Photovoltaic tracking support angle optimization and power generation efficiency improvement method for meteorological prediction

Abstract

The invention provides a photovoltaic tracking support angle optimization and power generation efficiency improvement method for meteorological prediction, and relates to the technical field of meteorological prediction, and the method comprises the steps: building a deep neural network model through receiving meteorological data, predicting future meteorological data, calculating a solar azimuth angle and an elevation angle, and building a photovoltaic tracking support motion constraint model for safety risk avoidance. And an optimal angle sequence is optimized and solved by adopting an improved dynamic programming algorithm, and an angle adjusting instruction is generated to control the bracket to adjust. According to the invention, the power generation efficiency of the photovoltaic module is improved, and the safety of the system in severe weather is ensured.

Description

Technical Field

[0001] The present invention relates to meteorological prediction technologies, and particularly to a method for optimizing the angle of a photovoltaic tracking bracket and improving the power generation efficiency in meteorological prediction. Background Art

[0002] With the global energy structure adjustment and the rapid development of renewable energy, as an important part of clean energy, photovoltaic power generation is widely used globally. To improve the power generation efficiency of photovoltaic power stations, photovoltaic tracking bracket technology has emerged. This technology adjusts the angle of photovoltaic modules to better track the sun's movement trajectory, maximize the reception of solar radiation energy, and thus increase power generation.

[0003] Traditional photovoltaic tracking systems mainly adopt simple tracking strategies based on time or light intensity, that is, adjust the bracket angle according to pre-calculated sun trajectory data or real-time light intensity sensor information. These systems usually trigger angle adjustments at fixed time intervals or when the sun position changes, and do not fully consider the impact of weather changes on power generation efficiency. With the development of meteorological prediction technologies and artificial intelligence technologies, optimizing the angle of photovoltaic tracking brackets by combining meteorological data has become a new direction for improving the power generation efficiency of photovoltaic power stations.

[0004] Existing technologies generally lack the ability to accurately predict and respond to complex meteorological conditions. Most systems rely only on historical data or simple weather forecasts, and cannot make timely and accurate predictions and adjustments for extreme weather conditions such as short-term cloud changes and sudden wind speed changes, resulting in poor tracking strategy effects under unstable weather conditions.

[0005] Existing tracking systems lack effective safety protection mechanisms under extreme weather conditions. When encountering severe weather such as strong winds and heavy rains, if the photovoltaic tracking brackets cannot be adjusted to a safe angle in time, it may cause equipment damage, which not only increases the maintenance cost but also reduces the overall reliability and lifespan of the power station.

[0006] Traditional optimization algorithms are inefficient in calculating and difficult to achieve global optimality when solving the optimal angle sequence of photovoltaic tracking brackets. Most systems adopt simple immediate optimization strategies and cannot consider the overall maximization goal of all-day power generation. As a result, although the angle adjustment in some time periods achieves immediate power maximization, it is not the optimal solution from the perspective of all-day power generation, and the power generation potential of the photovoltaic power station cannot be fully exploited. Summary of the Invention

[0007] The embodiments of the present invention provide a method for optimizing the angle of a photovoltaic tracking bracket and improving the power generation efficiency in meteorological prediction, which can solve the problems in the existing technologies.

[0008] In the first aspect of the embodiments of the present invention, a method for optimizing the angle of a photovoltaic tracking bracket and improving the power generation efficiency in weather prediction is provided, including: Receiving meteorological data of the target photovoltaic power station area collected by a meteorological monitoring station, constructing a deep neural network model based on the meteorological data, and using the deep neural network model to predict the meteorological data within a preset future time period to obtain predicted meteorological data; Based on the predicted meteorological data, calculating the solar azimuth angle and the solar altitude angle of the target photovoltaic power station area within the preset future time period; According to the cloud cover rate and wind speed data in the predicted meteorological data, establishing a motion constraint model of the photovoltaic tracking bracket, determining whether the cloud cover rate is greater than a preset constraint threshold and whether the wind speed data is greater than a safe wind speed threshold. When the judgment condition is met, calculating the safe avoidance angle of the photovoltaic tracking bracket and controlling the photovoltaic tracking bracket to rotate to the safe avoidance angle; According to the solar radiation intensity and temperature data in the predicted meteorological data, combining the solar azimuth angle and the solar altitude angle, using an improved dynamic programming algorithm to optimize and solve the optimal angle sequence of the photovoltaic tracking bracket within the preset future time period, where the optimal angle sequence is used to maximize the power generation efficiency of the photovoltaic module; According to the optimal angle sequence, generating an angle adjustment instruction for the photovoltaic tracking bracket, and sending the angle adjustment instruction to the controller of the photovoltaic tracking bracket to control the photovoltaic tracking bracket to perform angle adjustment according to the optimal angle sequence.

[0009] Constructing a deep neural network model based on the meteorological data, and using the deep neural network model to predict the meteorological data within a preset future time period to obtain predicted meteorological data includes: Normalizing the meteorological data according to an adaptive time window length to obtain normalized features; The deep neural network model includes a parallel feature extraction layer, an adaptive fusion layer and a dynamic attention layer; inputting the normalized features into the parallel feature extraction layer, and extracting multi-scale local features through a spatial pyramid convolution branch; inputting the multi-scale local features into the adaptive fusion layer, calculating adaptive weights based on feature discriminability and performing feature fusion to obtain enhanced features; inputting the enhanced features into the dynamic attention layer, and capturing key information at different time scales based on a multi-scale time attention mechanism to obtain attention-enhanced features; Constructing an adaptive multi-task loss function, where the adaptive multi-task loss function includes a prediction error loss term and a temporal correlation loss term; adaptively adjusting the weight coefficients of each loss term according to the difficulty of the prediction task and the data distribution characteristics; using the adaptive multi-task loss function to perform end-to-end training on the deep neural network model; Input the meteorological data to be predicted into the trained deep neural network model to generate predicted meteorological data.

[0010] Calculating the solar azimuth angle and solar altitude angle of the target photovoltaic power station area within the future preset time period based on the predicted meteorological data includes: Perform time series decomposition on the predicted meteorological data to extract time-varying features and periodic features; perform data correction on the predicted meteorological data based on the time-varying features and the periodic features to obtain corrected predicted data; Calculate the atmospheric refraction correction amount according to the predicted air pressure data, predicted temperature data and site altitude in the corrected predicted data; calculate the atmospheric scattering correction amount according to the predicted relative humidity data, predicted cloud cover data and site terrain height in the corrected predicted data; Perform weighted combination on the atmospheric refraction correction amount and the atmospheric scattering correction amount, dynamically adjust the weight coefficient according to the time-varying features of the corrected predicted data to obtain a comprehensive correction coefficient; calculate the initial solar azimuth angle and initial solar altitude angle of the target photovoltaic power station area within the future preset time period based on the comprehensive correction coefficient; Perform mutual constraint calculation on the initial solar azimuth angle and the initial solar altitude angle according to the geometric relationship between the solar azimuth angle and the solar altitude angle; correct the calculation result based on the historical prediction error to obtain the solar azimuth angle and solar altitude angle of the target photovoltaic power station area within the future preset time period.

[0011] According to the cloud cover rate and wind speed data in the predicted meteorological data, establish a motion constraint model for the photovoltaic tracking bracket, and judge whether the cloud cover rate is greater than the preset constraint threshold and whether the wind speed data is greater than the safe wind speed threshold. When the judgment condition is satisfied, calculate the safe avoidance angle of the photovoltaic tracking bracket and control the photovoltaic tracking bracket to rotate to the safe avoidance angle, including: The motion constraint model includes wind load constraint and moment balance constraint; generate a trigger signal based on the magnitude relationship between the cloud cover rate and the preset constraint threshold and the magnitude relationship between the wind speed data and the safe wind speed threshold; Obtain the structural parameters of the photovoltaic tracking bracket, and the structural parameters include bracket area, bracket mass, current bracket inclination angle and distance from the bracket center of gravity to the rotation axis; substitute the wind speed data, the bracket area and the current bracket inclination angle into the wind load constraint to calculate the wind load value, and generate a wind torque based on the wind load value; Calculate the gravity torque according to the bracket mass and the distance from the bracket center of gravity to the rotation axis; substitute the wind torque and the gravity torque into the moment balance constraint to calculate the safe avoidance angle that satisfies the motion constraint model; Judge whether the risk avoidance control condition is satisfied according to the trigger signal. When the risk avoidance control condition is satisfied, obtain the current angle of the photovoltaic tracking bracket; calculate the target angular velocity based on the safe risk avoidance angle and the current angle, and control the photovoltaic tracking bracket to rotate to the safe risk avoidance angle according to the target angular velocity.

[0012] Optimally solving the optimal angle sequence of the photovoltaic tracking bracket within a future preset time period according to the solar radiation intensity and temperature data in the predicted meteorological data, in combination with the solar azimuth angle and the solar altitude angle, by using an improved dynamic programming algorithm includes: Obtain the motion parameters of the photovoltaic tracking bracket, where the motion parameters include the minimum rotation angle, the maximum rotation angle, and the rotation power consumption coefficient; establish an initial angle state space based on the minimum rotation angle and the maximum rotation angle; Calculate the air mass index according to the solar azimuth angle and the solar altitude angle, and calculate the direct radiation intensity based on the air mass index and the solar radiation intensity; multiply the direct radiation intensity by the rotation power consumption coefficient to obtain the theoretical optimal power generation; Establish a state density function based on the theoretical optimal power generation; adaptively divide the initial angle state space according to the state density function, subdivide the area where the state density is higher than the preset density threshold to obtain an optimized angle state space; calculate the maximum power generation corresponding to each angle state in the optimized angle state space; Calculate the temperature correction coefficient according to the temperature data and the temperature coefficient in the predicted meteorological data, multiply the temperature correction coefficient by the maximum power generation to obtain the actual power generation; construct a power generation cost function by weighted combination of the actual power generation and the maximum power generation; In the optimized angle state space, substitute the power generation cost function into the dynamic programming recurrence equation, and obtain the optimal state value at each moment through iterative calculation; generate the optimal angle sequence within a future preset time period based on the optimal state value.

[0013] In the optimized angle state space, substitute the power generation cost function into the dynamic programming recurrence equation, and obtain the optimal state value at each moment through iterative calculation; generating the optimal angle sequence within a future preset time period includes: Calculate the normalized difference between the actual power generation and the predicted power generation to obtain the power prediction error; substitute the power prediction error into the exponential decay function, and multiply by the base discount rate to obtain the time-varying discount factor; add the negative value of the power generation cost function to the product of the time-varying discount factor and the maximum value of the next moment state value to construct the dynamic programming recurrence equation; Calculate the current angle state value based on the dynamic programming recurrence equation, and perform weighted averaging on the current angle state value and the historical angle state value according to a preset learning rate to obtain an updated angle state value; calculate the relative error between the updated angle state value and the historical angle state value, and when the relative error is less than a preset convergence threshold, record the updated angle state value as the final state value; Select the angle that maximizes the final state value from the optimized angle state space as the optimal angle at this moment; determine whether the difference between the optimal angles at adjacent prediction times exceeds a preset maximum angle change rate, and if it exceeds, correct the optimal angle based on the maximum angle change rate; connect the corrected optimal angles in sequence to generate an optimal angle sequence of the photovoltaic tracking bracket within a preset future time period.

[0014] According to the optimal angle sequence, generating the angle adjustment instruction for the photovoltaic tracking bracket includes: Obtain the motion constraint parameters of the photovoltaic tracking bracket, where the motion constraint parameters include the maximum adjustment speed, the maximum acceleration, and the minimum adjustment time interval; perform trajectory smoothing processing on the optimal angle sequence based on the motion constraint parameters to generate a transition angle sequence that satisfies the maximum adjustment speed and the maximum acceleration; perform time discretization sampling on the transition angle sequence according to the minimum adjustment time interval, and convert the discretized sampled transition angle sequence into an angle adjustment instruction.

[0015] In a second aspect of the embodiments of the present invention, there is provided an electronic device, including: A processor; A memory for storing instructions executable by the processor; Wherein, the processor is configured to call the instructions stored in the memory to execute the method described above.

[0016] In a third aspect of the embodiments of the present invention, there is provided a computer-readable storage medium, on which computer program instructions are stored, and when the computer program instructions are executed by a processor, the method described above is implemented.

[0017] The beneficial effects of this application are as follows: Predict future meteorological data through a deep neural network model, and calculate the solar azimuth angle and altitude angle based on the predicted meteorological data, which improves the perception ability of the photovoltaic tracking system for future lighting conditions, makes the angle optimization forward-looking, and avoids the lag problem caused by traditional real-time data adjustment.

[0018] A motion constraint model is established based on cloud cover rate and wind speed data. When the severe weather conditions meet the constraint threshold, the system can automatically calculate the safe avoidance angle and control the rotation of the bracket, effectively ensuring the safe operation of photovoltaic equipment under extreme weather conditions and extending the service life of the equipment.

[0019] An improved dynamic programming algorithm is used to optimize and solve the optimal angle sequence. Considering the solar radiation intensity and temperature factors comprehensively, compared with the traditional fixed angle or simple tracking algorithm, it can maximize the power generation efficiency of photovoltaic modules under various meteorological conditions, and improve the overall power generation and economic benefits of the photovoltaic power station. Brief Description of the Drawings

[0020] Figure 1 It is a schematic flow chart of the method for optimizing the angle of a photovoltaic tracking bracket and improving power generation efficiency in meteorological prediction according to an embodiment of the present invention; Figure 2 It is a combined effect diagram of the atmospheric refraction correction amount and the scattering correction amount according to an embodiment of the present invention; Figure 3 It is a working flow chart of the safe avoidance control system of a photovoltaic tracking bracket according to an embodiment of the present invention; Figure 4 It is a diagram showing the influence of the temperature correction coefficient on the power generation efficiency according to an embodiment of the present invention; Figure 5 It is a flow chart of the photovoltaic tracking control method optimized by dynamic programming according to an embodiment of the present invention. Detailed Embodiments

[0021] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the accompanying 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 the embodiments. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts shall fall within the protection scope of the present invention.

[0022] The technical solutions of the present invention will be described in detail below with specific embodiments. These specific embodiments can be combined with each other, and the same or similar concepts or processes may not be repeated in some embodiments.

[0023] Figure 1 It is a schematic flow chart of the method for optimizing the angle of a photovoltaic tracking bracket and improving power generation efficiency in meteorological prediction according to an embodiment of the present invention, as Figure 1 shown, the method includes: Receive the meteorological data of the target photovoltaic power station area collected by the meteorological monitoring station, construct a deep neural network model according to the meteorological data, and use the deep neural network model to predict the meteorological data within a preset future time period to obtain predicted meteorological data; Based on the predicted meteorological data, calculate the solar azimuth angle and solar altitude angle of the target photovoltaic power station area within the preset future time period; According to the cloud cover rate and wind speed data in the predicted meteorological data, establish a motion constraint model for the photovoltaic tracking bracket, determine whether the cloud cover rate is greater than a preset constraint threshold and whether the wind speed data is greater than the safe wind speed threshold. When the judgment condition is met, calculate the safe avoidance angle of the photovoltaic tracking bracket and control the photovoltaic tracking bracket to rotate to the safe avoidance angle; According to the solar radiation intensity and temperature data in the predicted meteorological data, combined with the solar azimuth angle and the solar altitude angle, use an improved dynamic programming algorithm to optimize and solve the optimal angle sequence of the photovoltaic tracking bracket within the preset future time period, where the optimal angle sequence is used to maximize the power generation efficiency of the photovoltaic module; According to the optimal angle sequence, generate an angle adjustment instruction for the photovoltaic tracking bracket, and send the angle adjustment instruction to the controller of the photovoltaic tracking bracket to control the photovoltaic tracking bracket to adjust the angle according to the optimal angle sequence.

[0024] In an alternative embodiment, constructing a deep neural network model according to the meteorological data and using the deep neural network model to predict the meteorological data within a preset future time period to obtain predicted meteorological data includes: Perform normalization processing on the meteorological data according to an adaptive time window length to obtain normalized features; The deep neural network model includes a parallel feature extraction layer, an adaptive fusion layer and a dynamic attention layer; input the normalized features into the parallel feature extraction layer, and extract multi-scale local features through a spatial pyramid convolution branch; input the multi-scale local features into the adaptive fusion layer, calculate adaptive weights based on feature discriminability and perform feature fusion to obtain enhanced features; input the enhanced features into the dynamic attention layer, and capture key information at different time scales based on a multi-scale time attention mechanism to obtain attention-enhanced features; Construct an adaptive multi-task loss function, where the adaptive multi-task loss function includes a prediction error loss term and a temporal correlation loss term; adaptively adjust the weight coefficients of each loss term according to the difficulty of the prediction task and the data distribution characteristics; use the adaptive multi-task loss function to perform end-to-end training on the deep neural network model; Input the meteorological data to be predicted into the trained deep neural network model to generate predicted meteorological data.

[0025] Perform adaptive time window normalization on the meteorological data. In this embodiment, the meteorological data includes multiple meteorological elements such as temperature, humidity, air pressure, wind speed, and precipitation. After obtaining the historical meteorological data, the system automatically determines the optimal time window length according to the data distribution characteristics and the requirements of the prediction task. For example, for temperature data, the system determines 72 hours as the best window length after analysis; while for precipitation data, a 48-hour window may be more suitable. After determining the window length, perform normalization on the data within each window, specifically including: calculating the mean and standard deviation of the data within the window, and then subtracting the mean from the original data and dividing by the standard deviation to obtain standardized features with a mean of 0 and a standard deviation of 1. Experiments show that compared with the normalization method with a fixed window length, the adaptive window can improve the prediction accuracy by about 5.8%.

[0026] Construct a deep neural network model composed of a parallel feature extraction layer, an adaptive fusion layer, and a dynamic attention layer. In the parallel feature extraction layer, use spatial pyramid convolution branches to extract multi-scale local features. The specific implementation is as follows: Set three parallel convolution branches with convolution kernel sizes of 3×3, 5×5, and 7×7 respectively. Each branch contains two layers of convolution operations, and each layer is followed by batch normalization and an activation function. Taking temperature prediction as an example, input 72-hour historical temperature data (dimension 72×1), and after passing through the three parallel branches, feature maps with dimensions of 70×32, 68×32, and 66×32 are obtained respectively. These feature maps contain local change patterns captured under different receptive fields.

[0027] Calculate the adaptive weights based on the feature discriminability and perform feature fusion. The feature discriminability is measured by calculating the correlation between the feature and the target variable. The specific implementation is as follows: For the features output by each branch, calculate its correlation coefficient with the historical target variable. The higher the correlation coefficient, the more important the feature is for prediction, and a higher weight should be assigned.

[0028] For example, in a certain prediction, the feature discriminabilities of the three branches are 0.65, 0.73, and 0.58 respectively, then the corresponding adaptive weights can be calculated as 0.33, 0.37, and 0.30. Then, perform weighted summation on the features of the three branches according to the weights to obtain the enhanced features after fusion. Experiments prove that compared with simple feature concatenation or average fusion, this adaptive fusion method can improve the prediction accuracy by about 4.2%.

[0029] Capture key information at different time scales based on the multi-scale time attention mechanism. The specific implementation is as follows: Input the enhanced features into three attention units for short-term (6 hours), medium-term (24 hours), and long-term (72 hours) respectively. Each attention unit calculates the importance weights of each time point at the corresponding time scale. The weight values range from 0 to 1, and ensure that the sum of all weights is 1.

[0030] For example, for predicting the temperature in the next 24 hours, the system may assign a total weight of 0.5 to the data in the most recent 6 hours, a weight of 0.3 to the data in the previous 24 hours, and a weight of 0.2 to the earlier data. In this way, the model can focus on the historical time periods most relevant to the prediction task and generate attention-enhanced features. Experiments show that the multi-scale time attention mechanism can improve the prediction accuracy of the model for weather mutation situations by 7.6%.

[0031] Construct an adaptive multi-task loss function to train the deep neural network model. This loss function includes a prediction error loss term and a temporal correlation loss term. The prediction error loss term calculates the difference between the predicted value and the true value using the mean absolute error; the temporal correlation loss term measures the consistency between the predicted sequence and the true sequence in terms of the time variation trend.

[0032] Adaptively adjust the weight coefficients of each loss term according to the difficulty of the prediction task and the characteristics of the data distribution. Specifically, for meteorological elements with drastic changes (such as precipitation), the system increases the weight of the temporal correlation loss term; for relatively stable elements (such as air pressure), it increases the weight of the prediction error loss term. For example, in the temperature prediction task, the weight ratio of the prediction error loss term to the temporal correlation loss term may be 0.7:0.3; while in the precipitation prediction, the weight ratio may be adjusted to 0.4:0.6.

[0033] Input the meteorological data to be predicted into the trained deep neural network model to generate predicted meteorological data. In practical applications, taking the prediction of the temperature in Haidian District, Beijing in the next 24 hours as an example: Input the historical temperature data in the most recent 72 hours. After adaptive time window normalization processing, it passes through the parallel feature extraction layer, adaptive fusion layer, and dynamic attention layer in sequence, and finally outputs the hourly temperature prediction values in the next 24 hours. Experimental results show that the average absolute error of this method in temperature prediction is 0.7°C, improving the prediction accuracy by 23.5% compared with traditional time series models; in the case of extreme weather changes, the prediction accuracy improvement is more significant, reaching 31.8%.

[0034] Through the above technical solutions, the present invention can effectively capture the multi-scale features and time dependence relationships of meteorological data, achieve high-precision prediction of future meteorological data, and provide strong support for fields such as meteorological forecasting, agricultural production, disaster prevention and mitigation.

[0035] In an optional implementation, based on the predicted meteorological data, calculating the solar azimuth and solar altitude angle of the target photovoltaic power station area within the future preset time period includes: Performing time series decomposition on the predicted meteorological data to extract time-varying features and periodic features; performing data correction on the predicted meteorological data based on the time-varying features and the periodic features to obtain corrected predicted data; Calculate the atmospheric refraction correction value based on the predicted air pressure data, predicted temperature data and site altitude in the corrected predicted data; calculate the atmospheric scattering correction value based on the predicted relative humidity data, predicted cloud cover data and site altitude in the corrected predicted data; The atmospheric refraction correction amount and the atmospheric scattering correction amount are weighted and combined, and the weight coefficient is dynamically adjusted according to the time-varying characteristics of the corrected prediction data to obtain a comprehensive correction coefficient; based on the comprehensive correction coefficient, the initial solar azimuth angle and the initial solar altitude angle of the target photovoltaic power station area in a future preset time period are calculated; The initial solar azimuth angle and the initial solar altitude angle are calculated to be mutually constrained according to the geometric relationship between the solar azimuth angle and the solar altitude angle; the calculation result is corrected based on the historical prediction error to obtain the solar azimuth angle and the solar altitude angle of the target photovoltaic power station area within the future preset time period.

[0036] The obtained forecast meteorological data is decomposed into time series to extract time-varying and periodic features. Specifically, the seasonal trend decomposition method is used to decompose the forecast meteorological data of the target photovoltaic power station area into three parts: trend term, seasonal term and residual term. Among them, the trend term represents the long-term change trend of the data as a time-varying feature; the seasonal term represents the periodic change of the data as a periodic feature; and the residual term represents random fluctuations. For example, for the temperature forecast data of a target photovoltaic power station area for a week, after decomposition, the daily variation periodic characteristics of gradually increasing daytime temperature and decreasing nighttime temperature, as well as the time-varying characteristics of gradually warming within a week can be obtained.

[0037] The forecasted meteorological data is corrected based on the extracted time-varying features and periodic features. Specifically, the historical meteorological data for the same period is compared with the forecasted meteorological data, the deviation value is calculated, and then the forecasted data is corrected according to the statistical distribution of the deviation value. For example, if the historical data shows that the temperature forecast value of the forecast system from 14:00 to 16:00 every day is usually 2°C higher than the actual value, the temperature forecast value for this time period is corrected by subtracting 2°C. In this way, more accurate corrected forecast data can be obtained.

[0038] Calculate the atmospheric refraction correction based on the predicted barometric pressure data, predicted temperature data, and station elevation in the corrected predicted data. Atmospheric refraction causes a deviation between the celestial position observed by the observer and the actual position. In practical applications, when the elevation is 500 meters, the barometric pressure is 101.3 kPa, and the temperature is 25 °C, the calculated atmospheric refraction correction is approximately 0.08 degrees, and this correction needs to be applied to the calculation of the solar altitude angle.

[0039] Calculate the atmospheric scattering correction based on the predicted relative humidity data, predicted cloud cover data, and station terrain elevation in the corrected predicted data. Atmospheric scattering causes sunlight to scatter when passing through the atmosphere, affecting the direct amount of solar radiation reaching the ground. For example, when the relative humidity is 60%, the cloud cover is 3 oktas, and the station terrain elevation is 200 meters, the calculated atmospheric scattering correction is approximately 0.12 degrees.

[0040] Perform a weighted combination of the atmospheric refraction correction and the atmospheric scattering correction to form a comprehensive correction coefficient. The weight coefficient is dynamically adjusted according to the time-varying characteristics of the corrected predicted data. For example, under clear weather conditions, the weight of the refraction correction can be set to 0.6, and the weight of the scattering correction can be set to 0.4; under cloudy weather conditions, the weight of the refraction correction can be adjusted to 0.4, and the weight of the scattering correction can be adjusted to 0.6. Through this dynamic adjustment method, a more realistic comprehensive correction coefficient can be obtained.

[0041] Calculate the initial solar azimuth and initial solar altitude angles in the target photovoltaic power station area within a preset future time period based on the comprehensive correction coefficient. For example, for a certain photovoltaic power station located at 30°N and 120°E, at 12:00 on a certain day, the calculated initial solar azimuth is 178.5 degrees, and the initial solar altitude angle is 68.3 degrees.

[0042] Perform mutual constraint calculations on the initial calculation results according to the geometric relationship between the solar azimuth and the solar altitude angle. There is a geometric constraint relationship between the solar azimuth and the solar altitude angle, and it is necessary to ensure that the calculation results of the two are coordinated with each other. Specifically, through the joint verification of parameters such as the declination angle, hour angle, and geographical latitude, ensure that the calculation results of the two angles satisfy the spherical geometric relationship. When it is found that the constraint relationship is not satisfied, adjust the other parameter according to the parameter with higher confidence. For example, if the solar altitude angle calculated based on the geographical location and time has a higher confidence, slightly adjust the azimuth angle to make the two satisfy the geometric constraint relationship.

[0043] The calculation result is corrected based on historical prediction errors. By analyzing the solar angle prediction errors at the same location under similar weather conditions over a past period of time, an error correction model is established. For example, if it is statistically found that under cloudy weather conditions, the predicted value of the solar altitude angle is usually 0.5 degrees lower than the actual value, then this systematic error is compensated in the prediction result. In this way, the solar azimuth angle and solar altitude angle in the target photovoltaic power station area within a preset future time period are finally obtained.

[0044] In the practical application of a certain photovoltaic power station, after calculating the solar angles for the next 24 hours using the above method, the prediction results are compared with the actual observed values. The results show that compared with the traditional method, the average error of the solar azimuth angle calculated by this method is reduced from 1.2 degrees to 0.5 degrees, and the average error of the solar altitude angle is reduced from 0.8 degrees to 0.3 degrees. The prediction accuracy is improved by more than 60%, effectively improving the power generation efficiency and the prediction accuracy of power generation in the photovoltaic power station.

[0045] Figure 2 The combined effect diagram of the atmospheric refraction correction amount and the scattering correction amount for the embodiments of the present invention is as follows: This figure shows the performance comparison of different photovoltaic power generation correction schemes within a complete working day (6:00 - 18:00). The figure contains four curves, respectively representing the time variation trends of the atmospheric refraction correction amount (the curve with circular markers), the atmospheric scattering correction amount (the curve with triangular markers), the comprehensive correction coefficient of this technical solution (the curve with square markers), and the single correction coefficient of the traditional method (the curve with cross markers). From the data performance, all correction coefficients show a "U" - shaped distribution characteristic, reaching the minimum value at noon and being larger in the early morning and evening periods. Specifically, the comprehensive correction coefficient of this technical solution is approximately 0.62 at 6:00 in the morning, then gradually decreases, reaches the lowest value of approximately 0.21 at 12:00, and then starts to rise, returning to approximately 0.65 at 18:00. The variation range of the atmospheric refraction correction amount is between 0.15 - 0.42, with the lowest value at noon being approximately 0.15; the atmospheric scattering correction amount is smaller, with a variation range between 0.09 - 0.29 and approximately 0.09 at noon. The single correction coefficient of the traditional method generally performs between the atmospheric refraction and the comprehensive correction of this solution, being approximately 0.50 at 6:00 in the morning, dropping to 0.21 at noon, and rising to 0.52 at 18:00 in the evening. Through comparison, it can be seen that this technical solution provides a more comprehensive correction ability by considering the combined effects of atmospheric refraction and scattering, especially showing obvious advantages in key periods such as sunrise and sunset.

[0046] In an alternative embodiment, a motion constraint model of the photovoltaic tracking bracket is established according to the cloud cover rate and wind speed data in the predicted meteorological data, and it is determined whether the cloud cover rate is greater than a preset constraint threshold and whether the wind speed data is greater than a safe wind speed threshold. When the judgment conditions are met, calculating the safe avoidance angle of the photovoltaic tracking bracket and controlling the photovoltaic tracking bracket to rotate to the safe avoidance angle includes: The motion constraint model includes wind load constraint and moment balance constraint; a trigger signal is generated based on the magnitude relationship between the cloud cover rate and the preset constraint threshold and the magnitude relationship between the wind speed data and the safe wind speed threshold; Obtain the structural parameters of the photovoltaic tracking bracket, where the structural parameters include the bracket area, the bracket mass, the current bracket inclination angle, and the distance from the bracket center of gravity to the rotation axis; substitute the wind speed data, the bracket area, and the current bracket inclination angle into the wind load constraint, calculate the wind load value, and generate a wind moment based on the wind load value; Calculate the gravity moment according to the bracket mass and the distance from the bracket center of gravity to the rotation axis; substitute the wind moment and the gravity moment into the moment balance constraint, and calculate the safe avoidance angle that satisfies the motion constraint model; Judge whether the avoidance control condition is met according to the trigger signal. When the avoidance control condition is met, obtain the current angle of the photovoltaic tracking bracket; calculate the target angular velocity based on the safe avoidance angle and the current angle, and control the photovoltaic tracking bracket to rotate to the safe avoidance angle according to the target angular velocity.

[0047] Obtain the cloud cover rate and wind speed data in the predicted meteorological data. The meteorological data can be obtained in real time from a meteorological service platform. For example, future 24-hour meteorological prediction data, including the cloud cover rate (in percentage form) and wind speed data (in meters per second) per hour, can be obtained from the meteorological bureau or a third-party meteorological service provider through an API interface.

[0048] Establish a motion constraint model of the photovoltaic tracking bracket. This model includes two parts: wind load constraint and moment balance constraint.

[0049] The wind load constraint mainly considers the force exerted by the wind on the photovoltaic bracket. The wind load value is proportional to the square of the wind speed, the bracket area, and the sine value of the bracket inclination angle. Specifically, the wind load value is equal to half of the product of the square of the wind speed, the bracket area, the sine value of the inclination angle, the wind resistance coefficient (usually taken as 0.6), and the air density (usually taken as 1.225 kg / m³). For example, when the wind speed is 15 m / s, the bracket area is 20 m², and the current bracket inclination angle is 30 degrees, the calculated wind load value is approximately 1378.1 Newtons.

[0050] The moment balance constraint considers the balance relationship between the wind moment and the gravity moment. The wind moment is equal to the value of the wind load multiplied by the perpendicular distance from the center of the support to the rotation axis. The gravity moment is equal to the mass of the support multiplied by the acceleration due to gravity multiplied by the horizontal distance from the center of gravity of the support to the rotation axis. The safety evacuation angle should be such that the wind moment is less than or equal to the gravity moment to ensure the stability of the system.

[0051] Judge whether the evacuation conditions are met. Set a preset constraint threshold (e.g., cloud cover rate is 80%) and a safety wind speed threshold (e.g., 10 m / s). When it is detected that the cloud cover rate is greater than the preset constraint threshold and the wind speed data is greater than the safety wind speed threshold, a trigger signal is generated, indicating that evacuation control needs to be executed.

[0052] Obtain the structural parameters of the photovoltaic tracking support, including the support area (e.g., 20 square meters), the support mass (e.g., 500 kg), the current support inclination angle (e.g., 35 degrees), and the distance from the center of gravity of the support to the rotation axis (e.g., 0.8 m).

[0053] Substitute the wind speed data, the support area, and the current support inclination angle into the wind load constraint. Assume the current wind speed is 12 m / s, the support area is 20 square meters, and the current support inclination angle is 35 degrees. Substitute these values into the wind load constraint formula to calculate the value of the wind load, which is approximately 982.7 Newtons.

[0054] Generate the wind moment based on the value of the wind load. Assume the perpendicular distance from the center of the support to the rotation axis is 0.5 m, then the wind moment is approximately 491.4 Newton - meters.

[0055] Calculate the gravity moment according to the support mass and the distance from the center of gravity of the support to the rotation axis. Assume the support mass is 500 kg, and the horizontal distance from the center of gravity of the support to the rotation axis is 0.8 m multiplied by the cosine of the current inclination angle, i.e., 0.8×cos(35°)=0.655 m, then the gravity moment is equal to 500×9.8×0.655 = 3209.5 Newton - meters.

[0056] Substitute the wind moment and the gravity moment into the moment balance constraint to calculate the safety evacuation angle that satisfies the motion constraint model. In this example, when the support inclination angle increases, the wind moment decreases and the gravity moment also decreases. By adjusting the support inclination angle, the maximum angle at which the wind moment is less than the gravity moment is the safety evacuation angle. After calculation, when the support angle is adjusted to 75 degrees, the wind moment drops to approximately 260 Newton - meters, and the gravity moment is 1270 Newton - meters. At this time, the system is stable, and it can be determined that 75 degrees is the safety evacuation angle.

[0057] Judge whether the evacuation control conditions are met according to the trigger signal. When the evacuation control conditions are met, obtain the current angle of the photovoltaic tracking support. Assume the current angle is 35 degrees.

[0058] Calculate the target angular velocity based on the safe - avoidance angle (75 degrees) and the current angle (35 degrees). The angle difference is 40 degrees. If the set adjustment time is 60 seconds, then the target angular velocity is 40 degrees / 60 seconds = 0.67 degrees per second.

[0059] Control the photovoltaic tracking bracket to rotate to the safe - avoidance angle according to the target angular velocity. Control the motor speed so that the bracket rotates at a speed of 0.67 degrees per second. After about 60 seconds, it reaches the safe - avoidance angle position of 75 degrees and maintains this position until the weather conditions improve.

[0060] In practical applications, to ensure the stability and reliability of the system, multi - level wind - speed thresholds and corresponding safe - avoidance angles can be set. For example, when the wind speed is 10 - 15 m / s, the safe - avoidance angle is 60 degrees; when the wind speed is 15 - 20 m / s, the safe - avoidance angle is 75 degrees; when the wind speed exceeds 20 m / s, the safe - avoidance angle is 90 degrees (fully vertical state).

[0061] In addition, according to the meteorological characteristics of different regions and different seasons, the preset constraint thresholds and safe wind - speed thresholds can be dynamically adjusted to meet the operating requirements under different environmental conditions. For example, in coastal areas prone to typhoons, the safe wind - speed threshold can be set lower to carry out early avoidance control.

[0062] Through the method of the present invention, it is possible to effectively prevent the damage of the photovoltaic tracking bracket caused by bad weather, extend the service life of the equipment, and improve the safety and reliability of the photovoltaic power generation system.

[0063] Figure 3 The following is the working flowchart of the safe - avoidance control system for the photovoltaic tracking bracket in the embodiment of the present invention: This is a safety hazard avoidance control flowchart for a photovoltaic tracking bracket, which details the complete control process from obtaining meteorological data to finally executing the hazard avoidance action. First, the system obtains predicted meteorological data, including cloud cover rate and wind speed data, and compares them with preset thresholds. When the cloud cover rate exceeds the preset threshold or the wind speed exceeds the safe wind speed threshold, the hazard avoidance control process is triggered; if not, the normal tracking mode is maintained without triggering the hazard avoidance control. After entering the hazard avoidance control process, the system first establishes a motion constraint model, including two key constraint conditions: wind load constraint and moment balance constraint. Subsequently, the structural parameters of the photovoltaic tracking bracket are obtained, including core parameters such as the bracket area, mass, inclination angle, and the distance from the center of gravity to the rotation axis. Based on these parameters, the system respectively performs wind load calculation and gravitational moment calculation: the wind load calculation mainly considers the influence of the wind moment on the bracket; the gravitational moment calculation is based on the bracket mass and structural parameters. Substitute the results of these two calculations into the moment balance constraint equation to solve for the safety hazard avoidance angle. Finally, in the control execution stage, the system obtains the current angle, calculates the target angular velocity, and controls the bracket to rotate to the safety hazard avoidance angle position. This process ensures the safe operation of the photovoltaic tracking bracket under adverse weather conditions through strict mechanical analysis and real-time control.

[0064] In an alternative embodiment, optimizing and solving the optimal angle sequence of the photovoltaic tracking bracket within the future preset time period according to the solar radiation intensity and temperature data in the predicted meteorological data, in combination with the solar azimuth angle and the solar altitude angle, using an improved dynamic programming algorithm includes: Obtain the motion parameters of the photovoltaic tracking bracket, where the motion parameters include the minimum rotation angle, the maximum rotation angle, and the rotation power consumption coefficient; establish an initial angle state space based on the minimum rotation angle and the maximum rotation angle; Calculate the air mass index according to the solar azimuth angle and the solar altitude angle, and calculate the direct radiation intensity based on the air mass index and the solar radiation intensity; multiply the direct radiation intensity by the rotation power consumption coefficient to obtain the theoretical optimal power generation; Establish a state density function based on the theoretical optimal power generation; adaptively divide the initial angle state space according to the state density function, subdivide the area with a state density higher than the preset density threshold to obtain an optimized angle state space; calculate the maximum power generation corresponding to each angle state in the optimized angle state space; Calculate the temperature correction coefficient according to the temperature data and the temperature coefficient in the predicted meteorological data, multiply the temperature correction coefficient by the maximum power generation to obtain the actual power generation; construct a power generation cost function by weighted combination of the actual power generation and the maximum power generation; In the optimized angular state space, substitute the power generation cost function into the dynamic programming recurrence equation, and obtain the optimal state values at each moment through iterative calculation; generate an optimal angle sequence within a preset future time period based on the optimal state values.

[0065] Obtain the motion parameters of the photovoltaic tracking bracket, including the minimum rotation angle, the maximum rotation angle, and the rotation power consumption coefficient. In this embodiment, the minimum rotation angle of the photovoltaic tracking bracket is -60 degrees, the maximum rotation angle is 60 degrees, and the rotation power consumption coefficient is 0.02 kWh / degree. Based on these parameters, establish an initial angular state space, that is, the angular range from -60 degrees to 60 degrees.

[0066] Calculate the air mass index according to the solar azimuth angle and the solar altitude angle. In practical applications, when the solar altitude angle is 45 degrees, the air mass index is approximately 1.4; when the solar altitude angle drops to 30 degrees, the air mass index increases to approximately 2.0. Taking 10 am on a certain day as an example, assuming the solar altitude angle is 35 degrees and the azimuth angle is 135 degrees, the air mass index can be calculated to be 1.7.

[0067] Calculate the direct radiation intensity based on the air mass index and the predicted solar radiation intensity. For example, when the predicted total radiation intensity is 800 W / m² and the air mass index is 1.7, the direct radiation intensity can be calculated to be approximately 650 W / m² through relevant calculations. Multiply this direct radiation intensity by the rotation power consumption coefficient of 0.02 kWh / degree to obtain the theoretical optimal power generation of 13 kW.

[0068] Establish a state density function based on the theoretical optimal power generation. This function reflects the distribution of power generation at different angular states. In this embodiment, when the angle of the photovoltaic bracket is close to the solar incidence angle, the state density value is high; when the angle deviates greatly, the state density value drops rapidly.

[0069] Adaptively divide the initial angular state space according to the state density function. Set a preset density threshold of 0.8 and subdivide the area where the state density is higher than this threshold. For example, between the angular range of 25 degrees and 40 degrees, the state density is generally higher than 0.8, so this area is subdivided at 1-degree intervals; while in the range from -60 degrees to -20 degrees, the state density is low, and a 5-degree interval is maintained. In this way, an optimized angular state space is obtained, which not only ensures the calculation accuracy of the high-density area but also reduces the calculation amount.

[0070] After completing the optimization of the angle state space, calculate the maximum power generation corresponding to each angle state in the optimized angle state space. Taking a certain moment as an example, when the bracket angle is 35 degrees, considering factors such as the incident angle and the area of the photovoltaic panel, the calculated maximum power generation is 12.5 kilowatts; when the angle is 20 degrees, the maximum power generation drops to 11.2 kilowatts; when the angle is -30 degrees, the maximum power generation is only 5.6 kilowatts.

[0071] Calculate the temperature correction coefficient according to the temperature data and the temperature coefficient in the predicted meteorological data. Assume that the temperature coefficient of the photovoltaic panel is -0.4% / °C. When the ambient temperature is 30°C and the temperature of the photovoltaic panel rises to 45°C, compared with 25°C under the standard test conditions, the temperature has increased by 20°C, and the temperature correction coefficient is 1 - 0.4%×20 = 0.92. Multiply this temperature correction coefficient by the maximum power generation to obtain the actual power generation. For example, when the angle is 35 degrees, the actual power generation is 12.5×0.92 = 11.5 kilowatts.

[0072] Construct a power generation cost function by weighted combination of the actual power generation and the maximum power generation. In this embodiment, the weight coefficients are 0.7 and 0.3 respectively, that is, the cost function value = 0.7×actual power generation + 0.3×maximum power generation. For example, when the angle is 35 degrees, the cost function value is: 0.7×11.5 + 0.3×12.5 = 11.8 kilowatts.

[0073] In the optimized angle state space, substitute the power generation cost function into the dynamic programming recurrence equation, and obtain the optimal state value at each moment through iterative calculation. Assume that the preset time period is 1 day, with an interval of 1 hour, a total of 24 time points. Through recursive calculation, determine the optimal angle state at each time point.

[0074] Generate an optimal angle sequence for the future preset time period based on the optimal state value. For example, the optimal angle sequence from 6 am to 6 pm on a certain day is: -45 degrees (6 am), -35 degrees (7 am), -25 degrees (8 am), -10 degrees (9 am), 5 degrees (10 am), 18 degrees (11 am), 30 degrees (12 pm), 35 degrees (1 pm), 32 degrees (2 pm), 25 degrees (3 pm), 10 degrees (4 pm), -5 degrees (5 pm), -20 degrees (6 pm).

[0075] Through the above steps, the method of the present invention comprehensively considers the change of the sun position, the predicted meteorological data, the temperature influence and the energy consumption of the bracket rotation, and uses an improved dynamic programming algorithm to realize the optimized control of the angle of the photovoltaic tracking bracket. Compared with the traditional method, the power generation of this method has increased by 8.5% in actual tests, reflecting good practical value.

[0076] Figure 4Influence diagram of temperature correction coefficient on power generation efficiency in the embodiments of the present invention: This diagram shows the changing trends of various performance parameters of the photovoltaic system within a complete working day (8:00 - 17:00). There are four curves in the diagram, respectively representing the time-varying relationships of the ambient temperature (the line marked with a cross), the temperature correction coefficient (the line marked with a triangle), the power generation efficiency after correction by the present technical solution (the line marked with a circle), and the uncorrected power generation efficiency of the traditional solution (the line marked with a square). From the data performance, the ambient temperature gradually rises from about 22°C at 8:00 in the morning, reaches the daily maximum value of about 36°C at 13:00, and then starts to decline; the temperature correction coefficient shows an opposite changing trend to the ambient temperature, gradually decreasing from about 0.97 in the morning to about 0.90 during the noon period. The power generation efficiency after correction by the present technical solution generally maintains at a relatively high level, starting from 92% at 8:00 in the morning, reaching a peak of about 95% at 10:00, and remaining above 92% even during the hottest noon period until it significantly drops to about 86% after 16:00. In contrast, the uncorrected power generation efficiency of the traditional solution generally performs poorly, starting from 86% in the morning, although it slightly increases to 88% during the period from 9:00 to 10:00, but significantly drops during the high-temperature noon period, reaching the lowest point of about 82% at 13:00, and only recovering to about 84% even after the temperature drops in the afternoon. This shows that the present technical solution can effectively improve the power generation efficiency of the system through temperature correction, especially having obvious advantages during high-temperature periods.

[0077] In an alternative embodiment, in the optimized angular state space, substituting the power generation power cost function into the dynamic programming recurrence equation, and obtaining the optimal state value at each moment through iterative calculation; generating an optimal angle sequence within a preset future time period based on the optimal state value includes: Calculating the normalized difference between the actual power generation power and the predicted power generation power to obtain the power prediction error; substituting the power prediction error into the exponential decay function and multiplying it by the base discount rate to obtain the time-varying discount factor; adding the negative value of the power generation power cost function to the maximum value of the time-varying discount factor and the state value at the next moment to construct the dynamic programming recurrence equation; Calculating the current angular state value based on the dynamic programming recurrence equation, performing weighted averaging on the current angular state value and the historical angular state value according to a preset learning rate to obtain the updated angular state value; calculating the relative error between the updated angular state value and the historical angular state value, and when the relative error is less than the preset convergence threshold, recording the updated angular state value as the final state value; Select the angle that maximizes the final state value from the optimized angle state space as the optimal angle at this moment; determine whether the difference between the optimal angles at adjacent prediction moments exceeds the preset maximum angle change rate, and if it does, correct the optimal angle based on the maximum angle change rate; connect the corrected optimal angles in sequence to generate the optimal angle sequence of the photovoltaic tracking bracket within a preset future time period.

[0078] In the optimized angle state space, substitute the power generation cost function into the dynamic programming recurrence equation, and obtain the optimal state value at each moment through iterative calculation. Based on the optimal state value, an optimal angle sequence within a preset future time period can be generated.

[0079] The initial angle state space of the photovoltaic system can be set as an angle set at intervals of 5° within the range of 0° to 90°, that is, {0°, 5°, 10°, 15°, ..., 90°}. After considering the actual light conditions, the angle state space can be optimized to a smaller range. For example, during the period from 8 am to 4 pm, it may be optimized to an angle set such as {30°, 35°, 40°, ..., 60°} to improve the calculation efficiency.

[0080] The first step in the dynamic programming solution process is to calculate the normalized difference between the actual power generation and the predicted power generation to obtain the power prediction error. Suppose the current actual power generation is 580W and the predicted power generation is 600W, then the normalized difference is calculated as (580 - 600) / 600 = -0.033, that is, the power prediction error is -3.3%.

[0081] Substitute the power prediction error into the exponential decay function and multiply it by the base discount rate to obtain the time-varying discount factor. Assume that the exponential decay function is in the form of a negative exponent, the base discount rate is 0.95, and the decay coefficient is 2, then the time-varying discount factor is calculated as 0.95×(1 - 0.033)^2≈0.944, which represents the weight coefficient of the current decision on the future.

[0082] Construct the dynamic programming recurrence equation, and add the negative value of the power generation cost function, the time-varying discount factor, and the maximum value of the next moment state value. For example, when the angle is 45°, the power generation cost function value is -590W, the time-varying discount factor is 0.944, and the maximum value of the next moment state value is 620, then the current recurrence value is -590 + 0.944×620≈ -5.72.

[0083] Calculate the current angle state value based on the dynamic programming recurrence equation, and perform weighted averaging on the current angle state value and the historical angle state value according to the preset learning rate to obtain the updated angle state value. Suppose the currently calculated state value is 595, the historical state value is 590, and the learning rate is 0.2. Then the updated state value is 590 + 0.2×(595 - 590) = 591, which represents the comprehensive evaluation value at an angle of 45°.

[0084] Calculate the relative error between the updated angle state value and the historical angle state value. The relative error is calculated as |(591 - 590) / 590| = 0.0017. If the preset convergence threshold is set to 0.001, then at this time the relative error is greater than the threshold, and iterative calculation needs to be continued. Assume that after multiple iterations, the updated state value is 590.5, the historical state value is 590.4, and the relative error is |(590.5 - 590.4) / 590.4| ≈ 0.00017 < 0.001. At this time, the convergence condition is reached, and 590.5 is recorded as the final state value of this angle.

[0085] Repeat the above steps for all angles in the optimized angle state space. For example, the final state values calculated for different angles are: 580.2 for 40°, 590.5 for 45°, and 585.6 for 50°. Select the angle of 45° with the largest final state value as the optimal angle at the current moment.

[0086] Judge whether the difference between the optimal angles at adjacent prediction times exceeds the preset maximum angle change rate. Suppose the optimal angles at two adjacent times are 45° and 60° respectively, the interval time is 10 minutes, and the preset maximum angle change rate is 1° / minute. Then the angle difference of 15° exceeds the maximum allowable change value of 10°, and correction is required. The angle at the latter moment after correction should be 45° + 10° = 55°, rather than the original 60°.

[0087] Connect all the corrected optimal angles in sequence to generate the optimal angle sequence of the photovoltaic tracking bracket within the preset future time period. For example, from 8 am to 4 pm one day, the optimal angle sequence at each half-hour moment may be {35°, 40°, 45°, 50°, 55°, 60°, 55°, 50°, 45°, 40°, 35°, 30°, 25°}. This sequence takes into account the change of the sun's trajectory and system constraints, and can enable the photovoltaic system to obtain the maximum power generation.

[0088] In practical applications, various parameters can be adjusted according to specific system parameters. For example, the calculation of power prediction error can adopt the moving window average to reduce the influence of instantaneous fluctuations; the decay coefficient of the time-varying discount factor can vary according to weather conditions, set to 1.5 on sunny days and 2.5 on cloudy days; the learning rate can be adjusted according to the system response speed, set to 0.1 for large systems and 0.3 for small systems; the convergence threshold can be set according to the accuracy requirements, generally set to 0.0005 for high-precision applications and 0.002 for ordinary applications; the maximum angular change rate can be set according to mechanical properties, set to 2° / minute for high-performance drivers and 1° / minute for ordinary drivers.

[0089] This method effectively solves the optimization problem of photovoltaic tracking control through the dynamic programming algorithm, and realizes generating the optimal angle sequence for the photovoltaic tracking bracket under the conditions of considering real-time power generation, prediction error and system constraints, thus improving the overall power generation efficiency of the system.

[0090] Figure 5 The following is the flowchart of the photovoltaic tracking control method optimized by dynamic programming in the embodiment of the present invention: The power prediction error is obtained by calculating the normalized difference between the actual power generation and the predicted power generation, and this error is substituted into the exponential decay function and multiplied by the base discount rate to obtain the time-varying discount factor. Then, the negative value of the power generation cost function is added to the maximum value of the time-varying discount factor and the state value at the next moment to construct a dynamic programming recurrence equation, laying a foundation for subsequent optimization. Based on the constructed dynamic programming recurrence equation, the state value of the current angle is calculated, and it is weighted and averaged with the historical angle state value according to the preset learning rate to obtain the updated angle state value. Subsequently, the relative error between the updated and historical angle state values is calculated. When this error is less than the preset convergence threshold, the updated angle state value is recorded as the final state value. In the optimization space, the angle that can maximize the final state value at each prediction moment is selected as the optimal angle at that moment, and constraints are imposed by judging whether the difference between the optimal angles at adjacent prediction moments exceeds the preset maximum angular change rate. If it exceeds the threshold, it is corrected based on the maximum angular change rate. Finally, the corrected optimal angles at each moment are connected in sequence to generate a complete optimal angle control sequence.

[0091] In an alternative embodiment, generating the angle adjustment instruction for the photovoltaic tracking bracket according to the optimal angle sequence includes: The motion constraint parameters of the photovoltaic tracking bracket are obtained, wherein the motion constraint parameters include a maximum adjustment speed, a maximum acceleration and a minimum adjustment time interval; based on the motion constraint parameters, the optimal angle sequence is subjected to trajectory smoothing processing to generate a transition angle sequence that satisfies the maximum adjustment speed and the maximum acceleration; the transition angle sequence is subjected to time discretization sampling according to the minimum adjustment time interval, and the transition angle sequence after the discretization sampling is converted into an angle adjustment instruction.

[0092] Obtaining the motion constraint parameters of the photovoltaic tracking bracket is the basic step in generating angle adjustment instructions. The motion constraint parameters mainly include the maximum adjustment speed, maximum acceleration and minimum adjustment time interval. The maximum adjustment speed refers to the maximum angular velocity allowed by the photovoltaic tracking bracket during the adjustment process. It is usually determined based on the characteristics of the bracket's drive motor and the mechanical structure strength. In practical applications, the typical value is set to 1.5 degrees / second. The maximum acceleration refers to the maximum angular acceleration allowed by the bracket during the angle adjustment process. This parameter is related to the smoothness of the bracket adjustment process and the mechanical stress condition. The typical value is set to 0.3 degrees / second 2 The minimum adjustment time interval refers to the minimum time required between two angle adjustment instructions. This parameter takes into account the processing capacity and execution efficiency of the control system. The typical value is set to 5 seconds. These parameters can be read through the bracket system configuration file and can also be dynamically adjusted according to the actual operating environment.

[0093] Based on the motion constraint parameters, the trajectory of the optimal angle sequence is smoothed to generate a transition angle sequence that meets the maximum adjustment speed and maximum acceleration. The trajectory smoothing adopts the cubic spline interpolation method to build a continuous curve between adjacent angle points to ensure that the angle change transitions smoothly within the speed and acceleration constraints. In the specific implementation, the angle difference is first calculated for adjacent angle points in the optimal angle sequence, and then the minimum transition time is determined according to the maximum adjustment speed. When the calculated transition time is less than the minimum time calculated according to the acceleration constraint, the acceleration-uniform speed-deceleration adjustment mode is adopted; when the calculated transition time is greater than the minimum time calculated according to the acceleration constraint, the acceleration-deceleration mode is directly adopted without going through the uniform speed stage.

[0094] Assume that the two adjacent angle points in the optimal angle sequence are 30 degrees and 45 degrees respectively, and the angle difference is 15 degrees. According to the maximum adjustment speed of 1.5 degrees / second, it takes at least 10 seconds to complete the adjustment; and according to the maximum acceleration of 0.3 degrees / second 2Calculation shows that it takes 5 seconds to accelerate to the maximum speed and the same 5 seconds for deceleration, totaling 10 seconds. In this case, the acceleration - deceleration mode is adopted without a constant - speed segment. The obtained transition angle sequence after trajectory smoothing is a set of angle values that smoothly transition from 30 degrees to 45 degrees within 10 seconds, ensuring that the angular velocity does not exceed 1.5 degrees per second and the angular acceleration does not exceed 0.3 degrees per second during the whole process. 2 。

[0095] The transition angle sequence is discretely sampled in time according to the minimum adjustment time interval, and the discretely sampled transition angle sequence is converted into angle adjustment instructions. Time discretization sampling is to sample the continuous transition angle sequence at the time intervals that the control system can handle, converting it into a discrete sequence of angle points. The sampling time interval needs to meet the requirements of the minimum adjustment time interval and also consider the real - time requirements of the control system. In practical applications, if the minimum adjustment time interval is 5 seconds, then from the continuous transition angle sequence after smoothing, an angle value is extracted every 5 seconds to form a discrete angle sequence.

[0096] Converting the discretely sampled transition angle sequence into angle adjustment instructions includes two aspects: instruction format construction and instruction queue management. Instruction format construction is to encapsulate the angle value into an instruction format recognizable by the control system, usually including information such as timestamp, target angle, execution priority, etc. Instruction queue management is responsible for storing the constructed instructions into the instruction queue in the execution order for the control system to execute sequentially.

[0097] Assuming the minimum adjustment time interval is 5 seconds, then through discretization sampling, 3 angle adjustment points are obtained: the angle is 30 degrees at 0 seconds, 37.5 degrees at 5 seconds, and 45 degrees at 10 seconds. Converting these angle values into angle adjustment instructions, the following instruction sequence is formed: [timestamp_1, angle: 30 degrees, priority: normal], [timestamp_2, angle: 37.5 degrees, priority: normal], [timestamp_3, angle: 45 degrees, priority: normal]. These instructions will be sent to the actuator in sequence to drive the photovoltaic tracking bracket to adjust the angle according to the predetermined trajectory.

[0098] In a more complex application scenario, the optimal angle sequence may contain multiple angle points, and the interval between adjacent angle points may be very small. For example, the optimal angle sequence of a certain photovoltaic power station within a day contains 48 angle points from east to west, with an average of one angle point every 30 minutes. When directly generating adjustment instructions based on these angle points, it may cause the bracket to adjust frequently, increasing energy consumption and shortening the equipment life. Through trajectory smoothing processing and time discretization sampling, these 48 angle points can be converted into a smooth transition sequence that meets the motion constraints, reducing unnecessary adjustment times and optimizing the bracket motion trajectory.

[0099] In practical applications, this method significantly improves the angle adjustment efficiency and stability of the photovoltaic tracking bracket. By obtaining the motion constraint parameters in real time, dynamically calculating the transition angle sequence, and performing time-discretized sampling, the photovoltaic tracking bracket can move along the optimal trajectory, reducing mechanical stress, extending the equipment life, and ensuring the efficient operation of the photovoltaic power generation system at the same time.

[0100] In a second aspect of the embodiments of the present invention, there is provided an electronic device, including: A processor; A memory for storing instructions executable by the processor; Wherein, the processor is configured to call the instructions stored in the memory to execute the method described above.

[0101] In a third aspect of the embodiments of the present invention, there is provided a computer-readable storage medium, on which computer program instructions are stored, and when the computer program instructions are executed by a processor, the method described above is implemented.

[0102] The present invention may be a method, an apparatus, a system, and / or a computer program product. The computer program product may include a computer-readable storage medium, on which computer-readable program instructions for executing various aspects of the present invention are loaded.

[0103] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit them; although the present invention has been described in detail with reference to the foregoing embodiments, those of ordinary skill in the art should understand that they can still modify the technical solutions described in the foregoing embodiments, or perform equivalent replacements for some or all of the technical features; and these modifications or replacements do not make the essence of the corresponding technical solutions deviate from the scope of the technical solutions of the embodiments of the present invention.

Claims

1. A method for optimizing the angle of a photovoltaic tracking bracket for meteorological prediction and improving power generation efficiency, characterized in that Including: Receiving meteorological data of the target photovoltaic power station area collected by a meteorological monitoring station, constructing a deep neural network model according to the meteorological data, and using the deep neural network model to predict meteorological data within a preset future time period to obtain predicted meteorological data; Based on the predicted meteorological data, calculating the solar azimuth angle and solar altitude angle of the target photovoltaic power station area within the preset future time period; According to the cloud cover rate and wind speed data in the predicted meteorological data, establishing a motion constraint model of the photovoltaic tracking bracket, judging whether the cloud cover rate is greater than a preset constraint threshold and whether the wind speed data is greater than a safe wind speed threshold, and when the judgment condition is met, calculating the safe avoidance angle of the photovoltaic tracking bracket and controlling the photovoltaic tracking bracket to rotate to the safe avoidance angle; According to the solar radiation intensity and temperature data in the predicted meteorological data, combined with the solar azimuth angle and the solar altitude angle, using an improved dynamic programming algorithm to optimize and solve the optimal angle sequence of the photovoltaic tracking bracket within the preset future time period, wherein the optimal angle sequence is used to maximize the power generation efficiency of the photovoltaic module; According to the optimal angle sequence, generating an angle adjustment instruction for the photovoltaic tracking bracket and sending the angle adjustment instruction to the controller of the photovoltaic tracking bracket to control the photovoltaic tracking bracket to perform angle adjustment according to the optimal angle sequence.

2. The method according to claim 1, characterized in that, Constructing a deep neural network model according to the meteorological data and using the deep neural network model to predict meteorological data within a preset future time period to obtain predicted meteorological data includes: Normalizing the meteorological data according to an adaptive time window length to obtain normalized features; The deep neural network model includes a parallel feature extraction layer, an adaptive fusion layer and a dynamic attention layer; inputting the normalized features into the parallel feature extraction layer, and extracting multi-scale local features through a spatial pyramid convolution branch; inputting the multi-scale local features into the adaptive fusion layer, calculating an adaptive weight based on the feature discriminability and performing feature fusion to obtain enhanced features; inputting the enhanced features into the dynamic attention layer, and capturing key information at different time scales based on a multi-scale time attention mechanism to obtain attention-enhanced features; Constructing an adaptive multi-task loss function, where the adaptive multi-task loss function includes a prediction error loss term and a temporal correlation loss term; adaptively adjusting the weight coefficients of each loss term according to the difficulty of the prediction task and the data distribution characteristics; using the adaptive multi-task loss function to perform end-to-end training on the deep neural network model; Inputting the meteorological data to be predicted into the trained deep neural network model to generate predicted meteorological data.

3. The method according to claim 1, wherein Based on the predicted meteorological data, calculating the solar azimuth angle and solar altitude angle of the target photovoltaic power station area within the preset future time period includes: Performing time series decomposition on the predicted meteorological data, and extracting time-varying features and periodic features; performing data correction on the predicted meteorological data based on the time-varying features and the periodic features to obtain corrected predicted data; Calculate the atmospheric refraction correction amount based on the predicted barometric pressure data, predicted temperature data, and site altitude in the corrected predicted data; calculate the atmospheric scattering correction amount based on the predicted relative humidity data, predicted cloud cover data, and site terrain height in the corrected predicted data; Perform weighted combination on the atmospheric refraction correction amount and the atmospheric scattering correction amount, dynamically adjust the weight coefficient according to the time-varying characteristics of the corrected predicted data, and obtain a comprehensive correction coefficient; calculate the initial solar azimuth angle and initial solar altitude angle in the target photovoltaic power station area within a preset future time period based on the comprehensive correction coefficient; Perform mutual constraint calculation on the initial solar azimuth angle and the initial solar altitude angle according to the geometric relationship between the solar azimuth angle and the solar altitude angle; correct the calculation result based on the historical prediction error to obtain the solar azimuth angle and solar altitude angle in the target photovoltaic power station area within the preset future time period.

4. The method according to claim 1, wherein Based on the cloud cover rate and wind speed data in the predicted meteorological data, establish a motion constraint model for the photovoltaic tracking bracket, and determine whether the cloud cover rate is greater than a preset constraint threshold and whether the wind speed data is greater than the safe wind speed threshold. When the judgment condition is satisfied, calculate the safe avoidance angle of the photovoltaic tracking bracket, and control the photovoltaic tracking bracket to rotate to the safe avoidance angle, including: The motion constraint model includes wind load constraint and moment balance constraint; generate a trigger signal based on the magnitude relationship between the cloud cover rate and the preset constraint threshold and the magnitude relationship between the wind speed data and the safe wind speed threshold; Obtain the structural parameters of the photovoltaic tracking bracket, where the structural parameters include the bracket area, bracket mass, current bracket inclination angle, and distance from the bracket center of gravity to the rotating shaft; substitute the wind speed data, the bracket area, and the current bracket inclination angle into the wind load constraint, calculate the wind load value, and generate a wind moment based on the wind load value; Calculate the gravity moment according to the bracket mass and the distance from the bracket center of gravity to the rotating shaft; substitute the wind moment and the gravity moment into the moment balance constraint to calculate the safe avoidance angle that satisfies the motion constraint model; Judge whether the avoidance control condition is satisfied according to the trigger signal. When the avoidance control condition is satisfied, obtain the current angle of the photovoltaic tracking bracket; calculate the target angular velocity based on the safe avoidance angle and the current angle, and control the photovoltaic tracking bracket to rotate to the safe avoidance angle according to the target angular velocity.

5. The method according to claim 1, characterized in that, Based on the solar radiation intensity and temperature data in the predicted meteorological data, combined with the solar azimuth angle and the solar altitude angle, use an improved dynamic programming algorithm to optimize and solve the optimal angle sequence of the photovoltaic tracking bracket within the preset future time period, including: Obtain the motion parameters of the photovoltaic tracking bracket, where the motion parameters include the minimum rotation angle, the maximum rotation angle, and the rotation power consumption coefficient; establish an initial angle state space based on the minimum rotation angle and the maximum rotation angle; Calculate the air mass index according to the solar azimuth angle and the solar altitude angle, and calculate the direct radiation intensity based on the air mass index and the solar radiation intensity; multiply the direct radiation intensity by the rotation power consumption coefficient to obtain the theoretical optimal power generation; Establish a state density function based on the theoretical optimal power generation; adaptively divide the initial angle state space according to the state density function, and subdivide the area where the state density is higher than the preset density threshold to obtain the optimized angle state space; calculate the maximum power generation corresponding to each angle state in the optimized angle state space; Calculate the temperature correction coefficient according to the temperature data and the temperature coefficient in the predicted meteorological data, and multiply the temperature correction coefficient by the maximum power generation to obtain the actual power generation; weight and combine the actual power generation and the maximum power generation to construct a power generation cost function; In the optimized angle state space, substitute the power generation cost function into the dynamic programming recurrence equation, and obtain the optimal state value at each moment through iterative calculation; generate an optimal angle sequence within a preset future time period based on the optimal state value.

6. The method according to claim 5, characterized in that, In the optimized angle state space, substitute the power generation cost function into the dynamic programming recurrence equation, and obtain the optimal state value at each moment through iterative calculation; Generating an optimal angle sequence within a preset future time period based on the optimal state value includes: Calculate the normalized difference between the actual power generation and the predicted power generation to obtain the power prediction error; substitute the power prediction error into the exponential decay function and multiply it by the base discount rate to obtain the time-varying discount factor; add the negative value of the power generation cost function to the maximum value of the time-varying discount factor and the state value at the next moment to construct the dynamic programming recurrence equation; Calculate the current angle state value based on the dynamic programming recurrence equation, and perform weighted average on the current angle state value and the historical angle state value according to the preset learning rate to obtain the updated angle state value; calculate the relative error between the updated angle state value and the historical angle state value, and when the relative error is less than the preset convergence threshold, record the updated angle state value as the final state value; Select the angle that maximizes the final state value from the optimized angle state space as the optimal angle at this moment; judge whether the difference between the optimal angles at adjacent prediction moments exceeds the preset maximum angle change rate, and if it exceeds, correct the optimal angle based on the maximum angle change rate; connect the corrected optimal angles in sequence to generate the optimal angle sequence of the photovoltaic tracking bracket within a preset future time period.

7. The method according to claim 1, characterized in that Generating an angle adjustment command for the photovoltaic tracking bracket according to the optimal angle sequence includes: Obtain the motion constraint parameters of the photovoltaic tracking bracket, where the motion constraint parameters include the maximum adjustment speed, the maximum acceleration, and the minimum adjustment time interval; perform trajectory smoothing on the optimal angle sequence based on the motion constraint parameters to generate a transition angle sequence that satisfies the maximum adjustment speed and the maximum acceleration; perform time discretization sampling on the transition angle sequence according to the minimum adjustment time interval, and convert the discretization sampled transition angle sequence into an angle adjustment instruction.

8. An electronic device, characterized in that, Comprising: A processor; A memory for storing instructions executable by the processor; Wherein, the processor is configured to call the instructions stored in the memory to execute the method according to any one of claims 1 to 7.

9. A computer-readable storage medium having computer program instructions stored thereon, characterized in that, When the computer program instructions are executed by the processor, the method according to any one of claims 1 to 7 is implemented.

Images