Photovoltaic water pump inverter control method and system based on deep learning

Abstract

The invention relates to the field of photovoltaic inverter control, and discloses a photovoltaic water pump inverter control method and system based on deep learning, and the method comprises the steps: obtaining the time sequence illumination intensity and time sequence cell panel temperature of a photovoltaic solar panel in a working condition, and obtaining the time sequence output voltage and time sequence output current of an inverter, preset time sequence output pressure of water flow output by the water pump is obtained; wherein electric energy output by the photovoltaic solar panel is converted by the inverter and then is supplied to the water pump; according to the time sequence illumination intensity, the time sequence cell panel temperature, the time sequence output voltage, the time sequence output current and the preset time sequence output pressure, the time sequence prediction output voltage and the time sequence prediction output current of the inverter are determined through a time sequence prediction model, and the time sequence prediction output voltage and the time sequence prediction output current are used for conducting maximum power tracking on the inverter. The system can be stabilized near a real maximum power point.

Description

Technical Field

[0001] The present application relates to the technical field of photovoltaic inverter control, and more specifically, to a photovoltaic water pump inverter control method and system based on deep learning. Background Art

[0002] The optimal operating point of a solar panel is called the maximum power point, which mainly depends on the operating temperature of the panel and the light level at the time. The maximum power point of a solar panel is different under different temperatures and light intensities. To make the solar panel work at the maximum power point as much as possible, photovoltaic maximum power point tracking (MPPT) is needed. The most important thing about MPPT technology is to find a suitable MPPT control algorithm that can effectively track the maximum power point under rapidly changing weather conditions and control the panel to work at the maximum power point as much as possible.

[0003] Patent CN106936362B (application number: CN201710121665.1) provides a method and system for fast MPPT of photovoltaic water pump inverter. By presetting a voltage close to the maximum power point, PID adjusts the output frequency of the inverter within a motor acceleration time, increases the upper limit of the output frequency, and gradually adjusts the maximum power point voltage according to the increase or decrease of the output frequency, so that the solar panel can work as close to the maximum power point as possible. In the method of patent CN106936362B, the dynamic response characteristics of PID regulation and the mismatch between the fixed step adjustment strategy under rapidly changing lighting conditions lead to the risk of algorithm convergence. For example, when intermittent shadows are caused by cloud movement, the power-voltage (PV) curve of the solar panel will fluctuate violently and may show multi-peak characteristics. At this time, the frequency adjustment of the PID controller based on historical error data (integral term) and change trend (differential term) will produce a hysteresis response, and the fixed step voltage adjustment mechanism is superimposed with the discretized disturbance. The phase difference between the two causes the system to produce a positive feedback effect when approaching the maximum power point - the inertial regulation of PID will continue to cross the actual maximum power point, and the fixed step size will aggravate the voltage span, especially in the steep area of ​​the PV curve. This dual regulation synergistic imbalance will cause continuous overshoot oscillation, making the system unable to stabilize near the true maximum power point, resulting in poor control of the photovoltaic water pump inverter. Summary of the invention

[0004] The purpose of this application is to provide a photovoltaic water pump inverter control method and system based on deep learning, which solves the technical problem that the system cannot be stabilized near the true maximum power point and achieves the technical effect of stabilizing the system near the true maximum power point.

[0005] An embodiment of the present application provides a photovoltaic water pump inverter control method based on deep learning, and the method includes: obtaining the time-series light intensity and time-series battery panel temperature of the photovoltaic solar panel under working conditions, and obtaining the time-series output voltage and time-series output current of the inverter, and obtaining the preset time-series output pressure of the water flow output by the water pump; wherein, the electric energy output by the photovoltaic solar panel is converted by the inverter to supply power to the water pump; through the time-series prediction model, according to the time-series light intensity, the time-series battery panel temperature, the time-series output voltage, the time-series output current and the preset time-series output pressure, the time-series predicted output voltage and the time-series predicted output current of the inverter are determined, and the time-series predicted output voltage and the time-series predicted output current are used to perform maximum power tracking on the inverter.

[0006] In a possible implementation, the timing prediction output voltage and the timing prediction output current of the inverter are determined through a timing prediction model according to the timing light intensity, the timing battery panel temperature, the timing output voltage, the timing output current and the preset timing output pressure, including: determining the timing light intensity weight according to the timing light intensity attention unit, determining the timing battery panel temperature weight according to the timing battery panel temperature attention unit, determining the timing output voltage weight according to the timing output voltage attention unit, determining the timing output current weight according to the timing output current attention unit, and determining the preset timing output pressure weight according to the preset timing output pressure attention unit; through the timing prediction model, according to the timing light intensity, the timing light intensity weight, the timing battery panel temperature, the timing battery panel temperature weight, the timing output voltage, the timing output voltage weight, the timing output current, the timing output current weight, the preset timing output pressure and the preset timing output pressure weight, the timing prediction output voltage and the timing prediction output current of the inverter are determined.

[0007] In another possible implementation, the timing prediction output voltage and the timing prediction output current of the inverter are determined according to the timing light intensity, the timing battery panel temperature, the timing output voltage, the timing output current and the preset timing output pressure through the timing prediction model, and also include: determining the light intensity fusion feature through the light fusion processing unit according to the timing light intensity and the timing light intensity weight; determining the first fusion feature through the first fusion unit according to the light intensity fusion feature, the timing battery panel temperature and the timing battery panel temperature weight; determining the second fusion feature through the second fusion unit according to the light intensity fusion feature, the timing output voltage and the timing output voltage weight; determining the second fusion feature through the third fusion unit according to the light intensity fusion feature, the timing output current and the timing output current weight; determining the third fusion feature through the third fusion unit according to the light intensity fusion feature, the preset timing output pressure and the preset timing output pressure weight; determining the timing prediction output voltage and the timing prediction output current of the inverter through the timing prediction model according to the light intensity fusion feature, the first fusion feature, the second fusion feature and the third fusion feature.

[0008] In another possible implementation, the method further includes: determining weather classification information according to the time-series light intensity through a weather classification model, and obtaining a first time window corresponding to the weather classification information, and obtaining the time-series light intensity within the first time window; and obtaining the time-series battery panel temperature, time-series output voltage, time-series output current and preset time-series output pressure within a second time window; wherein the length of the second time window is greater than the length of the first time window, and the first time window and the second time window are both time windows that are traced back from the current moment; determining the light intensity fusion feature through a light fusion processing unit according to the time-series light intensity and the time-series light intensity weight within the first time window; and determining the light intensity fusion feature through the first fusion unit according to the light intensity fusion feature and the time-series light intensity weight within the second time window. The first fusion feature is determined by the timing battery panel temperature and the timing battery panel temperature weight; the second fusion feature is determined by the second fusion unit according to the light intensity fusion feature, the timing output voltage in the second time window, and the timing output voltage weight; the second fusion feature is determined by the third fusion unit according to the light intensity fusion feature, the timing output current in the second time window and the timing output current weight; the third fusion feature is determined by the third fusion unit according to the light intensity fusion feature, the preset timing output pressure in the second time window and the preset timing output pressure weight; the timing prediction output voltage and the timing prediction output current of the inverter are determined by the timing prediction model according to the light intensity fusion feature, the first fusion feature, the second fusion feature and the third fusion feature.

[0009] In another possible implementation, the method also includes: respectively obtaining weather status information within the first time window and the third time window, and determining, through a weather status identification unit, weighted summation weights corresponding to the first time window and the third time window respectively according to the weather status information within the first time window and the third time window; wherein the third time window is the time window before the first time window, and the duration of the third time window is 3 to 5 times the duration of the first time window; and according to the weighted summation weights corresponding to the first time window and the third time window, summing the temporal light intensity of the first time window and the third time window as the temporal light intensity within the first time window.

[0010] In another possible implementation, the method also includes: respectively obtaining the time-series light intensity of multiple associated photovoltaic solar panels of the target photovoltaic solar panel within the first time window; respectively determining multiple first light intensity variance values ​​of the multiple first time-series light intensities of the target photovoltaic solar panel and the multiple associated photovoltaic solar panels, and determining the minimum first light intensity variance value among the multiple first light intensity variance values; wherein the multiple associated photovoltaic solar panels are multiple photovoltaic solar panels that are located close to the target photovoltaic solar panel; obtaining the first time window of the photovoltaic solar panel corresponding to the minimum first light intensity variance value as the target first time window; determining the target light intensity fusion feature through the light fusion processing unit according to the time-series light intensity and the time-series light intensity weight within the target first time window, and determining the unified time-series predicted output voltage and time-series predicted output current of the inverter of the target photovoltaic solar panel and the multiple associated photovoltaic solar panels according to the target light intensity fusion feature.

[0011] In another possible implementation, the method further includes: respectively obtaining the time-series illumination intensity of multiple associated photovoltaic solar panels of the target photovoltaic solar panel within the third time window; respectively determining multiple third illumination intensity variance values ​​of multiple third time-series illumination intensities of the target photovoltaic solar panel and multiple associated photovoltaic solar panels, and determining the minimum third illumination intensity variance value among the multiple third illumination intensity variance values; obtaining the third time window of the photovoltaic solar panel corresponding to the minimum third illumination intensity variance value as the target third time window; determining the weighted summation weight corresponding to the target first time window and the target third time window according to the weather state information within the target first time window and the target third time window through the weather state recognition unit; summing the time-series illumination intensities of the target first time window and the target third time window according to the weighted summation weights corresponding to the target first time window and the target third time window as the time-series illumination intensity within the target first time window, and determining the unified time-series predicted output voltage and time-series predicted output current of the inverter of the target photovoltaic solar panel and multiple associated photovoltaic solar panels according to the time-series illumination intensity within the target first time window.

[0012] In another possible implementation, the method also includes: respectively acquiring weather classification information of multiple associated photovoltaic solar panels of the target photovoltaic solar panel, and when the weather classification information of the target photovoltaic solar panel and the multiple associated photovoltaic solar panels is the same, determining the time-averaged light intensity within a first time window, the time-averaged battery panel temperature within a second time window, the time-averaged output voltage, the time-averaged output current and the time-averaged output pressure of the target photovoltaic solar panel and the multiple associated photovoltaic solar panels; and determining a unified time-predicted output voltage and time-predicted output current of the inverter of the target photovoltaic solar panel and the multiple associated photovoltaic solar panels based on the time-averaged light intensity within the first time window, the time-averaged battery panel temperature within the second time window, the time-averaged output voltage, the time-averaged output current and the time-averaged output pressure of the target photovoltaic solar panel and the multiple associated photovoltaic solar panels.

[0013] In another possible implementation, the method also includes: obtaining multiple weighted summation weights corresponding to the target photovoltaic solar panel and multiple associated photovoltaic solar panels in the first time window and the third time window, respectively, and determining the target weighted summation weight with the largest value in the first time window among the multiple weighted summation weights; according to the target weighted summation weights corresponding to the first time window and the third time window, summing and averaging the time-series light intensities of the target photovoltaic solar panel and multiple associated photovoltaic solar panels in the first time window and the third time window as the time-series average light intensity of the target photovoltaic solar panel and multiple associated photovoltaic solar panels in the first time window.

[0014] An embodiment of the present application also provides a photovoltaic water pump inverter control system based on deep learning, comprising a unit for executing any of the methods described above.

[0015] Compared with the prior art, the embodiments of the present invention have the following beneficial effects: The embodiment of the present application provides a photovoltaic water pump inverter control method based on deep learning, and the method includes: obtaining the timed light intensity and timed battery panel temperature of the photovoltaic solar panel under working conditions, and obtaining the timed output voltage and timed output current of the inverter, and obtaining a preset timed output pressure of the water flow output by the water pump; wherein, the electric energy output by the photovoltaic solar panel is converted by the inverter to supply power to the water pump; through the timed prediction model, according to the timed light intensity, the timed battery panel temperature, the timed output voltage, the timed output current and the preset timed output pressure, the timed predicted output voltage and the timed predicted output current of the inverter are determined, and the timed predicted output voltage and the timed predicted output current are used to perform maximum power tracking on the inverter. The embodiments of the present application can control the timing predicted output voltage and timing predicted output current of the inverter in combination with the timing characteristics of the timing light intensity, timing battery panel temperature, timing output voltage, timing output current and preset timing output pressure, and can directly combine the timing characteristics of the photovoltaic solar panels, water pumps and inverters to comprehensively control the working state of the inverter, thereby improving the control effect of the inverter. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] In order to more clearly illustrate the technical solutions in the embodiments of the present application, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative labor.

[0017] Figure 1 A schematic flow chart of a first photovoltaic water pump inverter control method based on deep learning provided in an embodiment of the present application; Figure 2 A schematic diagram of the workflow of the first photovoltaic water pump inverter control method based on deep learning provided in an embodiment of the present application; Figure 3 A schematic flow chart of a second photovoltaic water pump inverter control method based on deep learning provided in an embodiment of the present application; Figure 4 A schematic flow chart of a third photovoltaic water pump inverter control method based on deep learning provided in an embodiment of the present application; Figure 5 A schematic flow chart of a fourth photovoltaic water pump inverter control method based on deep learning provided in an embodiment of the present application; Figure 6 A schematic flow chart of a fifth photovoltaic water pump inverter control method based on deep learning provided in an embodiment of the present application; Figure 7 A schematic flow chart of a sixth photovoltaic water pump inverter control method based on deep learning provided in an embodiment of the present application; Figure 8 A flowchart of a seventh photovoltaic water pump inverter control method based on deep learning provided in an embodiment of the present application; Fig. 9 A schematic diagram of the logical structure of a photovoltaic water pump inverter control system based on deep learning provided in an embodiment of the present application. DETAILED DESCRIPTION

[0018] It should be understood that when used in the present specification and the appended claims, the term "comprising" indicates the presence of described features, wholes, steps, operations, elements and / or components, but does not exclude the presence or addition of one or more other features, wholes, steps, operations, elements, components and / or combinations thereof.

[0019] It should also be understood that the term “and / or” used in the specification and appended claims refers to any and all possible combinations of one or more of the associated listed items, and includes these combinations.

[0020] As used in the specification and appended claims of this application, the term "if" can be interpreted as "when" or "uponce" or "in response to determining" or "in response to detecting", depending on the context. Similarly, the phrase "if it is determined" or "if [described condition or event] is detected" can be interpreted as meaning "uponce it is determined" or "in response to determining" or "uponce [described condition or event] is detected" or "in response to detecting [described condition or event]", depending on the context.

[0021] In addition, in the description of the present application specification and the appended claims, the terms "first", "second", "third", etc. are only used to distinguish the descriptions and cannot be understood as indicating or implying relative importance.

[0022] References to "one embodiment" or "some embodiments" etc. described in the specification of this application mean that one or more embodiments of the present application include specific features, structures or characteristics described in conjunction with the embodiment. Therefore, the statements "in one embodiment", "in some embodiments", "in some other embodiments", "in some other embodiments", etc. that appear in different places in this specification do not necessarily refer to the same embodiment, but mean "one or more but not all embodiments", unless otherwise specifically emphasized in other ways. The terms "including", "comprising", "having" and their variations all mean "including but not limited to", unless otherwise specifically emphasized in other ways.

[0023] In the existing fast MPPT methods for photovoltaic water pump inverters, the frequency regulation based on the PID controller based on historical error data (integral term) and change trend (differential term) will produce a lagging response, while the fixed-step voltage adjustment mechanism superimposes the discrete disturbance. The phase difference between the two causes the system to produce a positive feedback effect when approaching the maximum power point - the inertial regulation of the PID will continue to exceed the actual maximum power point, and the fixed step size will aggravate the voltage span, especially in the steep area of ​​the PV curve. This dual-regulation synergistic imbalance will cause continuous overshoot oscillation, making the system unable to stabilize near the true maximum power point, resulting in poor control of the photovoltaic water pump inverter.

[0024] Based on the above reasons, an embodiment of the present application provides a photovoltaic water pump inverter control method based on deep learning, the method comprising: obtaining the timed light intensity and timed battery panel temperature of the photovoltaic solar panel under working conditions, and obtaining the timed output voltage and timed output current of the inverter, and obtaining a preset timed output pressure of the water flow output by the water pump; wherein, the electric energy output by the photovoltaic solar panel is converted by the inverter to supply power to the water pump; through the timed prediction model, according to the timed light intensity, the timed battery panel temperature, the timed output voltage, the timed output current and the preset timed output pressure, the timed predicted output voltage and the timed predicted output current of the inverter are determined, and the timed predicted output voltage and the timed predicted output current are used to perform maximum power tracking on the inverter. The embodiments of the present application can control the timing predicted output voltage and timing predicted output current of the inverter in combination with the timing characteristics of the timing light intensity, timing battery panel temperature, timing output voltage, timing output current and preset timing output pressure, and can directly combine the timing characteristics of the photovoltaic solar panels, water pumps and inverters to comprehensively control the working state of the inverter, thereby improving the control effect of the inverter.

[0025] In some scenarios, a photovoltaic water pump inverter control method based on deep learning in an embodiment of the present application can be applied to a control system coupled with photovoltaic solar panels, water pumps, and inverters, and can accurately control the working state of the inverter, thereby improving the accuracy of the inverter MPPT.

[0026] The following is a detailed description of a photovoltaic water pump inverter control method based on deep learning provided in an embodiment of the present application with reference to specific examples.

[0027] Figure 1 A schematic diagram of the flow chart of the first photovoltaic water pump inverter control method based on deep learning provided in the embodiment of the present application is as follows: Figure 1 As shown, the method includes S110 to S120, and S110 to S120 are described in detail below.

[0028] S110, obtaining the timed light intensity and timed panel temperature of the photovoltaic solar panel under working conditions, obtaining the timed output voltage and timed output current of the inverter, and obtaining the preset timed output pressure of the water flow output by the water pump. The electric energy output by the photovoltaic solar panel is converted by the inverter to supply power to the water pump.

[0029] Figure 2 A schematic diagram of the workflow of the first photovoltaic water pump inverter control method based on deep learning provided in the embodiment of the present application is shown in FIG. Figure 2 As shown, in the embodiment of the present application, the sequential light intensity and sequential battery panel temperature of the photovoltaic solar panel in the working conditions can be obtained, the sequential light intensity is a plurality of light intensities of the photovoltaic solar panel in a time sequence, and the sequential battery panel temperature is a plurality of battery panel temperatures of the photovoltaic solar panel in a time sequence. The sequential light intensity and sequential battery panel temperature can be used to control the working state of the photovoltaic inverter in combination with the working characteristics of the photovoltaic solar panel in a time sequence.

[0030] In an embodiment of the present application, the sequential output voltage and sequential output current of the inverter can also be obtained. The sequential output voltage is a plurality of output voltages of the inverter in a time sequence, and the sequential output current is a plurality of output currents of the inverter in a time sequence. The sequential output voltage and sequential output current of the inverter can be combined with the working characteristics of the inverter in a time sequence to control the working state of the photovoltaic inverter.

[0031] In an embodiment of the present application, a preset time sequence output pressure of the water flow output by the water pump can also be obtained. The preset time sequence output pressure is the pressure characteristic of the water flow output by the water pump in a time sequence. The preset time sequence output pressure is the preset time sequence output pressure for controlling the working state of the water pump. The preset time sequence output pressure is preset. The working state of the photovoltaic inverter can be controlled by combining the preset time sequence output pressure of the water pump with the working characteristics of the water pump in a time sequence.

[0032] When working, the preset timing output pressure of the water pump corresponds to the water flow pressure state pumped by the output voltage of the frequency converter. Compared with controlling the working state of the water pump only by the output voltage of the frequency converter, the present application can accurately feedback control the working state of the water pump in combination with the pressure of the water flow.

[0033] For example, the preset time-series output pressure of the water pump may be preset according to irrigation demand and water demand, and the preset time-series output pressure may be derived from a water pressure control curve of the water pump during operation.

[0034] It should be noted that the electric energy output by the photovoltaic solar panels is converted by the inverter to power the water pump. The photovoltaic solar panels are used to generate electric energy. The inverter is used to convert the electric energy generated by the photovoltaic solar panels into current and voltage suitable for the operation of the water pump. The water pump is used to pump and drive the water flow.

[0035] S120. Determine the inverter's predicted output voltage and predicted output current through a timing prediction model according to the timing light intensity, the timing battery panel temperature, the timing output voltage, the timing output current and the preset timing output pressure. The predicted output voltage and the predicted output current are used to perform maximum power tracking on the inverter.

[0036] like Figure 2 As shown, after obtaining the time-series light intensity, the time-series battery panel temperature, the time-series output voltage, the time-series output current and the preset time-series output pressure, the time-series prediction model can be used to determine the time-series prediction output voltage and the time-series prediction output current of the inverter according to the time-series light intensity, the time-series battery panel temperature, the time-series output voltage, the time-series output current and the preset time-series output pressure, thereby realizing the control of the time-series prediction output voltage and the time-series prediction output current of the inverter, and the time-series prediction output voltage and the time-series prediction output current are used to perform maximum power tracking on the inverter of the photovoltaic water pump.

[0037] Exemplarily, the timing prediction model is obtained by training the marked timing light intensity, timing battery panel temperature, timing output voltage, timing output current, preset timing output pressure, timing predicted output voltage and timing predicted output current of the inverter, and the marked timing predicted output voltage and timing predicted output current of the inverter are inverter control parameters corresponding to the marked timing light intensity, timing battery panel temperature, timing output voltage, timing output current and preset timing output pressure, wherein the preset timing output pressure is the preset timing output pressure preset when controlling the water pump.

[0038] It should be noted that the marked inverter's timing prediction output voltage and timing prediction output current are the maximum power tracking points of the photovoltaic water pump inverter. The photovoltaic water pump inverter can then be tracked for maximum power through the timing prediction model, thereby improving the effect of maximum power tracking of the photovoltaic water pump inverter.

[0039] The existing PID-based inertial regulation method will continuously exceed the actual maximum power point, especially in the steep area of ​​the PV curve. This dual regulation coordination disorder will cause continuous overshoot oscillation, making the system unable to stabilize near the true maximum power point, resulting in poor control of the PV water pump inverter.

[0040] In the embodiment of the present application, when the inverter of the photovoltaic water pump drives the water pump, as the speed of the water pump increases, the load becomes larger, and the output current becomes larger. It can be seen from the characteristic curve of the solar panel that when the current increases to a certain level, the voltage will drop rapidly, or in sudden situations such as clouds blocking the sun, the output voltage of the solar panel is too low. If it is not adjusted in time, it will cause the inverter to power off and restart. In the embodiment of the present application, the output voltage generated by the inverter of the solar panel is controlled, which can avoid the inverter power off and restart caused by poor control of the inverter's timing predicted output voltage and timing predicted output current.

[0041] The beneficial effect of the above implementation method is that, through the timing prediction model, the timing predicted output voltage and timing predicted output current of the inverter are determined according to the timing light intensity, the timing battery panel temperature, the timing output voltage, the timing output current and the preset timing output pressure. The timing prediction model can perform high-power tracking of the photovoltaic water pump inverter, improve the effect of maximum power tracking of the photovoltaic water pump inverter, and avoid the continuous overshoot oscillation caused by the inertial regulation method based on PID.

[0042] In some implementations, in the above S120, the timing prediction model is used to determine the timing predicted output voltage and timing predicted output current of the inverter according to the timing light intensity, the timing battery panel temperature, the timing output voltage, the timing output current and the preset timing output pressure, and also includes S121 to 122. S121 to 122 are described in detail below.

[0043] S121. Determine the timing light intensity weight according to the timing light intensity attention unit, determine the timing battery panel temperature weight according to the timing battery panel temperature attention unit, determine the timing output voltage weight according to the timing output voltage attention unit, determine the timing output current weight according to the timing output current attention unit, and determine the preset timing output pressure weight according to the preset timing output pressure attention unit.

[0044] In the embodiment of the present application, in order to further improve the effect of maximum power tracking of the photovoltaic water pump inverter, the time-series light intensity weight can be determined according to the time-series light intensity attention unit, and the time-series light intensity weight is used to determine the weight of the light intensity at different times in the time-series light intensity. The timing battery panel temperature weight is determined according to the timing battery panel temperature attention unit, and the timing battery panel temperature weight is used to determine the weight of the battery panel temperature at different times in the timing battery panel temperature; the timing output voltage weight is determined according to the timing output voltage attention unit, and the timing output voltage weight is used to determine the weight of the output voltage at different times in the timing output voltage; the timing output current weight is determined according to the timing output current attention unit, and the timing output current weight is used to determine the weight of the output current at different times in the timing output current; the preset timing output pressure weight is determined according to the preset timing output pressure attention unit, and the preset timing output pressure weight is used to determine the weight of the preset output pressure at different times in the preset timing output pressure, so as to further improve the accuracy of determining the timing predicted output voltage and the timing predicted output current.

[0045] Exemplarily, it is possible to normalize the time series light intensity, panel temperature, output voltage, output current and output pressure data to ensure that the data is at the same scale, and further divide the data into sequences of fixed length to adapt to the processing of the attention mechanism, and design multiple independent attention units for each time series data (including time series light intensity, time series panel temperature, time series output voltage, time series output current and preset time series output pressure), and determine the weight corresponding to each time series data through multiple independent attention units.

[0046] S122. Determine the inverter's predicted output voltage and predicted output current through a timing prediction model according to the timing light intensity, the timing light intensity weight, the timing battery panel temperature, the timing battery panel temperature weight, the timing output voltage, the timing output voltage weight, the timing output current, the timing output current weight, the preset timing output pressure, and the preset timing output pressure weight.

[0047] When controlling the timing predicted output voltage and the timing predicted output current, the timing prediction model can be used to determine the timing predicted output voltage and the timing predicted output current of the inverter according to the timing light intensity, the timing light intensity weight, the timing battery panel temperature, the timing battery panel temperature weight, the timing output voltage, the timing output voltage weight, the timing output current, the timing output current weight, the preset timing output pressure and the preset timing output pressure weight, so as to calculate the timing predicted output voltage and the timing predicted output current according to the respective weights of the timing light intensity, the timing battery panel temperature, the timing output voltage, the timing output current and the preset timing output pressure.

[0048] The beneficial effect of the above-mentioned implementation method is that the timing light intensity weight, the timing battery panel temperature weight, the timing output voltage weight, the timing output current weight and the preset timing output pressure weight are determined through the attention unit, and the timing predicted output voltage and the timing predicted output current of the inverter can be determined according to the timing light intensity, the timing light intensity weight, the timing battery panel temperature, the timing battery panel temperature weight, the timing output voltage, the timing output voltage weight, the timing output current, the timing output current weight, the preset timing output pressure and the preset timing output pressure weight. When determining the timing predicted output voltage and the timing predicted output current of the inverter, the accuracy of the timing predicted output voltage and the timing predicted output current of the inverter is improved according to the importance of different parameters, thereby improving the effect of controlling the timing predicted output voltage and the timing predicted output current of the inverter.

[0049] In some implementations, in the above S120, the timing prediction model is used to determine the timing predicted output voltage and timing predicted output current of the inverter according to the timing light intensity, the timing battery panel temperature, the timing output voltage, the timing output current and the preset timing output pressure, and also includes S123 to S124. S123 to S124 are described in detail below.

[0050] S123. Determine the illumination intensity fusion feature through the illumination fusion processing unit according to the sequential illumination intensity and the sequential illumination intensity weight. Determine the first fusion feature through the first fusion unit according to the illumination intensity fusion feature, the sequential battery panel temperature and the sequential battery panel temperature weight. Determine the second fusion feature through the second fusion unit according to the illumination intensity fusion feature, the sequential output voltage and the sequential output voltage weight. Determine the third fusion feature through the third fusion unit according to the illumination intensity fusion feature, the sequential output current and the sequential output current weight. Determine the fourth fusion feature through the fourth fusion unit according to the illumination intensity fusion feature, the preset sequential output pressure and the preset sequential output pressure weight.

[0051] When controlling the inverter, since light intensity is of greater importance, in order to improve the control effect of the inverter, in order to further optimize the feature extraction of light intensity, the embodiment of the present application determines the light intensity fusion feature through the light fusion processing unit according to the temporal light intensity and the temporal light intensity weight. The light intensity fusion feature fuses the temporal light intensity and the temporal light intensity weight information.

[0052] After obtaining the light intensity fusion feature, the first fusion feature can be determined through the first fusion unit according to the light intensity fusion feature, the sequential battery panel temperature and the sequential battery panel temperature weight. The first fusion feature includes information about the sequential battery panel temperature and the sequential battery panel temperature weight, thereby realizing a shallow fusion of the light intensity fusion feature, the sequential battery panel temperature and the sequential battery panel temperature weight.

[0053] Exemplarily, the first fusion unit may be a LSTM-based deep learning unit, through which the light intensity fusion feature, the sequential solar panel temperature and the sequential solar panel temperature weight can be fused.

[0054] After obtaining the light intensity fusion feature, the second fusion feature can be determined through the second fusion unit according to the light intensity fusion feature, the timing output voltage, and the timing output voltage weight. The second fusion feature contains information about the timing output voltage and the timing output voltage weight, thereby realizing a shallow fusion of the light intensity fusion feature, the timing output voltage, and the timing output voltage weight.

[0055] After obtaining the light intensity fusion feature, the third fusion feature can be determined through the third fusion unit according to the light intensity fusion feature, the timing output current and the timing output current weight. The third fusion feature contains information about the timing output current and the timing output current weight, thereby realizing a shallow fusion of the light intensity fusion feature, the timing output current and the timing output current weight.

[0056] After obtaining the light intensity fusion feature, the fourth fusion feature can be determined through the fourth fusion unit according to the light intensity fusion feature, the preset timing output pressure and the preset timing output pressure weight. The fourth fusion unit contains information of the preset timing output pressure and the preset timing output pressure weight, thereby realizing a shallow fusion of the light intensity fusion feature, the preset timing output pressure and the preset timing output pressure weight.

[0057] S124. Determine the timing predicted output voltage and the timing predicted output current of the inverter through the timing prediction model according to the light intensity fusion feature, the first fusion feature, the second fusion feature, the third fusion feature and the fourth fusion feature.

[0058] After obtaining the light intensity fusion feature, the first fusion feature, the second fusion feature, the third fusion feature and the fourth fusion feature, the timing prediction output voltage and the timing prediction output current of the inverter can be further determined through the timing prediction model according to the light intensity fusion feature, the first fusion feature, the second fusion feature, the third fusion feature and the fourth fusion feature. At this time, the light intensity fusion feature, the first fusion feature, the second fusion feature, the third fusion feature and the fourth fusion feature are fused again to generate control information for the timing prediction output voltage and the timing prediction output current, realizing the deep fusion of the light intensity fusion feature and the first fusion feature, the second fusion feature, the third fusion feature and the fourth fusion feature, realizing the shallow and deep secondary feature fusion of the light intensity fusion feature, improving the utilization effect of the light intensity information corresponding to the light intensity fusion feature, and improving the control effect of the inverter according to the key parameter light intensity.

[0059] The beneficial effects of the above-mentioned implementation method are as follows: shallow fusion of the light intensity fusion feature, the sequential battery panel temperature and the sequential battery panel temperature weight is achieved; shallow fusion of the light intensity fusion feature, the sequential output voltage and the sequential output voltage weight is achieved; shallow fusion of the light intensity fusion feature, the sequential output current and the sequential output current weight is achieved; shallow fusion of the light intensity fusion feature, the preset sequential output pressure and the preset sequential output pressure weight is achieved; deep fusion of the light intensity fusion feature and the first fusion feature, the second fusion feature, the third fusion feature and the fourth fusion feature is achieved; shallow and deep secondary feature fusion of the light intensity fusion feature is achieved, the utilization effect of the light intensity information corresponding to the light intensity fusion feature is improved, and the control effect of the inverter according to the key parameter light intensity is improved.

[0060] Figure 3 A schematic diagram of a second photovoltaic water pump inverter control method based on deep learning provided in an embodiment of the present application is shown in FIG. Figure 3 As shown, the above method also includes S210 to S220, and S210 to S230 are further described below.

[0061] S210. Determine weather classification information according to the time-series light intensity through the weather classification model, obtain a first time window corresponding to the weather classification information, and obtain the time-series light intensity within the first time window. Also obtain the time-series battery panel temperature, time-series output voltage, time-series output current, and preset time-series output pressure within the second time window. The length of the second time window is greater than the length of the first time window, and both the first time window and the second time window are time windows that are traced back from the current moment.

[0062] Because the changing characteristics of light intensity under different weather classifications are different, for example, under the weather classification of sunny days, the change of light intensity over time is not obvious. For example, under the weather classification of cloudy days, the change of light intensity over time will change dramatically with the change of cloud cover.

[0063] In order to further improve the utilization effect of light intensity information, the weather classification information is determined according to the time-series light intensity through the weather classification model. The weather classification information characterizes the change of light intensity, and then the first time window corresponding to the weather classification information can be obtained, and the time-series light intensity within the first time window can be obtained, wherein different first time windows have different time lengths.

[0064] Exemplarily, when the weather classification information is cloudy, the first time window is a time window within 1 hour starting from the current moment.

[0065] Exemplarily, when the weather classification information is sunny, the first time window is a time window within 5 hours starting from the current moment.

[0066] At the same time, in order to ensure the control effect of the inverter, the timing battery panel temperature, timing output voltage, timing output current and preset timing output pressure within the second time window can also be obtained, wherein the length of the second time window is greater than the length of the first time window. Compared with the first time window, the purpose of improving the control effect of the inverter is achieved according to the timing battery panel temperature, timing output voltage, timing output current and preset timing output pressure within the second time window of a longer time period.

[0067] It should be noted that both the first time window and the second time window are time windows that trace back from the current moment to ensure the control accuracy of the inverter.

[0068] S220. Determine the illumination intensity fusion feature through the illumination fusion processing unit according to the sequential illumination intensity and the sequential illumination intensity weight in the first time window. Determine the first fusion feature through the first fusion unit according to the illumination intensity fusion feature, the sequential battery panel temperature and the sequential battery panel temperature weight in the second time window. Determine the second fusion feature through the second fusion unit according to the illumination intensity fusion feature, the sequential output voltage and the sequential output voltage weight in the second time window. Determine the third fusion feature through the third fusion unit according to the illumination intensity fusion feature, the sequential output current and the sequential output current weight in the second time window. Determine the fourth fusion feature through the fourth fusion unit according to the illumination intensity fusion feature, the preset sequential output pressure and the preset sequential output pressure weight in the second time window.

[0069] After obtaining the temporal lighting intensity in the first time window, the lighting intensity fusion feature can be determined by the lighting fusion processing unit according to the temporal lighting intensity and the temporal lighting intensity weight in the first time window, wherein the temporal lighting intensity weight can be determined by step S121.

[0070] After obtaining the light intensity fusion feature, the first fusion feature can be determined by a method similar to that in 123, through the first fusion unit, according to the light intensity fusion feature, the sequential battery panel temperature in the second time window, and the sequential battery panel temperature weight. The second fusion feature is determined by the second fusion unit according to the light intensity fusion feature, the sequential output voltage in the second time window, and the sequential output voltage weight. The third fusion feature is determined by the third fusion unit according to the light intensity fusion feature, the sequential output current in the second time window, and the sequential output current weight. The fourth fusion feature is determined by the fourth fusion unit according to the light intensity fusion feature, the preset sequential output pressure in the second time window, and the preset sequential output pressure weight. Among them, the first fusion feature, the second fusion feature, the third fusion feature, and the fourth fusion feature all combine the control feature in the second time window and the light intensity feature in the first time window, so as to achieve the purpose of adjusting the light intensity feature according to the weather conditions, and further improve the control effect of the inverter under different weather conditions.

[0071] S230. Determine the time-series predicted output voltage and the time-series predicted output current of the inverter through a time-series prediction model according to the light intensity fusion feature, the first fusion feature, the second fusion feature, and the third fusion feature.

[0072] Furthermore, the timing prediction model can be used to determine the timing predicted output voltage and timing predicted output current of the inverter according to the light intensity fusion feature, the first fusion feature, the second fusion feature, the third fusion feature and the fourth fusion feature, so as to control the inverter according to the timing predicted output voltage and the timing predicted output current of the inverter.

[0073] The beneficial effect of the above-mentioned implementation method is that the light intensity characteristics are adjusted according to the weather conditions, and the first fusion feature, the second fusion feature, the third fusion feature and the fourth fusion feature all combine the control characteristics within the second time window and the light intensity characteristics within the first time window, thereby further improving the control effect of the inverter under different weather conditions.

[0074] Figure 4 A schematic diagram of a flow chart of a third photovoltaic water pump inverter control method based on deep learning provided in an embodiment of the present application, such as Figure 4 As shown, the above method also includes S240 to S250, and S240 to S250 are described in detail below.

[0075] S240, respectively obtaining weather status information in the first time window and the third time window, and determining weighted sum weights corresponding to the first time window and the third time window respectively according to the weather status information in the first time window and the third time window through a weather status identification unit. Wherein, the third time window is a time window before the first time window, and the duration of the third time window is 3 to 5 times the duration of the first time window.

[0076] When the light intensity information is processed by the methods in S210 to S230, the first time window corresponding to the sequential light intensity is obtained by simply truncating the time according to the weather conditions. This processing method may result in insufficient consideration of the sequential light intensity before the first time window, which in turn leads to poor control of the inverter only by the light intensity within the first time window.

[0077] In order to improve the utilization effect of the information on the time series light intensity before the first time window, the weather status information in the first time window and the third time window can be obtained respectively. Through the weather status identification unit, the weighted sum weights corresponding to the first time window and the third time window are determined according to the weather status information in the first time window and the third time window, respectively, so as to realize the fusion of the light intensity in the first time window and the third time window before the first time window, thereby improving the utilization effect of the information on the time series light intensity before the first time window.

[0078] Exemplarily, the weather status information may be local weather status information obtained through weather forecast.

[0079] It should be noted that the third time window is the time window before the first time window, and the length of the third time window is 3 to 5 times the length of the first time window to ensure the utilization effect of the information of the temporal light intensity before the first time window.

[0080] Exemplarily, the weather status recognition unit may be a LSTM-based neural network, which can determine the weighted sum weight at each moment corresponding to the first time window and the third time window respectively according to the weather status information in the first time window and the third time window.

[0081] S250. Sum the time series illumination intensities of the first time window and the third time window according to the weighted sum weights corresponding to the first time window and the third time window, as the time series illumination intensity within the first time window.

[0082] After obtaining the temporal light intensity of the first time window and the third time window before the first time window, the temporal light intensity of the first time window and the third time window can be summed according to the weighted summation weights corresponding to the first time window and the third time window, and the sum can be used as the temporal light intensity within the first time window, so that the temporal light intensity within the first time window represents the effect of the light intensity within the first time window and the third time window, thereby improving the effect of subsequent inverter control based on the temporal light intensity within the first time window.

[0083] Exemplarily, when the inverter is subsequently controlled according to the sequential light intensity within the first time window, the state of the inverter may be controlled according to the methods in S210 to S230 described above.

[0084] The beneficial effect of the above-mentioned implementation method is that it avoids obtaining the first time window corresponding to the sequential light intensity by only cutting off the time according to the weather conditions, and avoids insufficient consideration of the sequential light intensity before the first time window. The light intensity information of the first time window is optimized by weighted summing weights of the light intensity information in the first time window and the third time window corresponding to different weather status information, thereby improving the effect of subsequent inverter control according to the sequential light intensity in the first time window.

[0085] Figure 5 A schematic flow chart of a fourth photovoltaic water pump inverter control method based on deep learning provided in an embodiment of the present application is shown in FIG. Figure 5 As shown, the above method also includes S310 to S320, and S310 to S320 are described in detail below.

[0086] S310, respectively obtaining the time-series illumination intensities of multiple associated photovoltaic solar panels of the target photovoltaic solar panel within the first time window. Respectively determining multiple first illumination intensity variance values ​​of multiple first time-series illumination intensities of the target photovoltaic solar panel and multiple associated photovoltaic solar panels, and determining the minimum first illumination intensity variance value among the multiple first illumination intensity variance values. The multiple associated photovoltaic solar panels are multiple photovoltaic solar panels that are close to the position of the target photovoltaic solar panel.

[0087] In order to reduce the amount of calculation for maximum power tracking of multiple photovoltaic solar panels, the time-series light intensity of multiple associated photovoltaic solar panels of the target photovoltaic solar panel within the first time window can be obtained respectively, and then the time-series light intensity of the target photovoltaic solar panel and multiple associated photovoltaic solar panels within the first time window can be evaluated.

[0088] When the time-series light intensity of the target photovoltaic solar panel and multiple associated photovoltaic solar panels within the first time window is obtained, multiple first light intensity variance values ​​of the multiple first time-series light intensities of the target photovoltaic solar panel and the multiple associated photovoltaic solar panels can be determined respectively, and the first light intensity variance value represents the fluctuation of the first time-series light intensity. At the same time, the minimum first light intensity variance value among the multiple first light intensity variance values ​​can be further determined, and the photovoltaic solar panel corresponding to the minimum first light intensity variance value has the smallest light intensity fluctuation amplitude.

[0089] It should be noted that the multiple associated photovoltaic solar panels are multiple photovoltaic solar panels that are close to the position of the target photovoltaic solar panel. The multiple associated photovoltaic solar panels and the target photovoltaic solar panel have similar light intensities due to their close positions. Therefore, the inverters of the multiple associated photovoltaic solar panels and the target photovoltaic solar panels can be centrally controlled according to the light intensity to reduce the amount of calculation for controlling the inverters of the multiple associated photovoltaic solar panels and the target photovoltaic solar panels.

[0090] S320, obtaining the first time window of the photovoltaic solar panel corresponding to the minimum first illumination intensity variance value as the target first time window. Determine the target illumination intensity fusion feature through the illumination fusion processing unit according to the sequential illumination intensity and the sequential illumination intensity weight in the target first time window, and determine the unified sequential predicted output voltage and sequential predicted output current of the inverters of the target photovoltaic solar panel and multiple associated photovoltaic solar panels according to the target illumination intensity fusion feature.

[0091] After obtaining the minimum first light intensity variance value, it means that the light intensity fluctuation amplitude detected by the photovoltaic solar panel corresponding to the minimum first light intensity variance value is the smallest, and the detection value of the light intensity detected by the photovoltaic solar panel corresponding to the minimum first light intensity variance value may be the most accurate and most consistent with the actual local light intensity change. The photovoltaic solar panel corresponding to the minimum first light intensity variance value is least frequently blocked by clouds, and then the first time window of the photovoltaic solar panel corresponding to the minimum first light intensity variance value can be obtained as the target first time window, and the target photovoltaic solar panel and inverters of multiple associated photovoltaic solar panels can be controlled according to the target first time window.

[0092] Exemplarily, when obtaining the first time window of the photovoltaic solar panel corresponding to the minimum first illumination intensity variance value, the first time window of the photovoltaic solar panel corresponding to the minimum first illumination intensity variance value may be determined according to the method in step S210.

[0093] After obtaining the first time window of the photovoltaic solar panel corresponding to the minimum first light intensity variance value, the target light intensity fusion feature can be determined through the light fusion processing unit according to the sequential light intensity and the sequential light intensity weight within the target first time window, and the unified sequential predicted output voltage and sequential predicted output current of the inverters of the target photovoltaic solar panel and multiple associated photovoltaic solar panels can be determined according to the target light intensity fusion feature, and the control effect of the target photovoltaic solar panel and the inverters of multiple associated photovoltaic solar panels is improved according to the most accurate sequential light intensity of the detection result.

[0094] The beneficial effect of the above-mentioned implementation method is that the control effect of the inverter of the target photovoltaic solar panel and multiple related photovoltaic solar panels is improved according to the most accurate time-series light intensity of the detection results, and the calculation amount of the inverter control of the target photovoltaic solar panel and multiple related photovoltaic solar panels is reduced.

[0095] The beneficial effect of the above-mentioned implementation method is that the detection value of the light intensity detected by the photovoltaic solar panel corresponding to the minimum first light intensity variance value may be the most accurate and most consistent with the actual local light intensity changes. The photovoltaic solar panel corresponding to the minimum first light intensity variance value has the lowest frequency of being blocked by clouds. The target first time window of the photovoltaic solar panel corresponding to the minimum first light intensity variance value is used to control the inverter of the target photovoltaic solar panel and multiple related photovoltaic solar panels, thereby improving the stability of controlling the inverter.

[0096] Figure 6 A schematic diagram of a fifth photovoltaic water pump inverter control method based on deep learning provided in an embodiment of the present application is shown in FIG. Figure 6 As shown, the above method also includes S330 to S340, and S330 to S340 are described in detail below.

[0097] S330, respectively obtaining the time-series illumination intensities of the multiple associated photovoltaic solar panels of the target photovoltaic solar panel within the third time window, respectively determining multiple third illumination intensity variance values ​​of the multiple third time-series illumination intensities of the target photovoltaic solar panel and the multiple associated photovoltaic solar panels, and determining the minimum third illumination intensity variance value among the multiple third illumination intensity variance values.

[0098] In order to further improve the control effect of the inverters of the target photovoltaic solar panel and multiple associated photovoltaic solar panels, the time-series light intensity of multiple associated photovoltaic solar panels of the target photovoltaic solar panel within the third time window can also be obtained respectively. Specifically, the time-series light intensity within the third time window can be obtained through the above-mentioned step S240.

[0099] After obtaining the time-series light intensity within the third time window, multiple third light intensity variance values ​​of the multiple third time-series light intensities of the target photovoltaic solar panel and the multiple associated photovoltaic solar panels can be determined respectively. The multiple third light intensity variance values ​​characterize the fluctuation amplitude of the third light intensity of the target photovoltaic solar panel and the multiple associated photovoltaic solar panels. Further, the minimum third light intensity variance value among the multiple third light intensity variance values ​​can be determined. The minimum third light intensity variance value is the third light intensity with the smallest fluctuation amplitude.

[0100] S340, obtaining the third time window of the photovoltaic solar panel corresponding to the minimum third illumination intensity variance value as the target third time window. Determine the weighted sum weights corresponding to the target first time window and the target third time window through the weather state identification unit according to the weather state information in the target first time window and the target third time window. According to the weighted sum weights corresponding to the target first time window and the target third time window, sum the sequential illumination intensities of the target first time window and the target third time window as the sequential illumination intensity in the target first time window, and determine the unified sequential prediction output voltage and sequential prediction output current of the inverter of the target photovoltaic solar panel and multiple associated photovoltaic solar panels according to the sequential illumination intensity in the target first time window.

[0101] After obtaining the minimum third light intensity variance value, the third time window of the photovoltaic solar panel corresponding to the minimum third light intensity variance value can be obtained as the target third time window. The detection value of the light intensity in the target third time window may be the most accurate, most consistent with the actual local light intensity changes, and the frequency of being blocked by clouds is the lowest. Then, the weather state recognition unit can be used to determine the weighted sum weights corresponding to the target first time window and the target third time window according to the weather state information in the target first time window and the target third time window, thereby realizing the control of the target photovoltaic solar panel and the inverters of multiple associated photovoltaic solar panels according to the target first time window and the target third time window, in which the detection value of the light intensity may be the most accurate, most consistent with the actual local light intensity changes, and the frequency of being blocked by clouds is the lowest.

[0102] When controlling the inverters of a target photovoltaic solar panel and multiple associated photovoltaic solar panels, the temporal light intensities of the target first time window and the target third time window can be summed according to the weighted summation weights corresponding to the target first time window and the target third time window, and the temporal light intensity within the target first time window can be used as the temporal light intensity within the target first time window. Specifically, the temporal light intensity within the target first time window can be determined by the method in step S250.

[0103] After obtaining the time-series light intensity within the target first time window determined by the method in step S250, the unified time-series predicted output voltage and time-series predicted output current of the inverters of the target photovoltaic solar panel and multiple associated photovoltaic solar panels can be determined according to the time-series light intensity within the target first time window. Specifically, the time-series light intensity within the target first time window can be determined according to the above method, thereby realizing unified control of the inverters of the target photovoltaic solar panel and multiple associated photovoltaic solar panels, and successfully reducing the amount of calculation.

[0104] The beneficial effect of the above-mentioned implementation method is that the control of the target photovoltaic solar panel and the inverter of multiple associated photovoltaic solar panels is improved by improving the unified control of the inverter of the target photovoltaic solar panel and multiple associated photovoltaic solar panels, successfully reducing the control calculation amount, and improving the control effect of the system of the target photovoltaic solar panel and multiple associated photovoltaic solar panels according to the target first time window and the target third time window whose light intensity detection values ​​may be the most accurate, most in line with the actual local light intensity changes, and the lowest frequency of being blocked by clouds.

[0105] Figure 7 A flowchart of a sixth photovoltaic water pump inverter control method based on deep learning provided in an embodiment of the present application is shown in FIG. Figure 7 As shown, the above method further includes S410 to S420, and S410 to S420 are described in detail below.

[0106] S410. Obtain weather classification information of multiple associated photovoltaic solar panels of the target photovoltaic solar panel respectively. When the weather classification information of the target photovoltaic solar panel and the multiple associated photovoltaic solar panels are the same, determine the time-series average light intensity within the first time window, the time-series average panel temperature within the second time window, the time-series average output voltage, the time-series average output current and the time-series average output pressure of the target photovoltaic solar panel and the multiple associated photovoltaic solar panels.

[0107] When controlling a target photovoltaic solar panel and multiple associated photovoltaic solar panels, the weather classification information of the multiple associated photovoltaic solar panels of the target photovoltaic solar panel can be obtained respectively. When the weather classification information of the target photovoltaic solar panel and the multiple associated photovoltaic solar panels are the same, it means that the weather classification information corresponding to the target photovoltaic solar panel and the multiple associated photovoltaic solar panels is relatively accurate, and the inverters of the target photovoltaic solar panel and the multiple associated photovoltaic solar panels can be centrally controlled.

[0108] When the inverters of a target photovoltaic solar panel and multiple associated photovoltaic solar panels are centrally controlled, the sequential average light intensity within a first time window, the sequential average panel temperature within a second time window, the sequential average output voltage, the sequential average output current and the sequential average output pressure of the target photovoltaic solar panel and multiple associated photovoltaic solar panels can be determined, and then the target photovoltaic solar panel and multiple associated photovoltaic solar panels can be centrally controlled according to the average values ​​of the parameters of the target photovoltaic solar panel and multiple associated photovoltaic solar panels.

[0109] S420, determining a unified timing predicted output voltage and timing predicted output current of the inverters of the target photovoltaic solar panel and multiple associated photovoltaic solar panels based on the timing average light intensity within the first time window, the timing average panel temperature within the second time window, the timing average output voltage, the timing average output current and the timing average output pressure of the target photovoltaic solar panel and multiple associated photovoltaic solar panels.

[0110] When the target photovoltaic solar panel and the multiple associated photovoltaic solar panels are centrally controlled according to the mean values ​​of their parameters, a unified sequential predicted output voltage and sequential predicted output current of the inverters of the target photovoltaic solar panel and the multiple associated photovoltaic solar panels can be determined according to the sequential average light intensity of the target photovoltaic solar panel and the multiple associated photovoltaic solar panels within a first time window, the sequential average panel temperature within a second time window, the sequential average output voltage, the sequential average output current and the sequential average output pressure of the target photovoltaic solar panel and the multiple associated photovoltaic solar panels. Specifically, the inverter can be controlled according to S210 to S230 above, thereby reducing the amount of calculation for controlling the inverters of the target photovoltaic solar panel and the multiple associated photovoltaic solar panels. The beneficial effect of the above-mentioned implementation method is that when the weather classification information of the target photovoltaic solar panel and multiple associated photovoltaic solar panels is the same, it means that the weather classification information corresponding to the target photovoltaic solar panel and multiple associated photovoltaic solar panels is relatively accurate, and the target photovoltaic solar panel and multiple associated photovoltaic solar panels are centrally controlled according to the mean of the parameters of the target photovoltaic solar panel and multiple associated photovoltaic solar panels, thereby reducing the calculation amount of inverter control of the target photovoltaic solar panel and multiple associated photovoltaic solar panels, and improving the accuracy and control effect of inverter control of the target photovoltaic solar panel and multiple associated photovoltaic solar panels.

[0111] Figure 8 A flowchart of a seventh photovoltaic water pump inverter control method based on deep learning provided in an embodiment of the present application is shown in FIG. Figure 8 As shown, the above method also includes S510 to S520, and S510 to S520 are described in detail below.

[0112] S510, obtaining multiple weighted sum weights corresponding to the target photovoltaic solar panel and multiple associated photovoltaic solar panels in the first time window and the third time window respectively, and determining the largest target weighted sum weight in the first time window among the multiple weighted sum weights.

[0113] When the target photovoltaic solar panel and the inverters of multiple associated photovoltaic solar panels are centrally controlled, multiple weighted sum weights corresponding to the target photovoltaic solar panel and the multiple associated photovoltaic solar panels in the first time window and the third time window can be obtained, that is, multiple weighted sum weights corresponding to the target photovoltaic solar panel in the first time window and the third time window can be obtained, and multiple weighted sum weights corresponding to the multiple associated photovoltaic solar panels in the first time window and the third time window can be obtained. Further, the target weighted sum weight with the largest weight in the first time window among the multiple weighted sum weights can be determined. Among the target weighted sum weights with the largest weight in the first time window, the time sequence light intensity weight in the first time window corresponding to the target weighted sum weight is the largest. According to the target weighted sum weight, the time sequence light intensity in the first time window can be considered to the greatest extent. The target photovoltaic solar panel and the inverters of multiple associated photovoltaic solar panels can be centrally controlled by the time sequence light intensity in the first time window considered to the greatest extent, thereby ensuring the accuracy of the centralized control of the target photovoltaic solar panel and the inverters of multiple associated photovoltaic solar panels.

[0114] S520. According to the target weighted summation weights corresponding to the first time window and the third time window, the time-series light intensity of the target photovoltaic solar panel and multiple associated photovoltaic solar panels in the first time window and the third time window is summed and averaged as the time-series average light intensity of the target photovoltaic solar panel and multiple associated photovoltaic solar panels in the first time window.

[0115] When the target photovoltaic solar panel and the inverters of multiple associated photovoltaic solar panels are centrally controlled, the time-series light intensities of the target photovoltaic solar panel and the multiple associated photovoltaic solar panels in the first time window and the third time window are summed and averaged according to the target weighted summation weights corresponding to the first time window and the third time window, and the time-series average light intensity of the target photovoltaic solar panel and the multiple associated photovoltaic solar panels in the first time window is used as the time-series average light intensity of the target photovoltaic solar panel and the multiple associated photovoltaic solar panels in the first time window. The target photovoltaic solar panel and the inverters of multiple associated photovoltaic solar panels can be centrally controlled by the time-series light intensity in the first time window that is considered to the greatest extent, and the control accuracy of the target photovoltaic solar panel and the inverters of multiple associated photovoltaic solar panels is further improved by summing and averaging the time-series light intensity in the first time window that is considered to the greatest extent and the time-series light intensity in the third time window that are considered to the greatest extent.

[0116] The beneficial effect of the above-mentioned implementation method is that when the target photovoltaic solar panel and the inverters of multiple related photovoltaic solar panels are centrally controlled, the target photovoltaic solar panel and the inverters of multiple related photovoltaic solar panels are centrally controlled by maximally considering the sequential light intensity within the first time window, and the sequential light intensity within the first time window and the sequential light intensity within the third time window of the target photovoltaic solar panel and the multiple related photovoltaic solar panels are summed and averaged, thereby improving the control accuracy of the inverters of the target photovoltaic solar panel and the multiple related photovoltaic solar panels.

[0117] An embodiment of the present application also provides a photovoltaic water pump inverter control system based on deep learning, comprising a unit for executing any of the methods described above.

[0118] Fig. 9 A schematic diagram of the logical structure of a photovoltaic water pump inverter control system based on deep learning provided in an embodiment of the present application is shown in FIG. Fig. 9 As shown, the system 1 of this embodiment includes a processing unit 11, a storage unit 12 and a transceiver unit 13. The processing unit 11 is used to process data, the storage unit 12 is used to store data, and the transceiver unit 13 is used to send and receive data. The processing unit 11, the storage unit 12 and the transceiver unit 13 cooperate with each other to implement the above method. The beneficial effects of the embodiment of the present application have been described in the above method and will not be repeated here.

[0119] It should be noted that the information interaction, execution process, etc. between the above-mentioned devices / units are based on the same concept as the method embodiment of the present application. Their specific functions and technical effects can be found in the method embodiment part and will not be repeated here.

[0120] The technicians in the relevant field can clearly understand that for the convenience and simplicity of description, only the division of the above-mentioned functional units and modules is used as an example for illustration. In practical applications, the above-mentioned function allocation can be completed by different functional units and modules as needed, that is, the internal structure of the device can be divided into different functional units or modules to complete all or part of the functions described above. The functional units and modules in the embodiment can be integrated in a processing unit, or each unit can exist physically separately, or two or more units can be integrated in one unit. The above-mentioned integrated unit can be implemented in the form of hardware or in the form of software functional units. In addition, the specific names of the functional units and modules are only for the convenience of distinguishing each other, and are not used to limit the scope of protection of this application. The specific working process of the units and modules in the above-mentioned system can refer to the corresponding process in the aforementioned method embodiment, which will not be repeated here.

[0121] If the integrated unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the present application implements all or part of the processes in the above-mentioned embodiment method, which can be completed by instructing the relevant hardware through a computer program. The computer program can be stored in a computer-readable storage medium, and the computer program can implement the steps of the above-mentioned various method embodiments when executed by the processor. Among them, the computer program includes computer program code, and the computer program code can be in source code form, object code form, executable file or some intermediate form. The computer-readable medium can at least include: any entity or device that can carry the computer program code to the camera / terminal device, recording medium, computer memory, read-only memory (ROM, Read-Only Memory), random access memory (RAM, RandomAccess Memory), electric carrier signal, telecommunication signal and software distribution medium. For example, a USB flash drive, a mobile hard disk, a magnetic disk or an optical disk. In some jurisdictions, according to legislation and patent practice, computer-readable media cannot be electric carrier signals and telecommunication signals.

[0122] In the above embodiments, the description of each embodiment has its own emphasis. For parts that are not described or recorded in detail in a certain embodiment, reference can be made to the relevant descriptions of other embodiments.

[0123] Those of ordinary skill in the art will appreciate that the units and algorithm steps of each example described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professional and technical personnel can use different methods to implement the described functions for each specific application, but such implementation should not be considered to be beyond the scope of this application.

[0124] In the embodiments provided in the present application, it should be understood that the disclosed devices and methods can be implemented in other ways. For example, the device embodiments described above are only schematic. For example, the division of the modules or units is only a logical function division. There may be other division methods in actual implementation, such as multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some interfaces, indirect coupling or communication connection of devices or units, which can be electrical, mechanical or other forms.

[0125] The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed on multiple network units. Some or all of the units may be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0126] The embodiments described above are only used to illustrate the technical solutions of the present application, rather than to limit them. Although the present application has been described in detail with reference to the aforementioned embodiments, a person skilled in the art should understand that the technical solutions described in the aforementioned embodiments may still be modified, or some of the technical features may be replaced by equivalents. Such modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present application, and should all be included in the protection scope of the present application.

Claims

1. A photovoltaic water pump inverter control method based on deep learning, characterized in that: The method comprises: Obtain the timed light intensity and timed panel temperature of the photovoltaic solar panel under working conditions, obtain the timed output voltage and timed output current of the inverter, and obtain the preset timed output pressure of the water flow output by the water pump; wherein, the electric energy output by the photovoltaic solar panel is converted by the inverter to supply power to the water pump; Through the timing prediction model, the timing prediction output voltage and timing prediction output current of the inverter are determined according to the timing light intensity, timing battery panel temperature, timing output voltage, timing output current and preset timing output pressure. The timing prediction output voltage and timing prediction output current are used to perform maximum power tracking on the inverter.

2. The method according to claim 1, characterized in that Through the timing prediction model, according to the timing light intensity, timing solar panel temperature, timing output voltage, timing output current and preset timing output pressure, the timing prediction output voltage and timing prediction output current of the inverter are determined, including: Determine the sequential light intensity weight according to the sequential light intensity attention unit, determine the sequential battery panel temperature weight according to the sequential battery panel temperature attention unit, determine the sequential output voltage weight according to the sequential output voltage attention unit, determine the sequential output current weight according to the sequential output current attention unit, and determine the preset sequential output pressure weight according to the preset sequential output pressure attention unit; Through the timing prediction model, the timing predicted output voltage and timing predicted output current of the inverter are determined according to the timing light intensity, the timing light intensity weight, the timing panel temperature, the timing panel temperature weight, the timing output voltage, the timing output voltage weight, the timing output current, the timing output current weight, the preset timing output pressure and the preset timing output pressure weight.

3. The method according to claim 2, characterized in that Through the timing prediction model, according to the timing light intensity, the timing battery panel temperature, the timing output voltage, the timing output current and the preset timing output pressure, the timing prediction output voltage and the timing prediction output current of the inverter are determined, and also include: Through the illumination fusion processing unit, the illumination intensity fusion feature is determined according to the sequential illumination intensity and the sequential illumination intensity weight; through the first fusion unit, the first fusion feature is determined according to the illumination intensity fusion feature, the sequential battery panel temperature and the sequential battery panel temperature weight; through the second fusion unit, the second fusion feature is determined according to the illumination intensity fusion feature, the sequential output voltage and the sequential output voltage weight; through the third fusion unit, the second fusion feature is determined according to the illumination intensity fusion feature, the sequential output current and the sequential output current weight; through the third fusion unit, the third fusion feature is determined according to the illumination intensity fusion feature, the preset sequential output pressure and the preset sequential output pressure weight; Through the timing prediction model, the timing predicted output voltage and the timing predicted output current of the inverter are determined according to the light intensity fusion feature, the first fusion feature, the second fusion feature and the third fusion feature.

4. The method according to claim 3, characterized in that The method further comprises: Determine weather classification information according to the time-series light intensity through the weather classification model, obtain a first time window corresponding to the weather classification information, and obtain the time-series light intensity within the first time window; and obtain the time-series battery panel temperature, time-series output voltage, time-series output current and preset time-series output pressure within the second time window; wherein the length of the second time window is greater than the length of the first time window, and both the first time window and the second time window are time windows that are traced back from the current moment; The illumination fusion processing unit determines the illumination intensity fusion feature according to the sequential illumination intensity and the sequential illumination intensity weight in the first time window; the first fusion unit determines the first fusion feature according to the illumination intensity fusion feature, the sequential battery panel temperature and the sequential battery panel temperature weight in the second time window; the second fusion feature is determined according to the illumination intensity fusion feature, the sequential output voltage and the sequential output voltage weight in the second time window by the second fusion unit; the second fusion feature is determined according to the illumination intensity fusion feature, the sequential output current and the sequential output current weight in the second time window by the third fusion unit; the third fusion feature is determined according to the illumination intensity fusion feature, the preset sequential output pressure and the preset sequential output pressure weight in the second time window by the third fusion unit; Through the timing prediction model, the timing predicted output voltage and the timing predicted output current of the inverter are determined according to the light intensity fusion feature, the first fusion feature, the second fusion feature and the third fusion feature.

5. The method according to claim 4, characterized in that The method further comprises: Acquire weather status information in the first time window and the third time window respectively, and determine weighted sum weights corresponding to the first time window and the third time window respectively according to the weather status information in the first time window and the third time window through a weather status identification unit; wherein the third time window is a time window before the first time window, and the duration of the third time window is 3 to 5 times the duration of the first time window; According to the weighted summation weights corresponding to the first time window and the third time window, the temporal illumination intensity of the first time window and the third time window are summed up as the temporal illumination intensity within the first time window.

6. The method according to claim 5, characterized in that The method further comprises: Respectively obtain the time-series illumination intensities of multiple associated photovoltaic solar panels of the target photovoltaic solar panel within the first time window; respectively determine multiple first illumination intensity variance values ​​of multiple first time-series illumination intensities of the target photovoltaic solar panel and the multiple associated photovoltaic solar panels, and determine the minimum first illumination intensity variance value among the multiple first illumination intensity variance values; wherein the multiple associated photovoltaic solar panels are multiple photovoltaic solar panels that are located close to the position of the target photovoltaic solar panel; The first time window of the photovoltaic solar panel corresponding to the minimum first light intensity variance value is obtained as the target first time window; the target light intensity fusion feature is determined through the light fusion processing unit according to the temporal light intensity and the temporal light intensity weight in the target first time window, and the unified temporal predicted output voltage and temporal predicted output current of the inverters of the target photovoltaic solar panel and multiple associated photovoltaic solar panels are determined according to the target light intensity fusion feature.

7. The method according to claim 6, characterized in that The method further comprises: Respectively obtain the time-series illumination intensities of multiple associated photovoltaic solar panels of the target photovoltaic solar panel within the third time window; respectively determine multiple third illumination intensity variance values ​​of multiple third time-series illumination intensities of the target photovoltaic solar panel and the multiple associated photovoltaic solar panels, and determine the minimum third illumination intensity variance value among the multiple third illumination intensity variance values; The third time window of the photovoltaic solar panel corresponding to the minimum third light intensity variance value is obtained as the target third time window; the weighted sum weights corresponding to the target first time window and the target third time window are determined through the weather state recognition unit according to the weather state information in the target first time window and the target third time window; the time series light intensities of the target first time window and the target third time window are summed according to the weighted sum weights corresponding to the target first time window and the target third time window as the time series light intensity in the target first time window, and the unified time series predicted output voltage and time series predicted output current of the inverters of the target photovoltaic solar panel and multiple associated photovoltaic solar panels are determined according to the time series light intensity in the target first time window.

8. The method according to claim 7, characterized in that The method further comprises: Respectively obtain weather classification information of multiple associated photovoltaic solar panels of the target photovoltaic solar panel, and when the weather classification information of the target photovoltaic solar panel and the multiple associated photovoltaic solar panels are the same, determine the time-series average light intensity within a first time window, the time-series average panel temperature within a second time window, the time-series average output voltage, the time-series average output current, and the time-series average output pressure of the target photovoltaic solar panel and the multiple associated photovoltaic solar panels; Based on the sequential average light intensity of the target photovoltaic solar panel and multiple associated photovoltaic solar panels in the first time window, the sequential average panel temperature, the sequential average output voltage, the sequential average output current and the sequential average output pressure in the second time window, the unified sequential predicted output voltage and sequential predicted output current of the inverters of the target photovoltaic solar panel and multiple associated photovoltaic solar panels are determined.

9. The method according to claim 8, characterized in that The method further comprises: Obtain multiple weighted sum weights corresponding to the target photovoltaic solar panel and multiple associated photovoltaic solar panels in the first time window and the third time window respectively, and determine the target weighted sum weight with the largest value in the first time window among the multiple weighted sum weights; According to the target weighted summation weights corresponding to the first time window and the third time window, the time-series light intensities of the target photovoltaic solar panel and multiple associated photovoltaic solar panels in the first time window and the third time window are summed and averaged, which is used as the time-series average light intensity of the target photovoltaic solar panel and multiple associated photovoltaic solar panels in the first time window.

10. A photovoltaic water pump inverter control system based on deep learning, characterized in that: Comprising means for performing the method according to any one of claims 1 to 9.

Images