim 6, characterized in that In the multi-objective optimization algorithm: A user feedback module is introduced to receive users' satisfaction scores on shading angles in real time through mobile terminals and dynamically adjust the λ value.

8. The method according to claim 1, characterized in that The method is performed by an intelligent photovoltaic shading system, which includes: Data acquisition module: integrated with multi-spectral camera, temperature and humidity sensor, wind speed sensor and power metering unit; Edge computing module: deployed on the local server of the building, performing the data preprocessing, model prediction and threshold calculation described in claims 1-7; Cloud collaboration module: upload historical data to the cloud and optimize global model parameters through federated learning; User interaction interface: provides visual display of real-time shading angle, power supply and demand status, and abnormal alarms; Anti-typhoon module: receives typhoon warning information, activates anti-typhoon mode, and adjusts the photovoltaic panels to the state of minimum wind-exposed area.

9. The method according to claim 8, characterized in that The data acquisition module transmits real-time data to the edge computing module, which performs data preprocessing and model prediction, and uploads the results to the cloud collaboration module for global optimization, and finally displays real-time status and alarm information through the user interaction interface.

Images