Solar tracking holder dynamic adjustment method and system based on sun trajectory prediction

Abstract

The invention provides a solar tracking holder dynamic adjustment method and system based on sun trajectory prediction, and relates to the technical field of photovoltaic power generation, and the method specifically comprises the steps: obtaining the positioning data, time data and environment data of a solar panel; calculating to obtain sun position data through a sun position algorithm; substituting the sun position data and the environment data into a weight fusion formula, and calculating to obtain an environment influence factor; future environment data is predicted, and a holder control strategy is adjusted; inputting the sun position data, the environmental data and the environmental impact factors into the long short-term memory network model, outputting a predicted future sun trajectory, calculating a prediction error, and correcting data in the long short-term memory network model and / or a medium and short-term prediction algorithm; and optimizing the future sun trajectory and the pan-tilt control strategy, and calculating and adjusting the pitch angle and azimuth angle of the pan-tilt. According to the invention, the prediction precision of the sun position and environmental data is improved, the dynamic adjustment capability of the holder is enhanced, and the method can accurately adapt to the dynamic environment.

Description

Technical Field

[0001] The present invention relates to the technical field of photovoltaic power generation, and particularly to a dynamic adjustment method and system for a solar tracking cloud platform based on solar trajectory prediction. Background Art

[0003] Traditional cloud platform tracking methods mainly consider the collection of solar energy, but pay less attention to environmental impact and energy loss, and it is difficult to accurately adapt to the dynamic environment while improving power generation efficiency. Summary of the Invention

[0004] The purpose of the present invention is to provide a dynamic adjustment method and system for a solar tracking cloud platform based on solar trajectory prediction, so as to solve the problem that the existing technology in the above-mentioned background art is difficult to accurately adapt to the dynamic environment while improving power generation efficiency.

[0005] To achieve the above purpose, the present invention provides the following technical solution: A dynamic adjustment method for a solar tracking cloud platform based on solar trajectory prediction, the steps include: obtaining the positioning data, time data and environmental data of the solar panel; calculating the solar position data according to the positioning data and the time data through the solar position algorithm; substituting the solar position data and the environmental data into the weight fusion formula to calculate the environmental impact factor; predicting the future environmental data by using the medium and short-term prediction algorithm, and adjusting the cloud platform control strategy through a preset threshold; establishing a long short-term memory network model, inputting the solar position data, environmental data and the environmental impact factor into the long short-term memory network model, outputting the predicted future solar trajectory, calculating the prediction error and correcting the data in the long short-term memory network model and / or the medium and short-term prediction algorithm; optimizing the future solar trajectory and the cloud platform control strategy, and calculating and adjusting the pitch angle and azimuth angle of the cloud platform according to the optimized result.

[0006] Optionally, the step of optimizing the future solar trajectory and the cloud platform control strategy includes: optimizing the cloud platform control strategy and the future solar trajectory through the particle swarm optimization algorithm.

[0007] Optionally, the step of obtaining the positioning data, time data and environmental data of the solar panel specifically includes: obtaining the longitude, latitude and altitude of the solar panel, obtaining the current time, obtaining the light intensity, temperature, humidity and wind speed; the solar position data includes the solar altitude angle and the solar azimuth angle.

[0008] Optionally, after the step of calculating the solar position data, the following steps are further included: calculating the theoretical light intensity through the solar position data, collecting the real-time light intensity through a light intensity sensor, and correcting the light intensity by means of a dynamic weight method. The calculation formula is: I abjusted (h, A) = w1I predicted + w2I actual , w2 = 1 - w1; In the formula, the I abjusted (h, A) is the corrected light intensity, the I predicted is the theoretical light intensity, the I actual is the real-time light intensity, the I max is the maximum light intensity, the w1 is the weight of the theoretical light intensity, and the w2 is the weight of the real-time light intensity.

[0009] Optionally, the step of substituting the solar position data and the environmental data into the weight fusion formula to calculate the environmental impact factor specifically includes: calculating the environmental impact factor by using the weight fusion formula. The calculation formula is: E(h, A) = αI abjusted (h, A)+βH + γT + δW, In the formula, the E(h, A) is the environmental impact factor, the I abjusted (h, A) is the corrected light intensity, the h, A is the solar position data, the H is the humidity, the T is the temperature, the W is the wind speed, the α is the weight of the corrected light intensity, the β is the humidity weight, the β0 is the reference humidity weight, the γ is the temperature weight, the γ0 is the reference temperature weight, the δ is the wind speed weight, and the δ0 is the reference wind speed weight.

[0010] Optionally, the step of predicting the future environmental data by using the medium- and short-term prediction algorithm specifically includes: performing real-time prediction on the future environmental data through the Kalman filter algorithm, and performing medium- and short-term prediction on the future environmental data through the autoregressive integrated moving average model; calculating the environmental change rate V and the historical environmental stability σ 2 , and setting the environmental change rate threshold V th and the historical environmental stability threshold

[0011] Optionally, the step of adjusting the pan-tilt control strategy through a preset threshold specifically includes: if V < V th and then select the autoregressive integrated moving average model to perform medium- and short-term prediction on the future environmental data. If V < V th and then select the Kalman filter algorithm to perform real-time prediction on the future environmental data. If V ≥ V thAnd Then select the Kalman filter algorithm to perform real-time prediction on future environmental data. If V ≥ V th And Then select the Kalman filter algorithm to perform real-time prediction on future environmental data.

[0012] Optionally, the steps of predicting the future solar trajectory specifically include: combining solar position data and environmental impact factors as the input vector of the long short-term memory network model, and its expression is: X t =[h t , A t JD t φ, λ, E t , where the X t is the model input vector, the h t is the solar altitude angle, the A t is the solar azimuth angle, the JD t is the current time, the φ is the latitude, the λ is the longitude, and the E t is the environmental impact factor; the long short-term memory network model outputs the predicted value of the future solar trajectory by learning N historical data, and its calculation formula is: where the is the predicted solar altitude angle, the is the predicted solar azimuth angle, and (X t , X t-1 ,..., X t-N ) are historical data.

[0013] Optionally, the steps of calculating the prediction error and correcting the data in the long short-term memory network model and / or the medium-term prediction algorithm specifically include: comparing the predicted solar altitude angle and the predicted solar azimuth angle with the actual measured values, and correcting the future solar trajectory, and its calculation formula is: where the L is the mean square error loss function, the N is the total number of samples, the h i is the true value of the solar altitude angle of the i-th sample, the is the predicted value of the solar altitude angle of the i-th sample, the A i is the true value of the solar azimuth angle of the i-th sample, and the is the predicted value of the solar azimuth angle of the i-th sample; during the training process of the long short-term memory network model, the Adam optimizer is used to adjust the weights, and the loss is minimized by the gradient descent method, and the error range is calculated. If the error range is higher than the preset error range threshold, the data in the long short-term memory network model and / or the medium-term algorithm is adjusted.

[0014] On the other hand, the present invention also provides a dynamic adjustment system for a solar tracking pan-tilt based on solar trajectory prediction, including: an acquisition module for acquiring positioning data, time data, and environmental data of a solar panel; a solar position calculation module for calculating solar position data through a solar position algorithm according to the positioning data and the time data; an environmental impact factor calculation module for substituting the solar position data and the environmental data into a weight fusion formula to calculate an environmental impact factor; a pan-tilt control strategy module for predicting future environmental data by using a medium- and short-term prediction algorithm and adjusting the pan-tilt control strategy through a preset threshold; a future solar trajectory prediction module for establishing a long short-term memory network model, inputting the solar position data, environmental data, and the environmental impact factor into the long short-term memory network model, outputting a predicted future solar trajectory, calculating a prediction error, and correcting data in the long short-term memory network model and / or the medium- and short-term prediction algorithm; and a pan-tilt adjustment module for optimizing the future solar trajectory and the pan-tilt control strategy, and calculating and adjusting the pitch angle and azimuth angle of the pan-tilt according to the optimized result.

[0015] Compared with the prior art, the beneficial effects of the present invention are as follows:

[0016] In this application, solar position data and an environmental impact factor are calculated, future environmental data is predicted by a medium- and short-term prediction algorithm, and the pan-tilt control strategy is adjusted. Multiple data such as solar position data, environmental data, and the environmental impact factor are input into a long short-term memory network model to predict the future solar trajectory, and the data in the long short-term memory network model and / or the medium- and short-term prediction algorithm is corrected according to the error result, which can improve the prediction accuracy of solar position and environmental data, enhance the dynamic adjustment ability of the pan-tilt, and accurately adapt to the dynamic environment. Description of the Drawings

[0017] Figure 1 It is a schematic flowchart of the method steps of the present invention.

[0018] Figure 2 It is a schematic structural diagram of the system of the present invention.

[0019] In the figure: 10 - acquisition module, 20 - solar position calculation module, 30 - environmental impact factor calculation module, 40 - pan-tilt control strategy module, 50 - future solar trajectory prediction module, 60 - pan-tilt adjustment module. Detailed Embodiments

[0020] Next, the solutions of the present invention will be clearly and completely described in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments.

[0021] It should be noted that in the description and claims of this application and the above-mentioned drawings, terms such as "first" and "second" are used to distinguish similar objects, and do not necessarily describe a specific order or sequence. It should be understood that such data used can be interchanged under appropriate circumstances, so as to implement the embodiments of this application described herein. In addition, the terms "including" and "having" and any variations thereof are intended to cover non-exclusive inclusion. For example, a process, method, system, product, or device that includes a series of steps or units does not necessarily limit to those steps or units clearly listed, but may include other steps or units not clearly listed or inherent to these processes, methods, products, or devices.

[0022] Those skilled in the art of this technology can understand that unless specifically stated, the singular forms "a", "an", "the", and "said" used herein may also include the plural forms. It should be further understood that the term "including" used in the description of this application means the presence of features, integers, steps, operations, elements, and / or components, but does not exclude the presence or addition of one or more other features, integers, steps, operations, elements, components, and / or their groups. It should be understood that when we say an element is "connected" or "coupled" to another element, it can be directly connected or coupled to other elements, or there may also be intermediate elements. In addition, the "connection" or "coupling" used herein may include wireless connection or wireless coupling. The phrase "and / or" used herein includes all or any unit and all combinations of one or more related listed items.

[0023] Those skilled in the art of this technology can understand that unless otherwise defined, all terms used herein (including technical terms and scientific terms) have the same meaning as the general understanding of those of ordinary skill in the field to which this application belongs. It should also be understood that terms such as those defined in a general dictionary should be understood to have a meaning consistent with the meaning in the context of the prior art, and will not be interpreted with an idealized or overly formal meaning unless specifically defined as here.

[0024] It should be understood that the sequence numbers and magnitudes of the steps in this embodiment do not mean the sequence of execution. The execution sequence of each process is determined by its function and internal logic, and should not constitute any limitation to the implementation process of the embodiments of this application.

[0025] It should be noted that, without conflict, the embodiments in this application and the features in the embodiments can be combined with each other. The following will refer to the drawings and combine with the embodiments to detail this application.

[0026] Please refer to Figure 1 , a dynamic adjustment method of a solar tracking cloud platform based on solar trajectory prediction according to the present invention, the steps include:

[0027] S1. Obtain the positioning data, time data, and environmental data of the solar panel.

[0028] S2. Calculate the solar position data according to the positioning data and the time data through the solar position algorithm.

[0029] S3. Substitute the solar position data and the environmental data into the weight fusion formula to calculate the environmental impact factor.

[0030] S4. Use the medium - and short - term prediction algorithm to predict the future environmental data and adjust the pan - tilt control strategy through a preset threshold.

[0031] S5. Establish a long - short - term memory network model, input the solar position data, environmental data, and the environmental impact factor into the long - short - term memory network model, output the predicted future solar trajectory, calculate the prediction error, and correct the data in the long - short - term memory network model and / or the medium - and short - term prediction algorithm.

[0032] S6. Optimize the future solar trajectory and the pan - tilt control strategy, and calculate and adjust the pitch angle and azimuth angle of the pan - tilt according to the optimized results.

[0033] Specifically, in this application, the solar position data and the environmental impact factor are calculated, the future environmental data is predicted through the medium - and short - term prediction algorithm, and the pan - tilt control strategy is adjusted. Multiple data such as the solar position data, environmental data, and environmental impact factor are input into the long - short - term memory network model to predict the future solar trajectory, and the data in the long - short - term memory network model and / or the medium - and short - term prediction algorithm is corrected through the error result, which can improve the prediction accuracy of the solar position and environmental data, enhance the dynamic adjustment ability of the pan - tilt, and accurately adapt to the dynamic environment.

[0034] In some embodiments, the step of obtaining the positioning data, time data, and environmental data of the solar panel specifically includes: obtaining the longitude, latitude, and altitude of the solar panel, obtaining the current time, obtaining the light intensity, temperature, humidity, and wind speed; the solar position data includes the solar altitude angle and the solar azimuth angle.

[0035] Specifically, obtain the UTC time, longitude, latitude, and altitude through the Beidou and GPS dual - mode positioning module, and calculate the Julian day. The calculation formula is:

[0036] In the formula, the JD is the Julian day, the Y is the year, the M is the month, the D is the day, the UT is the universal time, the Take the floor value, the 367Y is the number of days in the calculated year part, the Is the floor value, used to adjust the relationship between the year and the month, the To calculate the number of days in the month part, the 1721013.5 is a constant term used to adjust the reference date, and the is to convert the universal time to days. This application corrects the error through satellite signals to improve the positioning accuracy, and adopts high-precision clock synchronization to ensure that the time error is less than 1 millisecond.

[0037] Optionally, when calculating the solar ecliptic longitude, in order to improve the accuracy of calculating the solar declination angle, this application uses the solar position algorithm for correction, and its calculation formula is:

[0038] a = 280.46646° + 0.98564736°(JD - 2451545) + C, C = 1.914602sin g + 0.019993sin2g + 0.000289sin3g, g = 357.52911 + 0.98560028(JD - 2451545), where a is the solar ecliptic longitude, JD is the Julian day, C is the orbital perturbation correction term, and g is the mean anomaly of the Earth's orbit.

[0039] Optionally, calculate the solar declination angle. The solar declination angle determines the north-south position of the sun in the sky, and its calculation formula is:

[0040] b = arcsin(sin c sin a), c = 23.4393° - 0.00000036(JD - 2451545), where b is the solar declination angle, c is the obliquity of the ecliptic of the Earth, and a is the solar ecliptic longitude. This application considers the eccentricity of the Earth's orbit, improves the calculation of the declination angle, and adopts a second-order error correction term to improve the calculation accuracy.

[0041] Optionally, calculate the solar hour angle. The solar hour angle reflects the position of the sun relative to the local meridian, and its calculation formula is:

[0042] d = 15°×(T local - 12), where d is the solar hour angle and T local is the local time. This application synchronizes the UTC time through Beidou and GPS to ensure high precision, and combines the altitude to correct the solar hour angle.

[0043] Optionally, calculate the solar altitude angle. The solar altitude angle represents the height of the sun relative to the horizon, and its calculation formula is:

[0044] h = arcsin(sin φ sin b + cos φ cos b cos d), where h is the solar altitude angle, φ is the local latitude, b is the solar declination angle, and d is the solar hour angle.

[0045] Optionally, calculate the solar azimuth angle, which describes the angle of the sun relative to the due north direction, and its calculation formula is:

[0046] In the formula, A is the solar azimuth angle, d is the solar hour angle, φ is the local latitude, and b is the solar declination angle.

[0047] In some embodiments, after the step of calculating the solar position data, it further includes: calculating the theoretical light intensity through the solar position data, collecting the real-time light intensity through a light intensity sensor, and correcting the light intensity by a dynamic weight method, and its calculation formula is: I abjusted (h, A) = w1I predicted +w2I actual , w2 = 1 - w1; in the formula, I abjusted (h, A) is the corrected light intensity, I predicted is the theoretical light intensity, I actual is the real-time light intensity, I max is the maximum light intensity, w1 is the weight of the theoretical light intensity, and w2 is the weight of the real-time light intensity.

[0048] Specifically, since the light intensity is affected by the solar altitude angle and the solar azimuth angle, the present application corrects the light intensity. When |I predicted -I actual | is small, the weight of the theoretical value I predicted is higher to ensure stability; when the error is large, the weight of the sensor measurement value I actual increases to improve adaptability. In addition to calculating the theoretical solar light intensity value, the present application also combines a light intensity sensor, a humidity sensor, and a wind speed sensor for dynamic weight correction to ensure that the calculated light intensity is more in line with the actual situation. When the calculated value has a large deviation, the weight of the sensor data increases to reduce the error, and it still has high accuracy in environments with cloudy, strong wind, and serious air pollution.

[0049] Optionally, the step of calculating the theoretical light intensity through the solar position data specifically includes: the calculation formula for calculating the light intensity is:

[0050] I predicted = I0×T q ×cos(h)×cos(θ - A), in the formula, I predicted is the theoretical light intensity, I0 is the solar constant, T q = e -k×AMis the atmospheric transmittance to correct the influence of air absorption, where k is an empirical coefficient, usually between 0.1 and 0.3, the AM is the air mass factor, the cos(h) is the influence of the projection of light on the ground to correct the influence of the solar altitude angle, and the cos(θ - A) is to correct the included angle between the solar azimuth angle and the orientation of the solar panel. In this application, if the solar panel always faces the solar azimuth angle A, then the light reception is maximized: cos(θ - A) = 1; if the panel is fixed and θ - A ≠ 0°, the incident light energy decreases. For example, if the panel faces south (θ = 180°) but the sun is in the west (A = 270°), the light intensity will decrease significantly.

[0051] In some embodiments, the step of substituting the solar position data and the environmental data into the weight fusion formula to calculate the environmental impact factor specifically includes: calculating the environmental impact factor by using the weight fusion formula, and its calculation formula is: E(h, A) = αI abjusted (h, A) + βH + γT + δW, In the formula, the E(h, A) is the environmental impact factor, and the I abjusted (h, A) is the corrected light intensity, the h, A are the solar position data, the H is the humidity, the T is the temperature, the W is the wind speed, the α is the weight of the corrected light intensity, the β is the humidity weight, the β0 is the reference humidity weight, the γ is the temperature weight, the γ0 is the reference temperature weight, the δ is the wind speed weight, and the δ0 is the reference wind speed weight.

[0052] Specifically, this application substitutes multiple data such as solar position data and environmental data into the weight fusion formula to calculate the environmental impact factor, which can dynamically adjust the pan-tilt control strategy according to the environmental impact factor, and use the environmental impact factor as the input of the long short-term memory network model to optimize the solar trajectory prediction of the long short-term memory network model.

[0053] In some embodiments, the step of predicting future environmental data by using the medium and short-term prediction algorithm specifically includes: performing real-time prediction on future environmental data through the Kalman filter algorithm, and performing medium and short-term prediction on future environmental data through the autoregressive integrated moving average model; calculating the environmental change rate V and the historical environmental stability σ 2 , and setting the environmental change rate threshold V th and the historical environmental stability threshold

[0054] Specifically, the Kalman filter algorithm is used to perform real-time prediction on future environmental data, which is suitable for quickly responding to weather changes and optimizing the pan-tilt control in real time. The autoregressive integrated moving average model is used to perform medium- and short-term prediction on future environmental data, which is suitable for predicting the weather change trend within 5 to 30 minutes. The system can adaptively select different strategies according to the actual situation, can more accurately adapt to the dynamic environment, and reduce the energy consumption waste caused by frequent adjustments.

[0055] In some embodiments, the step of adjusting the pan-tilt control strategy through a preset threshold specifically includes: if V < V th and then select the autoregressive integrated moving model to perform medium- and short-term prediction on future environmental data. If V < V th and then select the Kalman filter algorithm to perform real-time prediction on future environmental data. If V ≥ V th and then select the Kalman filter algorithm to perform real-time prediction on future environmental data. If V ≥ V th and then select the Kalman filter algorithm to perform real-time prediction on future environmental data.

[0056] Specifically, based on the change rate of environmental data and the stability of historical data, that is, the environmental change rate V and the historical environmental stability σ 2 judge and select the adjustment strategy. Among them, the calculation formula of the environmental change rate is: In the formula, V is the environmental change rate, E t is the environmental impact factor calculated at the current moment, E t-1 is the environmental impact factor at the previous moment, and Δt is the data sampling interval. By default, it is 1 second in this application.

[0057] Optionally, the calculation formula of the environmental change rate threshold is: In the formula, V th is the environmental change rate threshold, H is the humidity, W is the wind speed, V base is the adaptive setting method calculated based on historical data, and the average change rate in the past N seconds is used as the reference value. Its calculation formula is: In the formula, V base is the adaptive setting method calculated based on historical data, N is the total number of samples, E i is the environmental impact factor of the i-th sample, E i-1 is the environmental impact factor of the (i - 1)-th sample, and Δt is the data sampling interval.

[0058] Optionally, if V < V th, that is, if the environmental change is slow, the autoregressive integrated moving average model is used to perform medium- and short-term prediction on future environmental data. The autoregressive integrated moving average model is applicable to stationary time series and is suitable for situations where the weather changes slowly, such as sunny days; if V≥V th , that is, if the environmental change is fast, the Kalman filter algorithm is used to perform real-time prediction on future environmental data. The Kalman filter algorithm is applicable to situations with rapid changes, such as rapid changes in clouds and sudden changes in wind speed.

[0059] Optionally, the stability calculation formula of the historical environmental data is: In the formula, the σ 2 is the historical environmental stability, the E t is the current environmental impact factor, the E is the average value of the environmental impact factors in the most recent N times, and the N is the total amount of the data window in the past period of time.

[0060] Optionally, if that is, the weather state is relatively stable, the autoregressive integrated moving average model is used to perform medium- and short-term prediction on future environmental data. If that is, the weather state changes greatly, the Kalman filter algorithm is used to perform real-time prediction on future environmental data; wherein, the typical value of is, in a stable weather state, such as in the case of weak wind and sunny days, in a changeable weather state, such as in the case of strong wind and cloudy or rainy days, If the environmental factors fluctuate little for a long time, the data sequence is relatively stationary, and using the autoregressive integrated moving average model to perform medium- and short-term prediction on future environmental data will be more accurate; if the historical environmental stability is high, such as drastic weather changes and frequent sunny and cloudy transitions, then it is more appropriate to select the Kalman filter algorithm to perform real-time prediction on future environmental data, which can quickly adjust the estimation of the future state.

[0061] Optionally, the pan-tilt control strategy is dynamically adjusted according to the environmental impact factor; in a low-light state, that is, when the light intensity is less than the preset light intensity threshold, the system reduces the pan-tilt adjustment frequency to reduce the probability of misoperation, and its calculation formula is: In the formula, the F abjust is the adjusted pan-tilt operation frequency, the F base is the basic pan-tilt adjustment frequency, the I(h, A) is the currently calculated light intensity, and the I maxis the theoretical maximum light intensity. In the present application, when I(h, A) is 0, it can be understood as night, that is, the pan-tilt adjustment is stopped and the low-power mode is entered; when I(h, A) is lower than the preset light intensity threshold, such as on a cloudy day or in case of occlusion, the adjustment frequency is reduced to prevent misoperation; when I(h, A) resumes above the preset light intensity threshold, the normal adjustment frequency is restored.

[0062] Optionally, in a strong wind state, that is, when the wind speed is greater than the preset wind speed threshold, the solar panel is adjusted to the horizontal angle to reduce the wind resistance. If the wind speed is greater than 10 m / s, the pan-tilt adjustment frequency is reduced to reduce shaking. If the wind speed is greater than 20 m / s, the wind protection mode is entered and the solar panel is adjusted to the horizontal angle. When the wind speed drops below the preset wind speed threshold, the normal tracking mode is restored.

[0063] Optionally, in a high humidity state, when the humidity is high, clouds may cause errors in light measurement, and the pan-tilt adjustment frequency is reduced to reduce misadjustment. When the humidity is greater than 80%, the pan-tilt adjustment frequency is reduced to prevent frequent misoperations caused by cloud occlusion. When the humidity is less than 80%, the normal adjustment frequency is restored.

[0064] In some embodiments, the step of predicting the future solar trajectory specifically includes: combining the solar position data and the environmental impact factors as the input vector of the long short-term memory network model, and its expression is: X t =[h t , A t JD t φ, λ, E t , where the X t is the model input vector, the h t is the solar altitude angle, the A t is the solar azimuth angle, the JD t is the current time, the φ is the latitude, the λ is the longitude, and the E t is the environmental impact factor; the long short-term memory network model outputs the predicted value of the future solar trajectory by learning N historical data, and its calculation formula is: where the is the predicted solar altitude angle, the is the predicted solar azimuth angle, and (X t , X t-1 ,..., X t-N ) are historical data.

[0065] In some embodiments, the steps of calculating the prediction error and correcting the data in the long short-term memory network model and / or the medium and short-term prediction algorithm specifically include: comparing the predicted solar altitude angle and the predicted solar azimuth angle with the actual measured values, and correcting the future solar trajectory, and its calculation formula is: In the formula, L is the mean square error loss function, N is the total number of samples, and h i is the true value of the solar altitude angle of the i-th sample, and the is the predicted value of the solar altitude angle of the i-th sample, and A i is the true value of the solar azimuth angle of the i-th sample, and the is the predicted value of the solar azimuth angle of the i-th sample; during the training process of the long short-term memory network model, the Adam optimizer is used to adjust the weights, and the loss is minimized by the gradient descent method, and the error range is calculated. If the error range is higher than the preset error range threshold, the data in the long short-term memory network model and / or the medium and short-term algorithm is adjusted.

[0066] In some embodiments, the steps of optimizing the future solar trajectory and the pan-tilt control strategy include: optimizing the pan-tilt control strategy and the future solar trajectory through the particle swarm optimization algorithm.

[0067] Specifically, the future solar position predicted by the long short-term memory network model is used for pan-tilt angle optimization and input into the particle swarm optimization algorithm, and its calculation formula is: In the formula, the is the fitness function, and the ψ is the pan-tilt angle and azimuth angle of the pan-tilt, I is the light intensity, and the is the solar altitude angle at the next moment, and the is the solar azimuth angle at the next moment; in this application, the particle swarm optimization algorithm is used to iteratively optimize the pan-tilt angle and azimuth angle to maximize the light intensity received by the solar panel.

[0068] Optionally, according to the solar trajectory predicted by the long short-term memory network model, combined with the environmental impact factor, the pan-tilt adjustment frequency is adaptively adjusted; if the weather is stable, that is then the pan-tilt adjustment frequency is reduced, and the calculation formula of the adaptive adjustment function is: In the formula, F adjust is the adaptive adjustment frequency, F base is the reference adjustment frequency, E(h, A) is the environmental impact factor, and E max is the environmental impact factor at the optimal weather moment. This method is applicable to sunny days and low wind speed conditions.

[0069] Optionally, if the weather changes rapidly, that is Then increase the adjustment frequency, and the calculation formula of the adaptive adjustment function is as follows: In the formula, the F adjust is the adaptive adjustment frequency, and the F base is the reference adjustment frequency, the E(h, A) is the environmental impact factor, and the E max is the environmental impact factor at the optimal weather moment. This method is applicable to cloudy weather and can quickly respond to the change of the sun's trajectory.

[0070] Optionally, at regular intervals, for example, every 10 seconds, according to the error between the actual position of the sun and the prediction result, that is, the error result between the current sun position obtained by using the sun position algorithm and the predicted sun trajectory, adjust the data in the long short-term memory network model, and its calculation formula is as follows: In the formula, the L new is the error, the N is the total number of samples, the (h, A) i is the true value of the sun position, and the is the predicted value.

[0071] Optionally, through the particle swarm optimization algorithm, find the optimal solution of the adjustment angle of the pan-tilt head to maximize the light intensity received by the solar panel. The calculation formula of the objective function is as follows: In the formula, the is the objective function, the ψ is the pitch angle and azimuth angle of the pan-tilt head, the I(h, A) is the light intensity, which is affected by the solar altitude angle h and azimuth angle A, the is the predicted solar altitude angle and azimuth angle at the next moment, and the E(h, A) is the environmental impact factor. Through optimizing the pitch angle and azimuth angle of the pan-tilt head, this application makes and maximized, so that the solar panel faces the sun as much as possible, and through optimizing the environmental impact factor, the solar panel reduces adjustment in bad weather, reduces energy consumption, and reduces the risk of damage.

[0072] Optionally, set parameters for the particle swarm optimization algorithm. Initial particle swarm X: ψ i ∈[0°, 360°]; where each particle X i represents a pan-tilt head angle configuration The N is the number of particles, and the initial value is randomly generated within the set range.

[0073] Optionally, dynamically adjust the inertia weight, and its calculation formula is as follows: In the formula, the w is the inertia weight, the w max is the maximum value of the inertia weight and the maximum is 0.9, the w minis the minimum inertia weight and is at least 0.4, where t is the current iteration number, which can ensure sufficient search in the early stage and accelerated convergence in the later stage.

[0074] Optionally, the particle swarm optimization algorithm updates the velocity and position of each particle, and its calculation formula is: In the formula, the is the velocity of the (i + 1)-th particle, w is the inertia weight, the is the velocity of the i-th particle, the is the individual optimal solution, G t is the global optimal solution, r1 and r2 are random numbers between 0 and 1, which are used to increase the search randomness, and c1 and c2 are acceleration factors. Set a threshold. If the change in the global optimal value of the particle swarm is less than the threshold in multiple consecutive rounds of iteration, the iteration is terminated in advance.

[0075] Optionally, the optimization strategy is adjusted by combining the environmental impact factor. Specifically, the search range of the particle swarm optimization algorithm is dynamically adjusted according to the environmental impact factor. If the weather is stable, the default calculation range of the particle swarm optimization algorithm is maintained. If the weather is bad, the adjustment angle of the pan-tilt head is restricted and the movement frequency is reduced. After multiple rounds of iteration, the optimal pan-tilt head angle is found and adjusted.

[0076] On the other hand, please refer to Figure 2 , the present invention also provides a dynamic adjustment system for a solar tracking pan-tilt head based on solar trajectory prediction, including: an acquisition module 10, configured to acquire positioning data, time data, and environmental data of a solar panel; a solar position calculation module 20, configured to calculate solar position data according to the positioning data and the time data through a solar position algorithm; an environmental impact factor calculation module 30, configured to substitute the solar position data and the environmental data into a weight fusion formula to calculate an environmental impact factor; a pan-tilt head control strategy module 40, configured to predict future environmental data by using a medium-term and short-term prediction algorithm and adjust the pan-tilt head control strategy through a preset threshold; a future solar trajectory prediction module 50, configured to establish a long short-term memory network model, input the solar position data, environmental data, and the environmental impact factor into the long short-term memory network model, output a predicted future solar trajectory, calculate a prediction error, and correct data in the long short-term memory network model and / or the medium-term and short-term prediction algorithm; a pan-tilt head adjustment module 60, configured to optimize the future solar trajectory and the pan-tilt head control strategy, and calculate and adjust the pitch angle and azimuth angle of the pan-tilt head according to the optimized result.

[0077] When the integrated unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art, or all or part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions for causing a computer device (which may be a personal computer, a server, or a network device, etc.) to execute all or part of the steps of the methods of the various embodiments of the present invention. The aforementioned storage medium includes: various media such as USB flash drives, mobile hard disks, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical discs that can store program codes.

[0078] Those of ordinary skill in the art can understand that to implement all or part of the processes in the above-described embodiment methods, it can be completed by instructing relevant hardware through a computer program. The computer program can be stored in a non-volatile computer-readable storage medium 11. When the computer program is executed, it can include the processes of the embodiments of the above-described various methods. Among them, any reference to the memory 10, storage, database, or other media used in the various embodiments provided by the present invention can include non-volatile and / or volatile memories 10. The non-volatile memory 10 can include read-only memory 10 (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM), or flash memory. The volatile memory 10 can include random access memory 10 (RAM) or an external cache memory 10. By way of illustration and not limitation, RAM is available in various forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), double data rate SDRAM (DDR SDRAM), enhanced SDRAM (ESDRAM), synchronous link DRAM (SLDRAM), Rambus direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and Rambus dynamic RAM (RDRAM), etc.

[0079] The above are only the embodiments of the present invention, and do not limit the patent scope of the present invention. All equivalent transformations made by using the content of the specification and drawings of the present invention, directly or indirectly applied in the related technical fields, are equally included in the patent protection scope of the present invention.

Claims

1. A dynamic adjustment method for a solar tracking cloud platform based on solar trajectory prediction, characterized in that the steps It includes: Obtain the positioning data, time data, and environmental data of the solar panel; Calculate the solar position data according to the positioning data and the time data through the solar position algorithm; Substitute the solar position data and the environmental data into the weight fusion formula to calculate the environmental impact factor; Use the medium- and short-term prediction algorithm to predict future environmental data, and adjust the pan-tilt control strategy through a preset threshold; Establish a long short-term memory network model, input the solar position data, environmental data, and the environmental impact factor into the long short-term memory network model, output the predicted future solar trajectory, calculate the prediction error, and correct the data in the long short-term memory network model and / or the medium- and short-term prediction algorithm; Optimize the future solar trajectory and the pan-tilt control strategy, and calculate and adjust the pitch angle and azimuth angle of the pan-tilt according to the optimized results.

2. The dynamic adjustment method of the solar tracking cloud platform based on solar trajectory prediction according to claim 1, wherein, The step of optimizing the future solar trajectory and the pan-tilt control strategy includes: Optimize the pan-tilt control strategy and the future solar trajectory through the particle swarm optimization algorithm.

3. The dynamic adjustment method of the solar tracking cloud platform based on solar trajectory prediction according to claim 1, characterized in that The step of obtaining the positioning data, time data, and environmental data of the solar panel specifically includes: Obtain the longitude, latitude, and altitude of the solar panel, obtain the current time, and obtain the light intensity, temperature, humidity, and wind speed; The solar position data includes the solar altitude angle and the solar azimuth angle.

4. The dynamic adjustment method of the solar tracking cloud platform based on solar trajectory prediction according to claim 1, wherein, After the step of calculating the solar position data, it further includes: Calculate the theoretical light intensity through the solar position data, collect the real-time light intensity through a light intensity sensor, and correct the light intensity in a dynamic weight manner. The calculation formula is: I abjusted (h, A) = w1I predicted + w2I actual , w2 = 1 - w1; Wherein, the I abjusted (h, A) is the corrected light intensity, the I predicted is the theoretical light intensity, the I actual is the real-time light intensity, the I max is the maximum light intensity, the w1 is the weight of the theoretical light intensity, and the w2 is the weight of the real-time light intensity.

5. The dynamic adjustment method of the solar tracking cloud platform based on solar trajectory prediction according to claim 1, characterized in that, The step of substituting the solar position data and the environmental data into the weight fusion formula to calculate the environmental impact factor specifically includes: Calculate the environmental impact factor by using the weight fusion formula. The calculation formula is: E(h, A) = αI abjusted (h, A) + βH + γT + δW, Wherein, the E(h, A) is an environmental impact factor, and the I abjusted (h, A) is the corrected light intensity, the h, A are solar position data, the H is humidity, the T is temperature, the W is wind speed, the α is the weight of the corrected light intensity, the β is the humidity weight, the β0 is the reference humidity weight, the γ is the temperature weight, the γ0 is the reference temperature weight, the δ is the wind speed weight, and the δ0 is the reference wind speed weight.

6. The dynamic adjustment method of the solar tracking cloud platform based on solar trajectory prediction according to claim 1, characterized in that, The step of using the medium- and short-term prediction algorithm to predict future environmental data specifically includes: Perform real-time prediction of future environmental data through the Kalman filter algorithm, and perform medium- and short-term prediction of future environmental data through the autoregressive integrated moving average model; Calculate the environmental change rate V and the historical environmental stability σ based on the environmental impact factor and the environmental data 2 , and set the environmental change rate threshold V th and the historical environmental stability threshold 7. The dynamic adjustment method of the solar tracking cloud platform based on solar trajectory prediction according to claim 6, characterized in that, The step of adjusting the pan-tilt control strategy through a preset threshold specifically includes: If V < V th and then select an autoregressive integrated moving model to perform medium- and short-term prediction on future environmental data. If V < V th and then select the Kalman filter algorithm to perform real-time prediction on future environmental data. If V ≥ V th and then select the Kalman filter algorithm to perform real-time prediction on future environmental data. If V ≥ V th and then select the Kalman filter algorithm to perform real-time prediction on future environmental data.

8. The dynamic adjustment method of the solar tracking cloud platform based on solar trajectory prediction according to claim 3, characterized in that The step of predicting the future solar trajectory specifically includes: Combine the solar position data and the environmental impact factor as the input vector of the long short-term memory network model. The expression is: X t = [h t , A t JD t φ, λ, E t , where the X t is the model input vector, the h t is the solar altitude angle, the A t is the solar azimuth angle, the JD t is the current time, the φ is the latitude, the λ is the longitude, and the E t is the environmental impact factor; The long short-term memory network model outputs the predicted value of the solar trajectory at a future moment by learning N historical data. The calculation formula is: In the formula, the is the predicted solar altitude angle, and the is the predicted solar azimuth angle. The (X t , X t-1 ,..., X t-N ) are historical data.

9. The dynamic adjustment method of the solar tracking cloud platform based on solar trajectory prediction according to claim 8, characterized in that, The step of calculating the prediction error and correcting the data in the long short-term memory network model and / or the medium- and short-term prediction algorithm specifically includes: Compare the predicted solar altitude angle and the predicted solar azimuth angle with the actual measured values, and correct the future solar trajectory. The calculation formula is: Wherein, L is the mean square error loss function, N is the total number of samples, h i is the true value of the solar altitude angle of the i-th sample, and the is the predicted value of the solar altitude angle of the i-th sample, and A i is the true value of the solar azimuth angle of the i-th sample, and the is the predicted value of the solar azimuth angle of the i-th sample; During the training process of the long short-term memory network model, use the Adam optimizer to adjust the weights, minimize the loss through the gradient descent method, and calculate the error range. If the error range is higher than the preset error range threshold, adjust the data in the long short-term memory network model and / or the medium- and short-term algorithm.

10. A dynamic adjustment system for a solar tracking cloud platform based on solar trajectory prediction, characterized in that, It includes: An acquisition module, configured to acquire the positioning data, time data, and environmental data of a solar panel; A solar position calculation module, configured to calculate solar position data according to the positioning data and the time data through a solar position algorithm; An environmental impact factor calculation module, configured to substitute the solar position data and the environmental data into a weight fusion formula to calculate an environmental impact factor; A pan-tilt control strategy module, configured to predict future environmental data by using a medium- and short-term prediction algorithm and adjust the pan-tilt control strategy through a preset threshold; A future solar trajectory prediction module, configured to establish a long short-term memory network model, input the solar position data, environmental data, and the environmental impact factor into the long short-term memory network model, output a predicted future solar trajectory, calculate a prediction error, and correct data in the long short-term memory network model and / or the medium- and short-term prediction algorithm; A pan-tilt adjustment module, configured to optimize the future solar trajectory and the pan-tilt control strategy, and calculate and adjust the pitch angle and azimuth angle of the pan-tilt according to the optimized result.

Images