Method and apparatus for controlling rotation of a tracking axis

Abstract

The application provides a method and device for controlling rotation of a tracking axis, when a plurality of target control algorithms are detected, target data sources required by each target control algorithm are acquired; based on each target control algorithm and the target data source corresponding thereto, a tracking angle corresponding to each target control algorithm is calculated; each tracking angle is input into a pre-trained tracking angle prediction model, so that the pre-trained tracking angle prediction model determines a target tracking angle according to each tracking angle; wherein the pre-trained tracking angle prediction model is obtained by training a tracking angle prediction model to be trained by using historical data sources required by each target control algorithm to determine each historical tracking angle; and the tracking axis of a photovoltaic tracking system is controlled to rotate towards the target tracking angle.

Description

Technical Field

[0001] This invention relates to the field of tracking photovoltaic technology, and specifically to a method and apparatus for controlling the rotation of a tracking axis. Background Technology

[0002] How to adjust the tracking axis angle to maximize the power generation of a photovoltaic power plant is one of the most pressing issues in the field of photovoltaic tracking. Current photovoltaic tracking systems typically employ intelligent algorithms based on external sensor data to control the tracking axis to face the sun, thereby increasing the system's power output.

[0003] However, the method of using intelligent algorithms based on external sensor data to control the tracking axis does not take into account external factors such as dust accumulation, aging, and damage to the external sensors of the photovoltaic tracking system. This can easily lead to errors in the calculated tracking axis angle, making it impossible to control the tracking axis based on the calculated tracking axis angle to maximize the power generation of the photovoltaic power station, resulting in a significant loss of power generation. Summary of the Invention

[0004] In view of this, embodiments of this application provide a method and apparatus for controlling the rotation of a tracking axis to solve the problem that controlling the tracking axis using intelligent algorithms based on external sensor data can easily lead to significant power generation losses.

[0005] To achieve the above objectives, the embodiments of this application provide the following technical solutions:

[0006] The first aspect of this application provides a method for controlling the rotation of a tracking axis, applied to a tracking controller, the method comprising:

[0007] When multiple target control algorithms are detected, the target data source required by each target control algorithm is obtained;

[0008] Based on each target control algorithm and its corresponding target data source, calculate the tracking angle corresponding to each target control algorithm;

[0009] Each of the tracking angles is input into a pre-trained tracking angle prediction model so that the pre-trained tracking angle prediction model determines the target tracking angle based on each of the tracking angles; wherein, the pre-trained tracking angle prediction model is obtained by training the tracking angle prediction model to be trained using each historical tracking angle determined by the historical data source required by each target control algorithm.

[0010] The tracking axis of the photovoltaic tracking system is controlled to rotate toward the target tracking angle.

[0011] Optionally, the method further includes:

[0012] During the process of controlling the tracking axis to rotate toward the target tracking angle, the first tracking angle when the current of the photovoltaic subarray of the photovoltaic tracking system is at its maximum is recorded;

[0013] Calculate the corresponding root mean square error based on the first tracking angle and the target tracking angle;

[0014] If the root mean square error is not within the preset threshold range, each of the tracking angles is taken as the historical tracking angles, and the first tracking angle is taken as the actual optimal tracking angle.

[0015] The pre-trained tracking angle prediction model is updated and iterated using each of the historical tracking angles and the corresponding actual tracking angles.

[0016] Optionally, the tracking angle prediction model to be trained is trained using various historical tracking angles to obtain the pre-trained tracking angle prediction model, including:

[0017] Obtain the historical data source required for each target control algorithm, and use each target control algorithm and its corresponding historical data source to calculate the historical tracking angle corresponding to each target control algorithm;

[0018] Calculate the historical tracking angle difference between every two historical tracking angles;

[0019] If the maximum value of each of the historical tracking angle differences is greater than a preset threshold, the actual optimal tracking angle is determined using each of the historical tracking angles.

[0020] Input each of the historical tracking angles and the corresponding actual best tracking angle into the tracking angle prediction model to be trained;

[0021] The tracking angle prediction model to be trained predicts the second tracking angle based on each of the historical tracking angles.

[0022] If the historical root mean square error calculated based on the second tracking angle and the corresponding actual best tracking angle is within a preset threshold range, the parameters of the tracking angle prediction model to be trained are adjusted based on the second tracking angle and the corresponding actual best tracking angle until the tracking angle prediction model to be trained converges, thus obtaining the pre-trained tracking angle prediction model.

[0023] Optionally, the tracking angle prediction model to be trained is a neural network algorithm or a fitting algorithm.

[0024] Optionally, the method further includes:

[0025] If the maximum value among the tracking angle differences is not greater than the preset threshold, any one of the historical tracking angles will be taken as the actual optimal tracking angle.

[0026] Optionally, the actual optimal tracking angle can be determined using the various historical tracking angles, including:

[0027] The tracking axis is controlled to rotate sequentially based on the tracking angles from smallest to largest, and the actual optimal tracking angle when the current of the photovoltaic subarray of the photovoltaic tracking system is at its maximum during the rotation of the tracking axis is recorded.

[0028] Optionally, the target data source required for each of the target control algorithms includes:

[0029] For each target control algorithm, obtain the initial data source required by the target control algorithm within the current preset time interval;

[0030] Calculate the average value of the initial data source corresponding to the target control algorithm to obtain the target data source of the target control algorithm.

[0031] Optionally, before obtaining the target data source required for each of the target control algorithms, the method further includes:

[0032] By tracking the communicator, it can be determined whether there is a target control algorithm that cannot obtain the required target data source;

[0033] If no target control algorithm exists that can acquire the required target data source, the tracking communicator selects the control algorithm with the highest priority from among the control algorithms as the optimal control algorithm based on the priority of each control algorithm; wherein each control algorithm is a control algorithm that can acquire the required data source.

[0034] Accordingly, obtaining the target data source required for each of the target control algorithms includes:

[0035] If no target control algorithm exists and the required target data source cannot be obtained, obtain the target data source required for each target control algorithm.

[0036] Optionally, the method further includes:

[0037] When a target control algorithm or an optimal control algorithm is detected, the target data source required by the target control algorithm or the optimal control algorithm is obtained;

[0038] Based on the target control algorithm or the optimal control algorithm and its target data source, the tracking angle corresponding to the target control algorithm or the optimal control algorithm is determined, and the tracking axis is controlled to rotate toward the tracking angle corresponding to the target control algorithm or the optimal control algorithm.

[0039] A second aspect of this application provides a device for controlling the rotation of a tracking axis, applied to a tracking controller, the device comprising:

[0040] The first target data source acquisition unit is used to acquire the target data source required for each target control algorithm when multiple target control algorithms selected by the user are detected.

[0041] The tracking angle calculation unit is used to calculate the tracking angle corresponding to each target control algorithm based on each target control algorithm and its corresponding target data source;

[0042] The target tracking angle determination unit is used to input each of the tracking angles into a pre-trained tracking angle prediction model, so that the pre-trained tracking angle prediction model determines the target tracking angle based on each of the tracking angles; wherein, the pre-trained tracking angle prediction model is obtained by training the tracking angle prediction model to be trained based on each historical tracking angle and the corresponding actual best tracking angle determined by the training unit using the historical data source required by each target control algorithm.

[0043] The first control unit is used to control the tracking axis of the photovoltaic tracking system to rotate toward the target tracking angle.

[0044] Based on the above-described embodiments of the present invention, a method and apparatus for controlling the rotation of a tracking axis can be provided. This method pre-trains a tracking angle prediction model using historical tracking angles and corresponding optimal tracking angles determined from historical data sources required by each target control algorithm. This pre-trained model allows for the acquisition of target data sources required by each target control algorithm after multiple target control algorithms are detected. Based on each target control algorithm and its corresponding target data source, the tracking angle corresponding to each target control algorithm is calculated. Each tracking angle is then input into the pre-trained tracking angle prediction model, enabling the model to determine the target tracking angle. Finally, the tracking axis of the photovoltaic tracking system is controlled to rotate towards the target tracking angle. In this technical solution, each target control algorithm calculates the corresponding tracking angle based on the required target data source. The pre-trained tracking angle prediction model then determines the final target tracking angle based on the tracking angles calculated by multiple target control algorithms. This ensures more precise angle control of the photovoltaic tracking axis, and the corresponding control of the tracking axis to rotate towards the calculated target tracking angle further enhances the power generation of the photovoltaic tracking system, maximizing its power output. Attached Figure Description

[0045] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on the provided drawings without creative effort.

[0046] Figure 1 A flowchart illustrating a method for controlling the rotation of a tracking axis, provided in an embodiment of this application;

[0047] Figure 2 A flowchart illustrating a training method for a pre-trained tracking angle prediction model provided in an embodiment of this application;

[0048] Figure 3 A flowchart illustrating another method for controlling the rotation of a tracking axis provided in an embodiment of this application;

[0049] Figure 4 This is a schematic diagram of a device for controlling the rotation of a tracking axis, provided in an embodiment of this application. Detailed Implementation

[0050] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0051] In this application, the terms "comprising," "including," or any other variations thereof are intended to cover a non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitation, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes said element.

[0052] The inventors discovered that existing photovoltaic tracking systems typically use astronomical algorithms to calculate the sun's position based on time and geographical parameters, then adjust the tracking axis's rotation angle to align it directly with the sun. Alternatively, they use intelligent algorithms based on external sensor data to control the tracking axis, specifically calculating the tracking angle for different weather conditions based on astronomical algorithms. However, neither method considers external factors such as dust accumulation, aging, damage, and decreased accuracy of external sensors (e.g., irradiance meters, sky cameras). This can easily lead to misjudgment of data sources, resulting in significant power generation losses.

[0053] Therefore, embodiments of the present invention provide a method and apparatus for controlling the rotation of a tracking axis. After calculating the corresponding tracking angle based on the required target data source using various target control algorithms, the final target tracking angle is determined by using a pre-trained tracking angle prediction model based on the tracking angles calculated by multiple target control algorithms. This ensures more precise angle control of the tracking axis of the photovoltaic tracking system. The corresponding control of the tracking axis to rotate toward the calculated target tracking angle can further improve the power generation of the photovoltaic tracking system and maximize its power generation.

[0054] The following detailed description of the solution is provided through various embodiments.

[0055] See Figure 1 The diagram illustrates a flowchart of a method for controlling the rotation of a tracking axis according to an embodiment of this application. This method for controlling the rotation of a tracking axis is applied to a tracking controller and specifically includes the following steps:

[0056] S101: When multiple target control algorithms are detected, obtain the target data source required by each target control algorithm.

[0057] In the embodiments of this application, the multiple control algorithms include, but are not limited to, standard component or standard string voltage and current data control algorithms, inverter input or output power control algorithms, irradiator data algorithms, and sky camera data algorithms.

[0058] It should be noted that the data source required for the standard component or standard string voltage and current data control algorithm includes at least the real-time operating current and real-time operating voltage of the target device; wherein, the real-time operating current and real-time operating voltage can be the real-time operating current and real-time operating voltage of the target device at the maximum power point, or the real-time open-circuit current and real-time open-circuit voltage of the target device, or any real-time operating current and real-time operating voltage of the target device at the current moment, and the specific situation of the real-time operating current and real-time operating voltage of the target device is not limited here.

[0059] It should also be noted that the target device is either a photovoltaic module (standard module) or a photovoltaic string (standard string) of a photovoltaic tracking system. When the target device is a photovoltaic module, it can be a module of the same specifications as the power station or a customized module. The length of the target device is the same as the length of the module in the power station. When the target device is a photovoltaic string, it can be any photovoltaic string on the same tracking bracket in the power station.

[0060] The data source required for the inverter input or output power control algorithm includes at least the input or output power of the photovoltaic subarray of the photovoltaic tracking system; wherein, the input or output power of the photovoltaic subarray can be DC input power or DC output power, or AC input power or AC output power, and the specific situation of the input or output power of the photovoltaic subarray is not limited here.

[0061] The data source required for the irradiance meter data algorithm includes at least the solar irradiance near the photovoltaic subarray of the photovoltaic tracking system; wherein, the solar irradiance near the photovoltaic subarray can be the total horizontal irradiance near the photovoltaic subarray, or the total irradiance of the inclined plane where the tracking bracket is located near the photovoltaic subarray, or the total horizontal irradiance and the horizontal diffuse irradiance near the photovoltaic subarray. The specific situation of the solar irradiance of the photovoltaic subarray is not limited here.

[0062] The data source required for the sky camera data algorithm includes at least sky cloud data above the photovoltaic subarray; the sky cloud data may include the number of clouds in the sky, the thickness and color of each cloud, etc., and the specific content of the sky cloud data is not limited here.

[0063] In the specific implementation step S101, the user can send the target control algorithm selected from various control algorithms to the tracking controller through the tracking communicator; when the tracking controller detects that the user has selected multiple target control algorithms, it can determine the multiple target control algorithms selected by the user from the various control algorithms, and obtain the data source required for each target control algorithm (for ease of distinction, the data source corresponding to each target control algorithm is called the target data source). Among them, the multiple target control algorithms include at least two target control algorithms.

[0064] Optionally, in this embodiment of the application, for each target control algorithm, the initial data source required by the target control algorithm within the current preset time interval can be obtained, and the average value of the initial data source corresponding to the target control algorithm can be calculated to obtain the target data source of the target control algorithm.

[0065] In practical applications, before acquiring the target data source required by each target control algorithm, the tracking communicator determines whether there are target control algorithms that cannot acquire the required target data source; if there are no target control algorithms that cannot acquire the required target data source, the tracking controller acquires the target data source required by each target control algorithm.

[0066] If no target control algorithm exists that can acquire the required target data source, the tracking communicator selects the highest priority control algorithm as the optimal control algorithm based on the priority of each control algorithm. This allows the tracking controller to determine the tracking angle corresponding to the optimal control algorithm based on the target data source required by the optimal control algorithm. The tracking axis is then controlled to rotate toward the tracking angle corresponding to the optimal control algorithm. Here, each control algorithm is one that can acquire the required data source.

[0067] If no control algorithm exists that can obtain the required data source, the tracking controller can use an astronomical algorithm to calculate the sun's position based on information such as time and geographical parameters, and then adjust the rotation angle of the tracking axis to face the sun.

[0068] Furthermore, in this embodiment of the application, when a target control algorithm is detected and it can be determined through the tracking communicator that the target data source required by the target control algorithm can be obtained, the tracking controller obtains the target data source required by the target control algorithm, and determines the tracking angle corresponding to the target control algorithm based on the target control algorithm and its target data source, so as to control the tracking axis to rotate toward the tracking angle corresponding to the target control algorithm.

[0069] S102: Calculate the tracking angle corresponding to each target control algorithm based on each target control algorithm and its corresponding target data source.

[0070] During the specific execution of step S102, after the tracking axis controller obtains the target data source required by each target control algorithm, it can calculate the tracking angle corresponding to the target control algorithm based on the target control algorithm and its corresponding target data source.

[0071] As one embodiment of this application, if the target control algorithm is a standard component or standard string voltage and current data control algorithm, then the corresponding target data source includes at least the real-time operating voltage and real-time operating current of the target device. Therefore, based on the real-time operating voltage and current of the target device and the standard component or standard string voltage and current data control algorithm, the tracking angle corresponding to the standard component or standard string voltage and current data control algorithm can be calculated. The target device is a photovoltaic module or photovoltaic string in a photovoltaic tracking system.

[0072] As another implementation method in this application, if the target control algorithm is an inverter input or output power control algorithm, the corresponding target data includes at least the input power or output power of the photovoltaic subarray of the photovoltaic tracking system. Then, the tracking angle corresponding to the inverter input or output power control algorithm can be calculated based on the inverter input or output power control algorithm and the input power or output power of the photovoltaic subarray.

[0073] As another implementation method in this application, if the target control algorithm is an irradiance data algorithm, the corresponding target data source includes at least the solar irradiance of the photovoltaic subarray of the photovoltaic tracking system. Then, the tracking angle corresponding to the irradiance data algorithm can be calculated based on the solar irradiance of the photovoltaic subarray and the irradiance data algorithm.

[0074] As another embodiment of this application, if the target control algorithm is a sky camera data algorithm, the corresponding target data source includes the sky cloud data above the photovoltaic subarray of the photovoltaic tracking system. Then, the tracking angle corresponding to the sky camera data algorithm can be calculated based on the sky cloud data above the photovoltaic subarray and the sky camera data algorithm.

[0075] S103: Input each tracking angle into the pre-trained tracking angle prediction model so that the pre-trained tracking angle prediction model can determine the target tracking angle based on each tracking angle.

[0076] In this embodiment of the application, the tracking angle prediction model to be trained can be pre-trained by using the historical tracking angles and corresponding actual optimal tracking angles determined by the historical data sources required by each target control algorithm.

[0077] See Figure 2 The diagram illustrates a flowchart of a training method for a pre-trained tracking angle prediction model provided in an embodiment of this application. The method specifically includes the following steps:

[0078] S201: Obtain the historical data source required for each target control algorithm.

[0079] During the specific execution of step S201, after the tracking controller detects multiple target control algorithms selected by the user, it can obtain the historical initial data source required by the target control algorithm within a historical preset time interval for each target control algorithm, and calculate the average value of the historical initial data source to obtain the historical target data source of the target control algorithm.

[0080] It should be noted that multiple control algorithms may include standard component or standard string voltage and current data control algorithms, inverter input or output power control algorithms, irradiance data algorithms, and sky camera data algorithms.

[0081] The specific contents of the target data source required for each target control algorithm can be found in the embodiments disclosed above. Figure 1 The details of step S101 are not repeated here.

[0082] S202: Using each target control algorithm and its corresponding historical data source, calculate the historical tracking angle corresponding to each target control algorithm, and calculate the historical tracking angle difference between every two historical tracking angles.

[0083] During the specific execution of step S202, after the tracking axis controller obtains the historical target data source required by each target control algorithm, it can calculate the historical tracking angle corresponding to the target control algorithm based on the target control algorithm and its corresponding historical target data source, and calculate the historical tracking angle difference between every two historical tracking angles.

[0084] In some embodiments, when the tracking controller detects that multiple target control algorithms selected by the user include a standard component or standard string voltage and current data control algorithm, an inverter input or output power control algorithm, an irradiance data algorithm, and a sky camera data algorithm, the historical target data source required by the corresponding standard component or standard string voltage and current data control algorithm shall at least include the historical real-time operating voltage and historical real-time operating current of the target device; the historical target data source required by the inverter input or output power control algorithm shall at least include the historical input power or historical output power of the photovoltaic subarray of the photovoltaic tracking system; the historical target data source required by the irradiance data algorithm shall at least include the historical solar irradiance of the photovoltaic subarray; and the historical target data source required by the sky camera data algorithm shall at least include historical cloud data above the photovoltaic subarray.

[0085] Based on the target device's historical real-time operating voltage, historical real-time operating current, and standard component or standard string voltage and current data control algorithm, calculate the historical tracking angle corresponding to the standard component or standard string voltage and current data control algorithm. sta-pv Based on the historical input or output power of the photovoltaic subarray and the historical inverter input or output power control algorithm, calculate the historical tracking angle corresponding to the inverter input or output power control algorithm. inv Based on the historical solar irradiance of the photovoltaic subarray and the algorithm of the irradiometer data, the historical tracking angle corresponding to the irradiometer data algorithm is calculated. irra Based on historical cloud data above the photovoltaic subarray and sky camera data algorithms, the historical tracking angle corresponding to the sky camera data algorithm is calculated. camerra Finally, calculate the historical tracking angle difference between every two historical tracking angles.

[0086] S203: Determine whether the maximum value of each historical tracking angle difference is greater than a preset threshold; if the maximum value of each tracking angle difference is not greater than the preset threshold, proceed to step S204; if the maximum value of each historical tracking angle difference is greater than the preset threshold, proceed to step S205.

[0087] During the specific execution of step S203, a corresponding preset threshold can be set in advance so that after calculating the historical tracking angle difference between each historical tracking angle, it can be further determined whether the maximum value of each historical tracking angle difference is greater than the preset threshold.

[0088] If the maximum value among the differences of each historical tracking angle is not greater than the preset threshold, it indicates that the historical tracking angles calculated by each target control algorithm are all better. Then, step S204 can be executed to take any one of the historical tracking angles as the actual best tracking angle.

[0089] If the maximum value among the differences of each historical tracking angle is greater than the preset threshold, it indicates that there are poor historical tracking angles among the historical tracking angles calculated by each control algorithm. Then, step S205 can be executed to further determine the actual best tracking angle using each historical tracking angle.

[0090] S204: Take any one of the historical tracking angles as the actual best tracking angle.

[0091] S205: Determine the actual optimal tracking angle using various historical tracking angles.

[0092] In this embodiment of the application, if the maximum value of each historical tracking angle difference is not greater than a preset threshold, the historical tracking angles can be further sorted from smallest to largest so that the tracking axis can be controlled to rotate sequentially according to each historical tracking angle from smallest to largest, and the actual optimal tracking angle corresponding to the maximum current of the photovoltaic subarray during the rotation of the tracking axis can be recorded.

[0093] S206: Input the historical tracking angles and the actual best tracking angle into the tracking angle prediction model to be trained.

[0094] In the specific execution step S206, after recording the corresponding actual best tracking angle, the historical tracking angles and the actual best tracking angles can be further input into the tracking angle prediction model to be trained, so that the tracking angle prediction model to be trained can use the input historical tracking angles and the actual best tracking angles to be trained until the tracking angle prediction model to be trained converges, and the final pre-trained tracking angle prediction model is obtained.

[0095] S207: The tracking angle prediction model to be trained predicts the second tracking angle based on each historical tracking angle, and determines whether the historical root mean square error calculated based on the second tracking angle and the actual best tracking angle is within the preset threshold range; if the historical root mean square error is within the preset threshold range, proceed to step S208; if the historical root mean square error is not within the preset threshold range, return to step S201.

[0096] In this embodiment of the application, the tracking angle prediction model to be trained can be a neural network algorithm or a fitting algorithm, and this embodiment of the application does not limit it.

[0097] As a preferred embodiment of this application, when the tracking angle prediction model to be trained is a fitting algorithm, the specific training process can be as follows: the tracking angle prediction model to be trained can predict the second tracking angle according to a pre-set fitting formula and each historical tracking angle; based on each second tracking angle within the historical target time period and its corresponding actual best tracking angle, the corresponding historical root mean square error is calculated to determine whether the historical root mean square error is within a preset threshold range. Wherein, each second tracking angle within the target time period and its corresponding actual best tracking angle includes the second tracking angle corresponding to a historical preset time interval and its corresponding actual best tracking angle, as well as the second tracking angle corresponding to each historical preset time interval before that historical preset time interval and its corresponding actual best tracking angle.

[0098] If the historical root mean square error is not within the preset threshold range, it can be considered that the fitting relationship obtained at present has not yet reached the accuracy requirement. The fitting relationship is adjusted according to the current historical tracking angles, and the process returns to step S201 to continue training the tracking angle prediction model to be trained until the fitting relationship obtained reaches the corresponding accuracy requirement, at which point step S208 is executed.

[0099] If the historical root mean square error is within the preset threshold range, it can be considered that the fitting relationship obtained by the current fitting has reached the accuracy requirement, and then step S208 can be executed, that is, the parameters of the tracking angle prediction model to be trained are adjusted based on the second tracking angle and the actual best tracking angle to obtain the final pre-trained tracking angle prediction model.

[0100] In this embodiment, the fitting formula is obtained by fitting the second tracking angle corresponding to each historical preset time interval before the previous preset time interval with its corresponding actual optimal tracking angle. The fitting formula can be set according to the actual application, and this embodiment does not limit it.

[0101] It should be noted that the preset accuracy requirement can be: the historical root mean square error calculated based on the second tracking angle corresponding to each historical preset time interval within the historical target time period and its corresponding actual optimal tracking angle is within a preset threshold range. The corresponding accuracy requirement can be set according to the actual application, and this application embodiment does not limit it.

[0102] The calculation method for the root mean square error based on each second tracking angle and its corresponding actual optimal tracking angle within the target time period is shown in formula (1):

[0103]

[0104] Where RMES is the historical root mean square error, n is the number of historical preset time intervals included in the historical target time period, and angle is... best-i For the i-th historical preset time interval, angle is the actual optimal tracking angle. nesbest-i The second tracking angle is the i-th historical preset time interval.

[0105] As another preferred embodiment of this application, when the tracking angle prediction model to be trained is a neural network algorithm or a fitting algorithm, the specific training process can be as follows: the tracking angle prediction model to be trained can calculate the second tracking angle corresponding to a historical preset time interval based on the weight factors corresponding to each target control algorithm and each historical tracking angle set in advance; based on the second tracking angle corresponding to each historical preset time interval within the historical target period and its corresponding actual best tracking angle, the corresponding historical root mean square error can be calculated to determine whether the historical root mean square error is within a preset threshold range. The weight factors corresponding to each target control algorithm are determined based on the historical tracking angles corresponding to each target control algorithm in each historical preset time interval before the current historical preset time interval and their corresponding actual best tracking angles.

[0106] If the historical root mean square error is not within the preset threshold range, it can be considered that the currently set weight factors have not met the corresponding accuracy requirements. The weight factors corresponding to each target control algorithm can be adjusted according to the historical tracking angles corresponding to the preset historical time intervals and their corresponding actual best tracking angles. Then, return to step S201 to continue training the tracking angle prediction model to be trained until the corresponding accuracy requirements are met, and then execute step S208.

[0107] If the historical root mean square error is within the preset threshold range, it can be considered that the weight factors of each target control algorithm have reached the accuracy requirements. Then, step S208 can be executed, that is, the parameters of the tracking angle prediction model to be trained are adjusted based on the second tracking angle and the actual best tracking angle to complete the training or update iteration of the target tracking angle prediction model and obtain the final pre-trained tracking angle prediction model.

[0108] It should be noted that the tracking angle prediction model to be trained can be a neural network algorithm or other machine learning algorithm. The appropriate tracking angle prediction model to be trained can be selected according to actual needs, and this application embodiment does not limit it.

[0109] The above are merely preferred methods for training the tracking angle prediction model to be trained, as provided in the embodiments of this application. The specific training method for the tracking angle prediction model to be trained can be selected according to actual needs, and the embodiments of this application do not limit it.

[0110] S208: The parameters of the tracking angle prediction model to be trained are adjusted based on the second tracking angle and the actual best tracking angle until the tracking angle prediction model to be trained converges, thus obtaining the pre-trained tracking angle prediction model.

[0111] S104: Controls the tracking axis to rotate toward the target tracking angle.

[0112] During the specific execution of step S104, after the tracking controller determines the target tracking angle based on each tracking angle using the pre-trained tracking angle prediction model, it can control the tracking axis of the photovoltaic tracking system to rotate toward the target tracking angle.

[0113] Based on the above-described embodiment of the invention, a method for controlling the rotation of a tracking axis can be provided. This method utilizes various target control algorithms to calculate the corresponding tracking angle based on the required target data source. Then, a pre-trained tracking angle prediction model is used to determine the final target tracking angle based on the tracking angles calculated by multiple target control algorithms. This ensures more precise angle control of the tracking axis in the photovoltaic tracking system. The corresponding control of the tracking axis towards the calculated target tracking angle further enhances the power generation of the photovoltaic tracking system, maximizing its output. Furthermore, during the process of controlling the tracking axis to rotate towards the target tracking angle, the method can also record the first time the current of the photovoltaic subarray of the photovoltaic tracking system reaches its maximum. The tracking angle is calculated, and if the root mean square error calculated based on the first tracking angle and the target tracking angle is not within the preset threshold range, the pre-trained tracking angle prediction model is updated and iterated again using each tracking angle and its corresponding first tracking angle. This ensures that the pre-trained tracking angle prediction model is self-intersecting in a closed loop and can continuously learn itself, thereby adapting to the target data source corresponding to the target control algorithm in different regions, scenarios, seasons, and time periods. This further ensures that the angle control of the tracking axis of the photovoltaic tracking system is more accurate, and the corresponding control tracking axis rotates towards the calculated target tracking angle, which can further improve the power generation of the photovoltaic tracking system, and maximize the power generation of the photovoltaic tracking system.

[0114] Based on the method for controlling the rotation of the tracking axis provided in the above embodiments of this application, in order to further ensure the prediction accuracy of the pre-trained tracking angle prediction model, the embodiments of this application may further perform the following steps, such as... Figure 3 As shown, the specific steps include:

[0115] S301: When multiple target control algorithms are detected, obtain the target data source required by each target control algorithm.

[0116] S302: Calculate the tracking angle corresponding to each target control algorithm based on each target control algorithm and its corresponding target data source.

[0117] S303: Input each tracking angle into the pre-trained tracking angle prediction model so that the pre-trained tracking angle prediction model can determine the target tracking angle based on each tracking angle.

[0118] S304: Controls the tracking axis to rotate toward the target tracking angle.

[0119] In the specific execution of steps S301 to S304, the specific execution process and implementation principle of steps S301 to S304 are the same as those disclosed in the embodiments of this application above. Figure 1 The specific execution process and implementation principle of steps S101 to S104 are the same as those described above. Figure 1 The relevant content will not be repeated here.

[0120] S305: During the process of controlling the tracking axis to rotate toward the target tracking angle, record the first tracking angle when the current of the photovoltaic subarray of the photovoltaic tracking system is at its maximum.

[0121] In this embodiment of the application, in order to further ensure the prediction accuracy of the pre-trained tracking angle prediction model, the tracking angle when the current of the photovoltaic subarray of the photovoltaic tracking system is at its maximum can be recorded during the process of controlling the tracking axis to rotate toward the target tracking angle. The tracking angle recorded at this time is used as the first tracking angle so as to calculate the corresponding root mean square error based on the first tracking angle and the target tracking angle.

[0122] S306: Calculate the corresponding root mean square error based on the first tracking angle and the target tracking angle.

[0123] In the specific execution of step S306, after recording the first tracking angle of the photovoltaic subarray, the root mean square error can be calculated by obtaining the first tracking angle corresponding to each preset time interval within the current preset time period and its corresponding target tracking angle. The calculation method for the root mean square error based on the first tracking angle corresponding to each preset time interval within the preset time period and its corresponding target tracking angle is shown in formula (2).

[0124]

[0125] Where RMES1 is the historical root mean square error, n is the number of preset time intervals included in the preset time period, and angle1best-i For the first tracking angle during the i-th preset time interval, angle1 nesbest-i The target tracking angle is the i-th preset time interval.

[0126] It should be noted that each preset time interval within the preset time period includes the current preset time interval and each preset time interval preceding the current preset time interval.

[0127] For example, if the current system time is 14:35 on June 7, 2023, and the preset time interval is 5 minutes, then the preset time interval can be from 14:30 on June 7, 2023 to 14:35 on June 7, 2023. Similarly, if the preset time period is 30 minutes, then the preset time intervals within the preset time period can be the preset time intervals from 14:05 on June 7, 2023 to 14:35 on June 7, 2023.

[0128] S307: Determine whether the root mean square error is within the preset threshold range; if the root mean square error is not within the preset threshold range, proceed to step S308.

[0129] In the specific execution of step S307, after calculating the corresponding root mean square error, it can be further determined whether the root mean square error is within the preset threshold range. If the root mean square error is not within the preset threshold range, it can be considered that the prediction accuracy of the pre-trained tracking angle prediction model is insufficient. Then, each tracking angle can be used as each historical tracking angle, and the first tracking angle can be used as the actual best tracking angle. In this way, the pre-trained tracking angle prediction model can be updated and iterated using each historical tracking angle and the corresponding actual tracking angle, thereby further improving the prediction accuracy of the pre-trained tracking angle prediction model.

[0130] If the root mean square error is within the preset threshold range, it can be considered that the prediction accuracy of the pre-trained tracking angle prediction model meets the standard, and there is no need to update and iterate the pre-trained tracking angle prediction model. At this time, we can return to step S301 to continue the control of the rotation of the tracking axis in a new round.

[0131] S308: Take each tracking angle as a historical tracking angle and the first tracking angle as the actual best tracking angle, and use each historical tracking angle and the corresponding actual tracking angle to update and iterate the pre-trained tracking angle prediction model.

[0132] In this embodiment, when the root mean square error is not within a preset threshold range, each tracking angle can be used as a historical tracking angle, and the first tracking angle can be used as the actual optimal tracking angle. The pre-trained tracking angle prediction model is then updated and iterated using each historical tracking angle and the corresponding actual tracking angle. In other words, the pre-trained tracking angle prediction model is retrained using each tracking angle and its corresponding first tracking angle to obtain a pre-trained tracking angle prediction model with higher prediction accuracy. This allows the updated and iterated pre-trained tracking angle prediction model to predict the corresponding target tracking angle, thereby ensuring more precise angle control of the tracking axis of the photovoltaic tracking system. The corresponding control of the tracking axis to rotate toward the calculated target tracking angle can further improve the power generation of the photovoltaic tracking system, maximizing the power generation of the photovoltaic tracking system.

[0133] In this embodiment, the pre-trained tracking angle prediction model is updated and iterated again using each tracking angle and its corresponding first tracking angle. This ensures the closed-loop self-intersection of the pre-trained tracking angle prediction model, enabling it to continuously learn and adapt to the target data sources corresponding to the target control algorithms in different regions, scenarios, seasons, and time periods. This ensures more precise angle control of the tracking axis of the photovoltaic tracking system. The corresponding control tracking axis rotates towards the calculated target tracking angle, which can further improve the power generation of the photovoltaic tracking system, maximizing the power generation of the photovoltaic tracking system.

[0134] In this embodiment, during the process of controlling the tracking axis to rotate towards the tracking angle corresponding to the target control algorithm, the tracking angle at which the current of the photovoltaic subarray is at its maximum during the rotation of the tracking axis towards the tracking angle corresponding to the target control algorithm can be further recorded. If there is a large error between the recorded tracking angle and the tracking angle corresponding to the target control algorithm, it is considered that there is an error in the target data source corresponding to the target control algorithm. At this time, a corresponding warning command can be sent through the tracking communicator to alert the relevant technicians to perform corresponding maintenance and repair. If the error cannot be eliminated after maintenance, it can be determined that the target control algorithm is not suitable for the current project scenario. The target control algorithm can be removed from the aggregation optimization algorithm learning mechanism, that is, the target control algorithm can be deleted from each control algorithm to avoid using the target control algorithm to calculate the corresponding tracking angle in subsequent applications, which would lead to a significant loss of power generation of the photovoltaic tracking system.

[0135] Corresponding to the method for controlling the rotation of a tracking axis provided in the above embodiments of the present invention, see also... Figure 4 The present invention also provides a structural block diagram of a device for controlling the rotation of a tracking axis. This device is applied to a tracking controller and specifically includes the following steps;

[0136] The first target data source acquisition unit 41 is used to acquire the target data source required by each target control algorithm when multiple target control algorithms are detected.

[0137] The tracking angle calculation unit 42 is used to calculate the tracking angle corresponding to each target control algorithm based on each target control algorithm and its corresponding target data source.

[0138] The target tracking angle determination unit 43 is used to input each tracking angle into the pre-trained tracking angle prediction model so that the pre-trained tracking angle prediction model determines the target tracking angle based on each tracking angle; wherein, the pre-trained tracking angle prediction model is obtained by training the tracking angle prediction model to be trained based on each historical tracking angle determined by the training unit using the historical data source required by each target control algorithm.

[0139] The first control unit 44 is used to control the tracking axis of the photovoltaic tracking system to rotate toward the target tracking angle.

[0140] In this embodiment of the invention, after calculating the corresponding tracking angle based on the required target data source using various target control algorithms, the final target tracking angle is determined by using a pre-trained tracking angle prediction model based on the tracking angles calculated by multiple target control algorithms. This ensures that the angle control of the tracking axis of the photovoltaic tracking system is more precise, and the corresponding control tracking axis rotates toward the calculated target tracking angle, which can further improve the power generation of the photovoltaic tracking system and maximize the power generation of the photovoltaic tracking system.

[0141] Preferably, the device for controlling the rotation of the tracking axis further includes:

[0142] The first recording unit is used to record the first tracking angle when the current of the photovoltaic subarray of the photovoltaic tracking system is at its maximum during the process of controlling the tracking axis to rotate toward the target tracking angle;

[0143] The root mean square error calculation unit is used to calculate the corresponding root mean square error based on the first tracking angle and the target tracking angle.

[0144] The determining unit is used to determine each tracking angle as a historical tracking angle and the first tracking angle as the actual best tracking angle if the root mean square error is not within a preset threshold range.

[0145] The update iteration unit is used to update and iterate the pre-trained tracking angle prediction model using each historical tracking angle and the corresponding actual tracking angle.

[0146] Preferably, the training unit includes:

[0147] The historical tracking angle calculation unit is used to obtain the historical data source required by each target control algorithm, and to calculate the historical tracking angle corresponding to each target control algorithm using each target control algorithm and its corresponding historical data source.

[0148] The tracking angle difference calculation unit is used to calculate the historical tracking angle difference between every two historical tracking angles;

[0149] The first actual optimal tracking angle determination unit is used to determine the actual optimal tracking angle by using each historical tracking angle if the maximum value of the differences between each historical tracking angle is greater than a preset threshold.

[0150] The input unit is used to input each historical tracking angle and the corresponding actual best tracking angle into the tracking angle prediction model to be trained.

[0151] The training subunit is used to predict the second tracking angle based on each historical tracking angle using the tracking angle prediction model to be trained. If the historical root mean square error calculated based on the second tracking angle and the corresponding actual best tracking angle is within a preset threshold range, the parameters of the tracking angle prediction model to be trained are adjusted based on the second tracking angle and the corresponding actual best tracking angle until the tracking angle prediction model to be trained converges, thus obtaining the pre-trained tracking angle prediction model.

[0152] Preferably, the tracking angle prediction model to be trained is a neural network algorithm or a fitting algorithm.

[0153] Preferably, the device for controlling the rotation of the tracking axis further includes:

[0154] The second actual optimal tracking angle determination unit is used to determine any one of the historical tracking angles as the actual optimal tracking angle if the maximum value of the differences between the tracking angles is not greater than a preset threshold.

[0155] Preferably, the first actual optimal tracking angle determination unit is specifically used to control the tracking shaft to rotate sequentially based on each tracking angle from small to large, and to record the actual optimal tracking angle when the current of the photovoltaic subarray of the photovoltaic tracking system is at its maximum during the rotation of the tracking shaft.

[0156] Preferably, the first target data source acquisition unit is specifically used to acquire the initial data source required by the target control algorithm within the current preset time interval for each target control algorithm; calculate the average value of the initial data source corresponding to the target control algorithm to obtain the target data source of the target control algorithm.

[0157] If the target control algorithm is a standard component or standard string voltage and current data control algorithm, the tracking angle calculation unit is specifically used for:

[0158] Based on the real-time operating voltage and current of the target device, and the standard component or standard string voltage and current data control algorithm, calculate the tracking angle corresponding to the standard component or standard string voltage and current data control algorithm;

[0159] Among them, the target data source corresponding to the standard component or standard string voltage and current data control algorithm includes the real-time operating voltage and real-time operating current of the target device, and the target device is the photovoltaic module or photovoltaic string of the photovoltaic tracking system.

[0160] Preferably, if the target control algorithm is an inverter input or output power control algorithm, the tracking angle calculation unit is specifically used for:

[0161] Based on the inverter input or output power control algorithm and the input or output power of the photovoltaic subarray, the tracking angle corresponding to the inverter input or output power control algorithm is calculated; wherein, the target data source corresponding to the inverter input or output power control algorithm includes the input or output power of the photovoltaic subarray of the photovoltaic tracking system.

[0162] Preferably, if the target control algorithm is an irradiator data algorithm, the tracking angle calculation unit is specifically used for:

[0163] Based on the solar irradiance of the photovoltaic subarray and the irradiance meter data algorithm, the tracking angle corresponding to the irradiance meter data algorithm is calculated; wherein, the target data source corresponding to the irradiance meter data algorithm includes the solar irradiance of the photovoltaic subarray of the photovoltaic tracking system.

[0164] Preferably, if the target control algorithm is a sky camera data algorithm, the tracking angle calculation unit is specifically used for:

[0165] Based on the cloud data and sky camera data algorithm above the photovoltaic subarray of the photovoltaic tracking system, the tracking angle corresponding to the sky camera data algorithm is calculated; wherein, the target data source corresponding to the sky camera data algorithm includes the cloud data above the photovoltaic subarray.

[0166] Preferably, before executing the first target data source acquisition unit, the device for controlling the rotation of the tracking axis further includes:

[0167] The judgment unit is used to determine, through the tracking communicator, whether there is a target control algorithm that cannot obtain the required target data source;

[0168] The selection unit is used to select the highest priority control algorithm from the control algorithms as the optimal control algorithm if there is no target control algorithm that can obtain the required target data source, through the tracking communicator; wherein, each control algorithm is a control algorithm that can obtain the required data source.

[0169] Correspondingly, the first target data source acquisition unit is specifically used to acquire the target data source required by each target control algorithm if there is no target control algorithm for which the required target data source cannot be acquired.

[0170] Preferably, the device for controlling the rotation of the tracking axis further includes:

[0171] The second target data source acquisition unit is used to acquire the target data source required by the target control algorithm when a target control algorithm or optimal control algorithm is detected.

[0172] The second control unit is used to determine the tracking angle corresponding to the target control algorithm or the optimal control algorithm based on the target control algorithm or the optimal control algorithm and its target data source, and to control the tracking axis to rotate toward the tracking angle corresponding to the target control algorithm or the optimal control algorithm.

[0173] The various embodiments in this specification are described in a progressive manner. Similar or identical parts between embodiments can be referred to mutually. Each embodiment focuses on describing the differences from other embodiments. In particular, for system or system embodiments, since they are basically similar to method embodiments, the description is relatively simple, and relevant parts can be referred to the descriptions in the method embodiments. The systems and system embodiments described above are merely illustrative. The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the modules can be selected to achieve the purpose of this embodiment according to actual needs. Those skilled in the art can understand and implement this without creative effort.

[0174] Those skilled in the art will further recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, computer software, or a combination of both. To clearly illustrate the interchangeability of hardware and software, the components and steps of the various examples have been generally described in terms of functionality in the foregoing description. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementations should not be considered beyond the scope of this invention.

[0175] The above description of the disclosed embodiments enables those skilled in the art to make or use the invention. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the scope of the invention. Therefore, the invention is not to be limited to the embodiments shown herein, but is to be accorded the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. A method for controlling the rotation of a tracking axis, applied to a tracking controller, characterized in that, The method includes: When multiple target control algorithms are detected, the target data source required by each target control algorithm is obtained; Based on each target control algorithm and its corresponding target data source, calculate the tracking angle corresponding to each target control algorithm; Each of the tracking angles is input into a pre-trained tracking angle prediction model so that the pre-trained tracking angle prediction model determines the target tracking angle based on each of the tracking angles; wherein, the pre-trained tracking angle prediction model is obtained by training the tracking angle prediction model to be trained using each historical tracking angle determined by the historical data source required by each target control algorithm. The tracking axis of the photovoltaic tracking system is controlled to rotate toward the target tracking angle.

2. The method according to claim 1, characterized in that, The method further includes: During the process of controlling the tracking axis to rotate toward the target tracking angle, the first tracking angle when the current of the photovoltaic subarray of the photovoltaic tracking system is at its maximum is recorded; Calculate the corresponding root mean square error based on the first tracking angle and the target tracking angle; If the root mean square error is not within the preset threshold range, each of the tracking angles is taken as the historical tracking angle, and the first tracking angle is taken as the actual best tracking angle. The pre-trained tracking angle prediction model is updated and iterated using each of the historical tracking angles and the corresponding actual tracking angles.

3. The method according to claim 1, characterized in that, The tracking angle prediction model to be trained is obtained by training the model using various historical tracking angles, including: Obtain the historical data source required for each target control algorithm, and use each target control algorithm and its corresponding historical data source to calculate the historical tracking angle corresponding to each target control algorithm; Calculate the historical tracking angle difference between every two historical tracking angles; If the maximum value of each of the historical tracking angle differences is greater than a preset threshold, the actual optimal tracking angle is determined using each of the historical tracking angles. Input each of the historical tracking angles and the corresponding actual best tracking angle into the tracking angle prediction model to be trained; The tracking angle prediction model to be trained predicts a second tracking angle based on each of the historical tracking angles. If the historical root mean square error calculated based on the second tracking angle and the corresponding actual best tracking angle is within a preset threshold range, the parameters of the tracking angle prediction model to be trained are adjusted based on the second tracking angle and the corresponding actual best tracking angle until the tracking angle prediction model to be trained converges, thus obtaining the pre-trained tracking angle prediction model.

4. The method according to claim 3, characterized in that, The tracking angle prediction model to be trained is a neural network algorithm or a fitting algorithm.

5. The method according to claim 3, characterized in that, The method further includes: If the maximum value among the tracking angle differences is not greater than the preset threshold, any one of the historical tracking angles will be taken as the actual optimal tracking angle.

6. The method according to claim 3, characterized in that, Determining the actual optimal tracking angle using the aforementioned historical tracking angles includes: The tracking axis is controlled to rotate sequentially based on the tracking angles from smallest to largest, and the actual optimal tracking angle when the current of the photovoltaic subarray of the photovoltaic tracking system is at its maximum during the rotation of the tracking axis is recorded.

7. The method according to claim 1, characterized in that, Obtain the target data source required for each of the target control algorithms, including: For each target control algorithm, obtain the initial data source required by the target control algorithm within the current preset time interval; Calculate the average value of the initial data source corresponding to the target control algorithm to obtain the target data source of the target control algorithm.

8. The method according to claim 1, characterized in that, Before obtaining the target data source required for each of the target control algorithms, the method further includes: By tracking the communicator, it can be determined whether there is a target control algorithm that cannot obtain the required target data source; If no target control algorithm exists that can acquire the required target data source, the tracking communicator selects the control algorithm with the highest priority from among the control algorithms as the optimal control algorithm based on the priority of each control algorithm; wherein each control algorithm is a control algorithm that can acquire the required data source. Accordingly, obtaining the target data source required for each of the target control algorithms includes: If no target control algorithm exists and the required target data source cannot be obtained, obtain the target data source required for each target control algorithm.

9. The method according to claim 8, characterized in that, The method further includes: When a target control algorithm or an optimal control algorithm is detected, the target data source required by the target control algorithm or the optimal control algorithm is obtained; Based on the target control algorithm or the optimal control algorithm and its target data source, the tracking angle corresponding to the target control algorithm or the optimal control algorithm is determined, and the tracking axis is controlled to rotate toward the tracking angle corresponding to the target control algorithm or the optimal control algorithm.

10. A device for controlling the rotation of a tracking axis, applied to a tracking controller, characterized in that, The device includes: The first target data source acquisition unit is used to acquire the target data source required for each target control algorithm when multiple target control algorithms selected by the user are detected. The tracking angle calculation unit is used to calculate the tracking angle corresponding to each target control algorithm based on each target control algorithm and its corresponding target data source; The target tracking angle determination unit is used to input each of the tracking angles into a pre-trained tracking angle prediction model, so that the pre-trained tracking angle prediction model determines the target tracking angle based on each of the tracking angles; wherein, the pre-trained tracking angle prediction model is obtained by training the tracking angle prediction model to be trained based on each historical tracking angle and the corresponding actual best tracking angle determined by the training unit using the historical data source required by each target control algorithm. The first control unit is used to control the tracking axis of the photovoltaic tracking system to rotate toward the target tracking angle.

Images