Self-adaptive photovoltaic tracking control method and system based on artificial intelligence

Abstract

The invention discloses a self-adaptive photovoltaic tracking control method and system based on artificial intelligence, and the method comprises the steps: collecting the data of a photovoltaic module, dividing the data of the photovoltaic module into control data and adjustment data, and carrying out the preprocessing of the data; determining the main direction of the photovoltaic module according to the control data, and predicting the environment change according to the adjustment data; and establishing an adjustment model, taking the photovoltaic power generation efficiency as a target, calculating an angle and a direction adjustment value which need to be finely adjusted, and adjusting the angle of the photovoltaic module in real time. According to the invention, the environmental change is predicted by collecting and preprocessing the environmental data, and the photovoltaic module is finely adjusted in real time and in advance by using the adjustment model. According to the method, the power generation efficiency is remarkably improved, the system stability is enhanced, the energy consumption is effectively reduced, and the service life of equipment is prolonged.

Description

Technical Field

[0001] The present invention relates to the technical field of photovoltaic power generation, and in particular to an adaptive photovoltaic tracking control method and system based on artificial intelligence. Background Art

[0002] With the continuous growth of global energy demand and the improvement of environmental protection awareness, photovoltaic power generation technology, as a clean and renewable energy source, has received widespread attention and rapid development. Traditional photovoltaic power generation systems mostly adopt a fixed installation method, that is, photovoltaic modules are installed at a fixed angle to receive as much sunlight as possible. However, when the intensity and angle of sunlight vary greatly, the fixed photovoltaic system cannot ensure that the photovoltaic modules are always in the best power generation state, thereby reducing the overall power generation efficiency of the system. In order to solve this problem, photovoltaic tracking control technology came into being. By dynamically adjusting the angle of photovoltaic modules to keep them aligned with the sun, the energy capture ability and power generation efficiency of the photovoltaic system are improved. Existing photovoltaic tracking control systems mainly rely on preset trajectories or simple sensor feedback, but these methods are difficult to achieve the best tracking effect under complex and changing environmental conditions.

[0003] There are many shortcomings in the existing photovoltaic tracking control technology. First, the traditional preset trajectory control method cannot adapt to the dynamic changes of sunshine conditions and environmental factors, resulting in the photovoltaic modules not always being able to maintain the optimal power generation angle. Secondly, although the simple sensor feedback control method can improve the angle adjustment of photovoltaic modules to a certain extent, it is unable to provide comprehensive environmental information due to the lack of singleness and real-time nature of sensor data, resulting in limited response speed and control accuracy of the system. In addition, the existing photovoltaic tracking control system generally lacks adaptive learning capabilities and cannot be intelligently optimized according to historical data and environmental changes, resulting in poor adaptability and robustness of the photovoltaic system under different environmental conditions. In view of these shortcomings, an adaptive photovoltaic tracking control method based on artificial intelligence is proposed. By introducing reinforcement learning and multimodal sensor fusion technology, the intelligence level and control effect of the system can be significantly improved. Summary of the invention

[0004] In view of the above-mentioned problems, the present invention is proposed.

[0005] Therefore, the technical problem solved by the present invention is that the existing photovoltaic tracking control technology lacks the ability to fully perceive and intelligently optimize environmental changes.

[0006] To solve the above technical problems, the present invention provides the following technical solutions: an artificial intelligence-based adaptive photovoltaic tracking control method, comprising: collecting data of photovoltaic modules, dividing the data of photovoltaic modules into control data and adjustment data, and preprocessing the data; determining the main direction of the photovoltaic modules according to the control data, and predicting environmental changes according to the adjustment data; establishing an adjustment model, taking photovoltaic power generation efficiency as the goal, calculating the angle and direction adjustment values ​​that need to be fine-tuned, and adjusting the angle of the photovoltaic modules in real time.

[0007] As a preferred solution of the artificial intelligence-based adaptive photovoltaic tracking control method described in the present invention, the data of the photovoltaic components are divided into control data and adjustment data, the control data are light intensity, atmospheric thickness and geographical location data of the photovoltaic components, and the adjustment data include temperature, wind speed, humidity, and atmospheric pressure.

[0008] As a preferred solution of the adaptive photovoltaic tracking control method based on artificial intelligence described in the present invention, the main direction of the photovoltaic module includes using a solar position calculation algorithm to determine the real-time position of the sun in the sky and the ideal light intensity according to the geographical location and time of the photovoltaic module. The altitude angle g and azimuth angle α of the sun are expressed as:

[0009] g=arcsin(sin(Latitude)·sin(δ)+cos(Latitude)·cos(δ)·cos(H))

[0010]

[0011] Ideal light intensity I t It is expressed as:

[0012] I t =I0·cos(g)·e -τ·m

[0013] Among them, Latitude represents the latitude of the photovoltaic module; δ represents the solar declination; H represents the hour angle; I0 represents the solar constant; τ represents the atmospheric optical thickness; m represents the air mass coefficient; Under ideal conditions, the tilt angle β of the photovoltaic module is perpendicular to the altitude angle of the sun, and the azimuth angle γ is the same as the solar azimuth angle, which is expressed as:

[0014] β=90°-g

[0015] γ=α

[0016] In actual situations, light intensity is indirectly affected by temperature and humidity. The effects of temperature and humidity on environmental factor E are expressed as:

[0017]

[0018] Where T represents the total time; T(t) represents the function of temperature changing with time t; H represents the total humidity; H t Indicates the humidity at time t; calculates the deviation rate threshold ΔI between the actual light intensity and the theoretical light intensity based on environmental factors t , expressed as:

[0019]

[0020] Among them, ΔI base represents the artificially set benchmark deviation; k represents the adjustment coefficient; I t,t represents the ideal light intensity at time t; I actual,t Indicates the actual light intensity at time t; when I actual,t -I t,t ≤ΔI t When I actual,t -I t,t >ΔI t When , it means that the light intensity deviation is too large, and the PV panels are fine-tuned by adjusting the model.

[0021] As a preferred solution of the adaptive photovoltaic tracking control method based on artificial intelligence described in the present invention, wherein: the prediction of environmental changes includes taking the adjustment data as input, using random forest regression to predict the adjustment data after one hour, evaluating the model performance through cross-validation, dividing the adjustment data into K parts, each part is used as a validation set in turn, and the remaining K-1 parts are used as training sets, repeating K times, and calculating the average performance index, which is expressed as:

[0022]

[0023] Among them, CV score represents the cross-validation score; K represents the number of cross-validation folds; score i Represents the performance score of the i-th fold; a judgment is made based on the predicted adjustment data. If the predicted adjustment data will reduce the efficiency of the PV module, the PV module is fine-tuned in advance through the adjustment model to avoid efficiency reduction.

[0024] As a preferred solution of the adaptive photovoltaic tracking control method based on artificial intelligence described in the present invention, the adjustment model includes establishing an adjustment model based on the proximal strategy optimization PPO, and the output data is the angle adjustment amount β t , input data s tis the angle and adjustment data of the current photovoltaic module. When the light intensity deviation is too large, the input adjustment data is real-time data for real-time adjustment. When the predicted adjustment data will reduce the efficiency of the photovoltaic module, the input data is the predicted adjustment data for advance adjustment. Taking photovoltaic power generation efficiency as the goal, the reward function is designed according to photovoltaic power generation efficiency, system stability and energy consumption factors. Set constraints, use the reward function to iteratively train the adjustment model, use the trained adjustment model to output the best angle adjustment amount, and adjust the angle of the photovoltaic module. The adjustment model is expressed as:

[0025]

[0026] Where L(θ) represents the objective function of the PPO algorithm; represents the expected value; π θ (β t |s t ) means in state s t In this case, policy π uses parameters θ to select actions β t The probability of Indicates that in state s t Next, the strategy π uses the old parameters θ old Select action β t The probability of represents the advantage function; f represents the reward function; ∈ represents the clipping range parameter.

[0027] As a preferred solution of the adaptive photovoltaic tracking control method based on artificial intelligence described in the present invention, the reward function includes power generation efficiency reward, system stability reward and energy consumption penalty. The power generation efficiency reward is given according to the actual power generation, so that the model selects the angle adjustment that can maximize the power generation, which is expressed as:

[0028]

[0029] Among them, R P represents the power generation efficiency reward function; P(t) represents the actual power generation at time t; P max Indicates the maximum possible power generation; system stability reward, rewards the smoothness of the system's angle adjustment, avoids frequent and large angle adjustments, and protects the system's mechanical components, expressed as:

[0030]

[0031] Among them, R sIt represents the system stability reward function; δ represents the weight factor; β(t) represents the tilt angle of the photovoltaic module at time t; the energy consumption penalty is to penalize the energy consumption of each angle adjustment, encouraging the model to minimize the energy consumption of the control system while maintaining high-efficiency power generation, which is expressed as:

[0032]

[0033] Among them, R represents the energy consumption penalty function; It represents the weight factor; Q(t) represents the energy consumption at time t; σ represents the parameter of the adjustment exponential function; the reward function is expressed as:

[0034] f = R P + R s - R

[0035] Among them, f represents the reward function.

[0036] As a preferred solution of the artificial intelligence-based adaptive photovoltaic tracking control method described in the present invention, where: the limiting conditions include that the range of the angle of the photovoltaic module is limited to 0° to 90°, which is expressed as:

[0037] 0° ≤ β + β t ≤ 90°

[0038] Among them, β t represents the angle adjustment amount calculated by the adjustment model; the wind speed thresholds A and B are set according to the physical strength of the photovoltaic module, and A < B. When the wind speed exceeds A, the fine-tuning function of the adjustment model is locked, and only the main direction adjustment of the photovoltaic module is performed. When the wind speed exceeds B, the angle of the photovoltaic module is locked to the horizontal position and no adjustment is made; the upper limit of the adjustment frequency and the amplitude of each adjustment are set to avoid mechanical damage caused by frequent adjustment.

[0039] In the second aspect, the present invention also provides an artificial intelligence-based adaptive photovoltaic tracking control system, including a data acquisition module that acquires data of the photovoltaic module, divides the data into two parts: control data and adjustment data, and performs data cleaning to ensure the integrity and accuracy of the data; a direction control module that uses the preprocessed control data to analyze the direction information of the photovoltaic module during the day, determines the main orientation of the photovoltaic module, and predicts the impact of environmental changes on photovoltaic power generation; a fine-tuning module that targets the photovoltaic power generation efficiency, based on the environmental change prediction results and real-time data, establishes an adjustment model, calculates the angle adjustment value that needs to be fine-tuned, and adjusts the angle of the photovoltaic module in real time according to the output of the adjustment model to ensure that the photovoltaic module can maintain the best power generation efficiency under different environmental conditions.

[0040] In the third aspect, the present invention also provides a computing device, including: a memory and a processor;

[0041] The memory is used to store computer executable instructions, and the processor is used to execute the computer executable instructions. When the computer executable instructions are executed by the processor, the steps of the adaptive photovoltaic tracking control method based on artificial intelligence are implemented.

[0042] In a fourth aspect, the present invention further provides a computer-readable storage medium storing computer-executable instructions, which, when executed by a processor, implement the steps of the artificial intelligence-based adaptive photovoltaic tracking control method.

[0043] Beneficial effects of the invention: The invention collects and pre-processes environmental data, predicts environmental changes, and uses adjustment models to fine-tune photovoltaic modules in real time and in advance. The method significantly improves power generation efficiency, enhances system stability, effectively reduces energy consumption, and prolongs equipment life. BRIEF DESCRIPTION OF THE DRAWINGS

[0044] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings required for use in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other accompanying drawings can be obtained based on these accompanying drawings without paying creative work.

[0045] Figure 1 An overall flow chart of an artificial intelligence-based adaptive photovoltaic tracking control method provided by one embodiment of the present invention. DETAILED DESCRIPTION

[0046] In order to make the above-mentioned purposes, features and advantages of the present invention more obvious and easy to understand, the specific implementation methods of the present invention are described in detail below in conjunction with the drawings of the specification. Obviously, the described embodiments are part of the embodiments of the present invention, but not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary persons in the art without creative work should fall within the scope of protection of the present invention.

[0047] Example 1, reference Figure 1 , as an embodiment of the present invention, provides an adaptive photovoltaic tracking control method based on artificial intelligence, comprising:

[0048] S1: Collect the data of photovoltaic modules, divide the data of photovoltaic modules into control data and adjustment data, and pre-process the data.

[0049] The basic principle of the photovoltaic tracking control system is to adjust the angle of the photovoltaic module in real time through mechanical structure and control algorithm so that it always remains in a direction perpendicular to the sunlight. This can significantly improve the light energy capture ability of the photovoltaic module, especially in the morning and evening when the sunlight angle is low.

[0050] The photovoltaic tracking system is a dual-axis tracking system that can adjust the angles on both the horizontal and vertical axes simultaneously, allowing the photovoltaic modules to follow the sun at any position in the sky and provide the best light energy capture effect.

[0051] Furthermore, the data of the photovoltaic module is divided into control data and adjustment data. The control data includes light intensity, atmospheric thickness and geographical location data of the photovoltaic module, and the adjustment data includes temperature, wind speed, humidity and atmospheric pressure.

[0052] Light intensity is the most direct factor affecting photovoltaic power generation efficiency. The stronger the light, the more light energy the photovoltaic module can capture, and the higher the power generation efficiency. Light intensity is the main control data of the photovoltaic tracking system, which determines the optimal angle adjustment direction of the photovoltaic module. Therefore, light intensity is the main control direction data.

[0053] The atmospheric thickness and the geographic location data of the photovoltaic modules are used to assist in calculating the position of the sun and the light intensity under ideal conditions, and compared with the real-time collected light intensity to determine whether angle adjustment is needed.

[0054] Temperature, wind speed, humidity, and atmospheric pressure indirectly affect photovoltaic efficiency. For example, high temperature will cause the working efficiency of photovoltaic modules to decrease, increase the internal resistance of the modules, and reduce the output power; high wind speed may cause damage to the module structure, and the wind speed also affects the heat dissipation efficiency of the modules.

[0055] By comprehensively considering control data and adjustment data, the system can take into account a variety of environmental factors, perform more precise control and optimization, and improve overall power generation efficiency. In complex environmental conditions, multimodal data can help the system achieve adaptive adjustment and ensure the safe and stable operation of photovoltaic modules under different conditions.

[0056] The collected sensor data is cleaned and standardized to remove noise and abnormal data, ensure the accuracy and consistency of the data, and facilitate subsequent calculations.

[0057] S2: Determine the main direction of the PV module based on the control data and predict environmental changes based on the adjustment data.

[0058] Furthermore, the main direction control of the photovoltaic module is determined based on the light intensity data so that it is always aimed at the sun to maximize the capture of light energy. Light intensity refers to the luminous flux received per unit area, usually in lux (lx), where 1 lux is equal to 1 lumen per square meter.

[0059] The sun position algorithm SPA is used to determine the real-time position of the sun in the sky. It requires the input of the current date, time, geographic location (latitude and longitude) and other information, and outputs the sun's azimuth and altitude.

[0060] Get the current time and date at regular intervals (for example, every minute), and calculate the Julian date JD standardized time, expressed as:

[0061]

[0062] Among them, Y represents year; M represents month; D represents date; and UT represents Coordinated Universal Time.

[0063] According to the geographical location and time of the photovoltaic module, the real-time position of the sun in the sky and the ideal light intensity are determined. The altitude angle g and azimuth angle α of the sun are expressed as:

[0064] g=arcsin(sin(Latitude)·sin(δ)+cos(Latitude)·cos(δ)·cos(H))

[0065]

[0066] Ideal light intensity I t It is expressed as:

[0067] I t =I0·cos(g)·e -τ·m

[0068] Among them, Latitude represents the latitude of the photovoltaic module; δ represents the solar declination; H represents the hour angle; I0 represents the solar constant; τ represents the atmospheric optical thickness; and m represents the air quality coefficient.

[0069] It should be noted that atmospheric optical thickness is a comprehensive data, which represents the comprehensive influence of cloud coverage, atmospheric scattering, aerosols, etc. on light intensity. Considering that the calculation amount of this data is large and the accurate calculation is not very meaningful, cloud coverage is used instead, which can be directly obtained from nearby meteorological stations.

[0070] Ideally, the PV panel tilt angle β is perpendicular to the sun's altitude angle, and the azimuth angle γ is the same as the sun's azimuth angle, expressed as:

[0071] β=90°-g

[0072] γ=α

[0073] In actual situations, light intensity is indirectly affected by temperature and humidity. The effects of temperature and humidity on environmental factor E are expressed as:

[0074]

[0075] Where T represents the total time; T(t) represents the function of temperature changing with time t; H represents the total humidity; H t represents the humidity at time t.

[0076] Calculate the deviation rate threshold ΔI between the actual light intensity and the theoretical light intensity based on environmental factors t , expressed as:

[0077]

[0078] Among them, ΔI base represents the artificially set benchmark deviation; k represents the adjustment coefficient; I t,t represents the ideal light intensity at time t; I actual,t Indicates the actual light intensity at time t; when I actual,t -I t,t ≤ΔI t When I actual,t -I t,t >ΔI t When , it means that the light intensity deviation is too large, and the PV panels are fine-tuned by adjusting the model.

[0079] It should be noted that when calculating light intensity, temperature and wind speed have a greater impact on light intensity, so these two data are considered as environmental factors. Specifically, when the temperature rises, the water vapor and pollutant content in the air may increase, and these substances will absorb and scatter light, reducing the light intensity reaching the ground. In addition, high temperatures may cause more clouds to form, further reducing light intensity.

[0080] High wind speed helps to diffuse pollutants in the air, reduces the absorption and scattering of light, and thus increases light intensity. In addition, wind speed also affects the movement of clouds and the distribution of aerosols, leading to fluctuations in light intensity.

[0081] Furthermore, the prediction of environmental changes includes taking the adjustment data as input, using random forest regression to predict the adjustment data one hour later, evaluating the model performance through cross-validation, dividing the adjustment data into K parts, each of which is used as a validation set in turn, and the remaining K-1 parts are used as training sets, repeating K times, and calculating the average performance index, which is expressed as:

[0082]

[0083] Among them, CV score represents the cross-validation score; K represents the number of cross-validation folds; score i Represents the performance score of the i-th fold; a judgment is made based on the predicted adjustment data. If the predicted adjustment data will reduce the efficiency of the PV module, the PV module is fine-tuned in advance through the adjustment model to avoid efficiency reduction.

[0084] Random forest regression is constructed using scikit-learn. The specific steps are:

[0085] Feature extraction: Take the temperature, wind speed, humidity and air pressure data over a period of time (such as the past 7 days) as features, and extract time features such as hours, days, weeks, etc. to help the model understand the time pattern.

[0086] Dataset division: The historical data is divided into a training set (80%) and a test set (20%) for model training and verification.

[0087] Model selection and training: Build a random forest regression model, train the random forest regression model using training set data, and adjust model parameters to optimize prediction performance.

[0088] Model validation and evaluation: Use K-fold cross validation (such as 5-fold cross validation) to evaluate model performance, ensure the stability and generalization ability of the model, and calculate the mean square error (MSE) and determination coefficient (R 2 ) and other indicators to evaluate the performance of the model on the test set.

[0089] Deploy the trained model to the system and use real-time data to predict environmental parameters.

[0090] It should be noted that random forest regression can directly predict and output temperature, wind speed, humidity and air pressure data, and compare the values ​​in historical data with the predicted data to intuitively reflect the relationship between the predicted data and historical data, so as to decide whether to make adjustments.

[0091] In addition to being used as a basis for adjusting the model, the predicted data can also be used for extreme data warnings. For example, if it is predicted that the temperature is about to exceed the set temperature threshold, cooling measures can be prepared in advance and cooling systems can be activated, such as spraying water for cooling or increasing ventilation, to prevent the reduction of photovoltaic efficiency or even damage to photovoltaic modules.

[0092] S3: Establish an adjustment model, take photovoltaic power generation efficiency as the target, calculate the angle and direction adjustment values ​​that need to be fine-tuned, and adjust the angle of the photovoltaic module in real time.

[0093] Furthermore, an adjustment model is established based on the proximal policy optimization (PPO). PPO is a powerful reinforcement learning algorithm that is widely used in continuous control tasks due to its stability and efficiency. PPO ensures the smoothness and stability of the strategy by limiting the amplitude of the strategy update, and is suitable for the continuous control tasks of the photovoltaic power generation system.

[0094] First, set the input and output. The output data is the angle adjustment amount β t , input data s t It is the angle and adjustment data of the current PV module. When the light intensity deviation is too large, the input adjustment data is real-time data for real-time adjustment. When the predicted adjustment data will reduce the efficiency of the PV module, the input data is predicted adjustment data for advance adjustment.

[0095] Furthermore, taking photovoltaic power generation efficiency as the goal, a reward function is designed based on photovoltaic power generation efficiency, system stability and energy consumption factors, where the power generation efficiency reward is given according to the actual power generation, so that the model can choose the angle adjustment that can maximize the power generation, expressed as:

[0096]

[0097] Among them, R P represents the power generation efficiency reward function; P(t) represents the actual power generation at time t; P max Indicates the maximum possible power generation.

[0098] System stability reward, rewards the smoothness of the system's angle adjustment, avoids frequent and large angle adjustments, and protects the system's mechanical components. It is expressed as:

[0099]

[0100] Among them, R s represents the system stability reward function; δ represents the weight factor; β(t) represents the tilt angle of the PV module at time t.

[0101] Energy consumption penalty, which penalizes the energy consumption of each angle adjustment, encourages the model to minimize the energy consumption of the control system while maintaining efficient power generation, expressed as:

[0102]

[0103] Among them, R represents the energy consumption penalty function; represents the weight factor; Q(t) represents the energy consumption at time t; σ represents the parameter for adjusting the exponential function.

[0104] The reward function is expressed as:

[0105] f=RP +R s -R

[0106] Among them, f represents the reward function.

[0107] In the model, the design and limitations of specific environmental factors need to be considered to ensure the actual application effect and safety of the model. Physical limitations include adjusting speed limits and angle limits to avoid mechanical damage caused by excessive angle changes. The angle range of the photovoltaic module is limited to 0° to 90°, expressed as:

[0108] 0° ≤ β + β t ≤ 90°

[0109] Among them, β t represents the angle adjustment amount calculated by the adjusted model.

[0110] Meteorological factor limitations include angle adjustment limitations under high wind speed conditions and anti-condensation measures under high humidity conditions. Wind speed thresholds A and B are set according to the physical strength of the photovoltaic module, and A < B. When the wind speed exceeds A, the fine-tuning function of the adjusted model is locked, and only the main direction adjustment of the photovoltaic module is performed. When the wind speed exceeds B, the angle of the photovoltaic module is locked to the horizontal position without any adjustment;

[0111] System stability limitations include limiting the amplitude of each adjustment to avoid frequent angle changes; setting an upper limit on the adjustment frequency, such as adjusting at most once per hour.

[0112] Furthermore, the reward function is used to iteratively train the adjusted model, and the trained adjusted model is used to output the best angle adjustment amount to adjust the angle of the photovoltaic module. The adjusted model is expressed as:

[0113]

[0114] Among them, L(θ) represents the objective function of the PPO algorithm; represents the expected value; π θ (β t |s t ) represents the probability that the policy π selects the action β using the parameter θ in the state s t ; t The probability; represents the probability that the policy π selects the action β using the old parameter θ t in the state s old ; t The probability; represents the advantage function; f represents the reward function; ∈ represents the clipping range parameter.

[0115] The present embodiment also provides an adaptive photovoltaic tracking control system based on artificial intelligence, including: a data acquisition module, which collects data of photovoltaic components, divides the data into two parts: control data and adjustment data, and performs data cleaning to ensure the integrity and accuracy of the data; a direction control module, which uses the pre-processed control data to analyze the direction information of the photovoltaic components during the day, determines the main orientation of the photovoltaic components, and predicts the impact of environmental changes on photovoltaic power generation; a fine-tuning module, which takes photovoltaic power generation efficiency as the goal, establishes an adjustment model based on environmental change prediction results and real-time data, calculates the angle adjustment value that needs to be fine-tuned, and adjusts the angle of the photovoltaic components in real time according to the output of the adjustment model to ensure that the photovoltaic components can maintain the best power generation efficiency under different environmental conditions.

[0116] This embodiment also provides a computing device, including a memory and a processor; the memory is used to store computer executable instructions, and the processor is used to execute computer executable instructions to implement the adaptive photovoltaic tracking control method based on artificial intelligence as proposed in the above embodiment.

[0117] This embodiment further provides a storage medium on which a computer program is stored. When the program is executed by a processor, the adaptive photovoltaic tracking control method based on artificial intelligence proposed in the above embodiment is implemented.

[0118] The storage medium proposed in this embodiment and the adaptive photovoltaic tracking control method based on artificial intelligence proposed in the above embodiment belong to the same inventive concept. The technical details not fully described in this embodiment can be referred to the above embodiment, and this embodiment has the same beneficial effects as the above embodiment.

[0119] Through the above description of the implementation method, the technicians in the relevant field can clearly understand that the present invention can be implemented by means of software and necessary general hardware, and of course it can also be implemented by hardware, but in many cases the former is a better implementation method. Based on such an understanding, the technical solution of the present invention is essentially or the part that contributes to the prior art can be embodied in the form of a software product, and the computer software product can be stored in a computer-readable storage medium, such as a computer floppy disk, read-only memory (ReadOnly, Memory, ROM), random access memory (RandomAccessMemory, RAM), flash memory (FLASH), hard disk or optical disk, etc., including a number of instructions for a computer device (which can be a personal computer, server, or network device, etc.) to perform the methods of various embodiments of the present invention.

[0120] Example 2: The following is an embodiment of the present invention, which provides an adaptive photovoltaic tracking control method based on artificial intelligence. In order to verify the beneficial effects of the present invention, scientific demonstration is carried out through simulation experiments.

[0121] The experiment was conducted through data simulation. The geographical location was set at latitude 30.67 and longitude 104.06. A dual-axis tracking photovoltaic system was selected. The test time was from July 1 to July 31. The data collection interval was once every hour, and a total of 31 days of data were collected.

[0122] According to the geographical location, the solar position algorithm (SPA) is used to calculate the real-time position of the sun, determine the real-time position of the sun in the sky and the ideal light intensity, and the control data and adjustment data are the historical data of a photovoltaic power station. A simulation experiment is carried out according to the method of the present invention, and some data are shown in the following table.

[0123] Table 1 Experimental data table

[0124]

[0125] From the table data, we can see that at 12:00: the light intensity reaches 85000lx, the actual photovoltaic efficiency is 19.0%, and the theoretical light intensity is 88000lx, with a small difference, indicating that the photovoltaic tracking system can effectively capture solar energy. At 15:00: the light intensity drops to 78000lx, and the actual photovoltaic efficiency drops to 17.8% accordingly. Although the light intensity drops, the efficiency drops less after adjusting the model optimization and remains at a high level.

[0126] The azimuth angle indicates the horizontal rotation angle of the photovoltaic module relative to the north. The data shows that from 10:00 to 15:00, the azimuth angle gradually increases from 120° (southeast) to 195° (southwest), reflecting that the photovoltaic system constantly adjusts its direction throughout the day to track the movement of the sun.

[0127] The tilt angle refers to the vertical tilt angle of the PV module relative to the horizontal plane. The data show that from 10:00 to 15:00, the tilt angle gradually decreases from 30° to 10°, and then increases slightly to 15°, reflecting that the PV system adjusts the tilt angle of the module during the day to maximize the receipt of sunlight.

[0128] In summary, the method of the present invention can effectively adjust the azimuth and inclination of photovoltaic modules under different time periods and environmental conditions, maintain a high power generation efficiency, comprehensively consider multiple factors such as light intensity, temperature, wind speed, humidity and atmospheric pressure, and realize precise adjustment and optimization of photovoltaic modules through intelligent algorithms, which significantly improves the photovoltaic power generation efficiency and enhances the stability and reliability of the system.

[0129] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention rather than to limit it. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical solutions of the present invention may be modified or replaced by equivalents without departing from the spirit and scope of the technical solutions of the present invention, which should all be included in the scope of the claims of the present invention.

Claims

1. An adaptive photovoltaic tracking control method based on artificial intelligence, characterized in that: include: Collect the data of photovoltaic modules, divide the data of photovoltaic modules into control data and adjustment data, and pre-process the data; Determine the main orientation of the PV modules based on control data and predict environmental changes based on adjustment data; An adjustment model is established, with photovoltaic power generation efficiency as the target, to calculate the angle and direction adjustment values ​​that need to be fine-tuned, and adjust the angle of the photovoltaic module in real time.

2. The adaptive photovoltaic tracking control method based on artificial intelligence according to claim 1, characterized in that: The data of the photovoltaic module is divided into control data and adjustment data. The control data includes light intensity, atmospheric thickness and geographical location data of the photovoltaic module, and the adjustment data includes temperature, wind speed, humidity and atmospheric pressure.

3. The adaptive photovoltaic tracking control method based on artificial intelligence according to claim 2, characterized in that: The main directions of the photovoltaic module include using a solar position calculation algorithm to determine the real-time position of the sun in the sky and the ideal light intensity according to the geographical location and time of the photovoltaic module. The solar altitude angle g and azimuth angle α are expressed as: g=arcsin(sin(Latitude)·sin(δ)+cos(Latitude)·cos(δ)·cos(H)) Ideal light intensity I t It is expressed as: I t =I0·cos(g)·e -τ·m Among them, Latitude represents the latitude of the photovoltaic module; δ represents the solar declination; H represents the hour angle; I0 represents the solar constant; τ represents the atmospheric optical thickness; m represents the air quality coefficient; Ideally, the PV panel tilt angle β is perpendicular to the sun's altitude angle, and the azimuth angle γ is the same as the sun's azimuth angle, expressed as: β＝90°-g γ=α In actual situations, light intensity is indirectly affected by temperature and humidity. The effects of temperature and humidity on environmental factor E are expressed as: Where T represents the total time; T(t) represents the function of temperature changing with time t; H represents the total humidity; H t represents the humidity at time t; Calculate the deviation rate threshold ΔI between the actual light intensity and the theoretical light intensity based on environmental factors t , expressed as: Among them, ΔI base represents the artificially set benchmark deviation; k represents the adjustment coefficient; I t,t represents the ideal light intensity at time t; I actual,t Indicates the actual light intensity at time t; when I actual,t -I t,t ≤ΔI t When I actual,t -I t,t >ΔI t When , it means that the light intensity deviation is too large, and the PV panels are fine-tuned by adjusting the model.

4. The artificial intelligence-based adaptive photovoltaic tracking control method according to claim 3, characterized in that: The prediction of environmental changes includes taking the adjustment data as input, using random forest regression to predict the adjustment data one hour later, evaluating the model performance through cross-validation, dividing the adjustment data into K parts, each of which is used as a validation set in turn, and the remaining K-1 parts are used as training sets, repeating K times, and calculating the average performance index, which is expressed as: Among them, CV score represents the cross-validation score; K represents the number of cross-validation folds; score i Represents the performance score of the i-th fold; a judgment is made based on the predicted adjustment data. If the predicted adjustment data will reduce the efficiency of the PV module, the PV module is fine-tuned in advance through the adjustment model to avoid efficiency reduction.

5. The artificial intelligence-based adaptive photovoltaic tracking control method according to claim 4, characterized in that: The adjustment model includes establishing an adjustment model based on the proximal strategy optimization PPO, and the output data is the angle adjustment amount β t , Input data t The angle and adjustment data of the current photovoltaic module. When the light intensity deviation is too large, the input adjustment data is the real-time data for real-time adjustment. When the predicted adjustment data will reduce the efficiency of the photovoltaic module, the input data is the predicted adjustment data for advance adjustment. Taking photovoltaic power generation efficiency as the goal, the reward function is designed according to photovoltaic power generation efficiency, system stability and energy consumption factors; Set constraints, use the reward function to iteratively train the adjustment model, use the trained adjustment model to output the best angle adjustment amount, adjust the angle of the photovoltaic module, and the adjustment model is expressed as: Where L(θ) represents the objective function of the PPO algorithm; represents the expected value; π θ (β t |s t ) means in state s t In this case, policy π uses parameters θ to select actions β t The probability of Indicates that in state s t Under this condition, the strategy π uses the old parameters θ old Select action β t The probability of represents the advantage function; f represents the reward function; ∈ represents the clipping range parameter.

6. The artificial intelligence-based adaptive photovoltaic tracking control method according to claim 5, characterized in that: The reward function includes power generation efficiency reward, system stability reward and energy consumption penalty. The power generation efficiency reward is given according to the actual power generation, so that the model can choose the angle adjustment that can maximize the power generation, which is expressed as: Among them, R P represents the power generation efficiency reward function; P(t) represents the actual power generation at time t; P max Indicates the maximum possible power generation; System stability reward, rewards the smoothness of the system's angle adjustment, avoids frequent and large angle adjustments, and protects the system's mechanical components. It is expressed as: Among them, R s represents the system stability reward function; δ represents the weight factor; β(t) represents the tilt angle of the PV module at time t; Energy consumption penalty, which penalizes the energy consumption of each angle adjustment, encourages the model to minimize the energy consumption of the control system while maintaining efficient power generation, expressed as: Among them, R represents the energy consumption penalty function; ρ represents the weight factor; Q(t) represents the energy consumption at time t; σ represents the parameter for adjusting the exponential function; The reward function is expressed as: f=R P +R s -R Where f represents the reward function.

7. The artificial intelligence-based adaptive photovoltaic tracking control method according to claim 6, characterized in that: The restriction condition includes that the range of the photovoltaic module angle is limited to 0° to 90°, which is expressed as: 0°≤β+β t ≤90° Among them, β t Indicates the angle adjustment amount for adjusting the model calculation; Set wind speed thresholds A and B according to the physical strength of the photovoltaic module, and A < B. When the wind speed exceeds A, lock the fine-tuning function of the adjustment model and only perform the main direction adjustment of the photovoltaic module. When the wind speed exceeds B, lock the angle of the photovoltaic module to the horizontal position and do not perform any adjustment; Set the upper limit of the adjustment frequency and the amplitude of each adjustment to avoid mechanical damage caused by frequent adjustments.

8. An artificial intelligence-based adaptive photovoltaic tracking control system using the method as claimed in any one of claims 1 to 7, characterized in that: Including, A data acquisition module that acquires data of the photovoltaic module, divides the data into two parts: control data and adjustment data, and performs data cleaning to ensure the integrity and accuracy of the data; A direction control module that uses the preprocessed control data to analyze the direction information of the photovoltaic module during the day, determines the main orientation of the photovoltaic module, and predicts the impact of environmental changes on photovoltaic power generation; A fine-tuning module that aims at the photovoltaic power generation efficiency, based on the environmental change prediction results and real-time data, establishes an adjustment model, calculates the angle adjustment value that needs to be fine-tuned, and adjusts the angle of the photovoltaic module in real time according to the output of the adjustment model to ensure that the photovoltaic module can maintain the best power generation efficiency under different environmental conditions.

9. A computing device comprising: A memory and a processor; The memory is used to store computer-executable instructions, and the processor is used to execute the computer-executable instructions. When the computer-executable instructions are executed by the processor, the steps of the method described in any one of claims 1 to 7 are implemented.

10. A computer-readable storage medium that stores computer-executable instructions. When the computer-executable instructions are executed by a processor, the steps of the method described in any one of claims 1 to 7 are implemented.

Images