Photovoltaic power generation system reverse power regulation method, device, system and storage medium

Abstract

The application relates to a photovoltaic power generation system reverse power regulation method, device and system and a storage medium. According to real-time power of a grid-connected point and a preset power threshold gear, a gear interval of the real-time power is determined. If the gear interval is a target interval of stable power, a predicted power generation power is determined according to a pre-trained machine learning model, and the ascending and descending power of a photovoltaic unit is adjusted according to the predicted power generation power. If the gear interval is a first interval of power rise, a model of standby photovoltaic units that need to be started is determined, and the standby photovoltaic units are controlled to start power generation. If the gear interval is a second interval of power drop, a model of working photovoltaic units that need to be lowered or turned off is determined. Since the model of the photovoltaic unit that needs to be regulated can be accurately selected in different intervals, accurate reverse power regulation of a multi-model photovoltaic power generation system is realized, and power generation efficiency and power grid safety are considered.

Description

Technical Field

[0001] This application relates to the field of photovoltaic power generation technology, and in particular to a method, device, system and storage medium for reverse power regulation of a photovoltaic power generation system. Background Technology

[0002] In photovoltaic (PV) power generation systems, when the PV power output exceeds the local load consumption, the excess energy is fed back to the grid, forming reverse power. To avoid the impact of reverse power on the grid and equipment, reverse power control devices are typically installed to maintain power balance at the grid connection point by reducing PV power output or disconnecting generators. However, current reverse power control methods are mostly designed for single-type generators (such as single-type PV units), and the control strategies are simple, relying solely on simple threshold comparisons to reduce or cut off power. They lack the ability to fine-tune control for scenarios where multiple generator types work together, easily leading to frequent start-stop cycles, overshoot, or under-regulation, affecting power generation efficiency and equipment lifespan. Summary of the Invention

[0003] This application provides a method, device, system, and storage medium for reverse power regulation of a photovoltaic power generation system, in order to solve the technical problem of how to perform precise reverse power regulation on a photovoltaic power generation system containing multiple models.

[0004] In a first aspect, this application provides a method for reverse power regulation of a photovoltaic power generation system, applied to a photovoltaic power generation system comprising multiple photovoltaic units, wherein the multiple photovoltaic units include at least two models; the method includes: Collect the real-time power of the grid connection point of the photovoltaic power generation system; Based on the real-time power and the preset power threshold, the range in which the real-time power is located is determined; wherein, the range includes at least a target range representing power stability, a first range representing power increase, and a second range representing power decrease; When the gear range is the target range, the predicted power generation is determined according to a pre-trained machine learning model, and the power output of the photovoltaic unit is adjusted according to the predicted power generation. When the gear range is the first range, determine the model of the standby photovoltaic unit corresponding to the first range, and control the standby photovoltaic unit to start generating electricity; When the gear range is the second range, determine the model of the working photovoltaic unit corresponding to the second range, reduce the power generation of the working photovoltaic unit or control the working photovoltaic unit to stop generating electricity.

[0005] Optionally, determining the predicted power generation based on a pre-trained machine learning model, and adjusting the boost / down power of the photovoltaic unit based on the predicted power generation, includes: Acquire real-time meteorological data and real-time operating data of photovoltaic units; The predicted power generation within the target time period is determined based on the real-time meteorological data, the real-time operational data, and the pre-trained machine learning model. The predicted power trend is determined based on the predicted power generation and the real-time power. The power generation of the photovoltaic unit is adjusted or maintained based on the predicted power trend.

[0006] Optionally, the training process of the machine learning model includes: Acquire historical operational data, and acquire the corresponding historical meteorological data and time characteristics; Training samples are constructed based on the historical operational data, the historical meteorological data, and the time characteristics; The initial machine learning model is trained based on the training samples to obtain the trained machine learning model. After determining whether to adjust or maintain the power generation of the photovoltaic unit based on the predicted power trend, the method further includes: New samples are constructed based on the real-time meteorological data and the real-time operational data; The machine learning model is incrementally learned based on the new samples to optimize the parameters of the machine learning model.

[0007] Optionally, the plurality of photovoltaic units include at least a first type of photovoltaic unit, a second type of photovoltaic unit, and a third type of photovoltaic unit; the rated power of the first type of photovoltaic unit is greater than the rated power of the second type of photovoltaic unit, and the rated power of the second type of photovoltaic unit is greater than the rated power of the third type of photovoltaic unit; the first interval includes at least a first rise interval, a second rise interval, and a third rise interval; When the gear range is the first range, determine the model of the standby photovoltaic unit corresponding to the first range, and control the standby photovoltaic unit to start generating electricity, including: When the gear range is the first rising range, a first target standby unit is determined from the first type of photovoltaic units, and the first target standby unit is controlled to start generating electricity; When the gear range is the second increase range, a second target standby unit is determined from the second type of photovoltaic units, and the second target standby unit is controlled to start generating electricity; When the gear range is the third increase range, a third target standby unit is determined from the third type of photovoltaic units, and the third target standby unit is controlled to start generating electricity.

[0008] Optionally, the second interval includes at least a first reduction interval, a second reduction interval, a third reduction interval, and a fourth reduction interval; when the range is the second interval, determining the model of the working photovoltaic unit corresponding to the second interval, reducing the power generation of the working photovoltaic unit, or controlling the working photovoltaic unit to stop generating electricity, includes: When the gear range is the first reduction range, a first target working unit is determined from the third type of photovoltaic units, and the power generation of the first target working unit is reduced; When the gear range is the second reduction range, a second target working unit is determined from the second type of photovoltaic units, and the power generation of the second target working unit is reduced; When the gear range is the third reduction range, a third target working unit is determined from the first type of photovoltaic units, and the power generation of the third target working unit is reduced; When the gear range is the fourth reduction range, the power generation of all photovoltaic units is reduced.

[0009] Optionally, before determining the power level range of the real-time power based on the real-time power and a preset power threshold, the method further includes: A unit status model is constructed based on the real-time status of the photovoltaic unit. Determine whether there is a faulty unit in the unit status model; If there are no faulty units in the unit status model, the step of determining the range of the real-time power based on the real-time power and the preset power threshold is executed.

[0010] Optionally, if a faulty unit exists in the unit status model, the faulty unit is marked as unavailable in the unit status model.

[0011] Secondly, this application provides a reverse power regulation device for a photovoltaic power generation system, characterized in that it is applied to a photovoltaic power generation system, the photovoltaic power generation system comprising multiple photovoltaic units, and the multiple photovoltaic units including at least two models; the device comprises: The data acquisition module is used to acquire the real-time power of the grid connection point of the photovoltaic power generation system; The determining module is used to determine the range of the real-time power based on the real-time power and a preset power threshold; wherein the range of the power includes at least a target range representing power stability, a first range representing power increase, and a second range representing power decrease; The first control module is used to determine the predicted power generation based on a pre-trained machine learning model when the gear range is the target range, and to adjust the power increase / decrease of the photovoltaic unit based on the predicted power generation. The second control module is used to determine the standby photovoltaic unit corresponding to the first interval when the gear range is the first interval, and control the standby photovoltaic unit to start generating electricity. The third control module is used to determine the working photovoltaic unit corresponding to the second interval when the gear range is the second interval, and reduce the power generation of the working photovoltaic unit or control the working photovoltaic unit to stop generating electricity.

[0012] Thirdly, this application provides a photovoltaic power generation system, including a processor, a communication interface, a memory, and a communication bus, wherein the processor, the communication interface, and the memory communicate with each other through the communication bus; Memory, used to store computer programs; When the processor executes a program stored in the memory, it implements the reverse power regulation method for a photovoltaic power generation system as described in any embodiment of the first aspect.

[0013] Fourthly, this application provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the reverse power regulation method for a photovoltaic power generation system as described in any embodiment of the first aspect.

[0014] Compared with the prior art, the technical solution provided in this application has the following advantages: The method provided in this application collects the real-time power of the grid-connected point of the photovoltaic power generation system; determines the range of the real-time power based on the real-time power and a preset power threshold; wherein, the range includes at least a target range representing power stability, a first range representing power increase, and a second range representing power decrease; when the range is the target range, the predicted power generation is determined based on a pre-trained machine learning model, and the power increase / decrease of the photovoltaic unit is adjusted based on the predicted power generation; when the range is the first range, the model of the standby photovoltaic unit corresponding to the first range is determined, and the standby photovoltaic unit is controlled to start power generation; when the range is the second range, the model of the working photovoltaic unit corresponding to the second range is determined, and the power generation of the working photovoltaic unit is reduced or the working photovoltaic unit is controlled to stop power generation. This method determines the power level range based on the real-time power at the grid connection point and a preset power threshold. Different strategies are then selected for different power level ranges. If the range is a target range with stable power, the predicted power generation can be determined using a pre-trained machine learning model, and the power output of the photovoltaic (PV) units can be adjusted accordingly. If the range is the first range where power increases, the model of the standby PV unit to be activated can be determined, and that standby PV unit can be controlled to start generating power. If the range is the second range where power decreases, the model of the working PV unit to be reduced or shut down can be determined, and the power generation of that working PV unit can be reduced or stopped. Because the model of the PV unit to be regulated can be accurately selected in different ranges, precise reverse power regulation of multi-model PV power generation systems is achieved, balancing power generation efficiency and grid security. Attached Figure Description

[0015] The accompanying drawings, which are incorporated in and form part of this specification, illustrate embodiments consistent with this application and, together with the description, serve to explain the principles of this application.

[0016] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, for those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0017] One or more embodiments are illustrated by way of example with reference numerals in the accompanying drawings. These illustrations do not constitute a limitation on the embodiments. Elements with the same reference numerals in the drawings are denoted as similar elements. Unless otherwise stated, the figures in the drawings are not to be limited by scale.

[0018] Figure 1 A schematic flowchart of a photovoltaic power generation system reverse power regulation method provided in one embodiment of this application; Figure 2 This is a schematic diagram illustrating a power regulation process based on a machine learning model, provided as an embodiment of this application. Figure 3 A flowchart illustrating a method for reverse power regulation in a photovoltaic power generation system, provided as another embodiment of this application; Figure 4 This application provides a schematic diagram of the structure of a photovoltaic power generation system reverse power regulation device according to one embodiment. Figure 5 This is a schematic diagram of a photovoltaic power generation system provided in one embodiment of this application. Detailed Implementation

[0019] To make the objectives, technical solutions, and advantages of the embodiments of this application clearer, the technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0020] The following disclosure provides numerous different embodiments or examples for implementing various structures of this application. To simplify the disclosure, specific examples of components and arrangements are described below. These are merely examples and are not intended to limit the scope of this application. Furthermore, reference numerals and / or letters may be repeated in different examples. Such repetition is for simplification and clarity and does not in itself indicate a relationship between the various embodiments and / or arrangements discussed.

[0021] To address the technical challenge of precise reverse power regulation for photovoltaic (PV) power generation systems with multiple models, this application provides a method, apparatus, system, and storage medium for reverse power regulation of PV power generation systems, belonging to the solar energy industry within the new energy sector. Based on the real-time power at the grid connection point and a preset power threshold, the real-time power range can be determined. Different strategies are then selected for different ranges. If the range is a stable power target range, a pre-trained machine learning model can be used to determine the predicted power output, and the power output of the PV units can be adjusted accordingly. If the range is the first range of increasing power, the model of the standby PV unit that needs to be activated can be determined, and the standby PV unit can be started to generate power. If the range is the second range of decreasing power, the model of the working PV unit that needs to be reduced or shut down can be determined, and the power output of the working PV unit can be reduced or the working PV unit can be stopped. Because the model of the PV unit to be regulated can be accurately selected in different ranges, precise reverse power regulation of multi-model PV power generation systems is achieved, balancing power generation efficiency and grid security.

[0022] The first embodiment of this application provides a method for reverse power regulation of a photovoltaic power generation system. This method can be applied to a photovoltaic power generation system, which includes multiple photovoltaic units. The multiple photovoltaic units include at least two models, and the rated power of different models of photovoltaic units is different. For example, they may include rated small power units (such as photovoltaic multi-split units), rated medium power units (such as photovoltaic screw chillers), and rated large power units (such as photovoltaic centrifuges). The number of each model can be one or more, and it may include any two or more models.

[0023] Next, the reverse power regulation method of this photovoltaic power generation system will be explained in detail, such as... Figure 1 The reverse power regulation method for the photovoltaic power generation system includes: Step 101: Collect the real-time power of the photovoltaic power generation system at the grid connection point.

[0024] The grid connection point is the electrical connection point where multiple photovoltaic (PV) units of a photovoltaic (PV) power generation system (such as PV multi-unit PV systems, PV screw chillers, PV centrifugal chillers, etc.) are connected to the public power grid. All different types of PV units are connected to the grid at this connection point. The real-time power at the grid connection point is the interactive power at that point. A positive interactive power indicates that PV power generation is less than the local load and is drawing power from the grid; a negative interactive power indicates that PV power generation is greater than the local load and is transmitting power back to the grid.

[0025] Step 102: Determine the range of power levels based on real-time power and preset power threshold levels; wherein the range of power levels includes at least a target range representing power stability, a first range representing power increase, and a second range representing power decrease.

[0026] The power threshold range can include multiple ranges, such as at least a target range representing stable power, a first range representing increased power, and a second range representing decreased power. The first range can further include multiple ranges, and the second range can also include multiple ranges, with no limit on the number. For example, there are seven preset power threshold ranges: A > B > C > D > E > F > G. All range values ​​can be adaptively calibrated based on the rated capacity and load conditions of the grid-connected photovoltaic power generation system and stored in the reverse power control controller. The target range can be between C and D. The first range can include multiple ranges such as those greater than A, greater than B but less than or equal to A, and greater than C but less than or equal to B. The second range can include multiple ranges such as those less than D but greater than or equal to E, less than E but greater than or equal to F, less than F but greater than or equal to G, and less than G.

[0027] The target range refers to a stable power range. Whether the target range involves drawing power from the grid or feeding surplus power into the grid depends on whether the current scenario allows for grid connection of surplus power and the permitted power output. For example, if grid connection is allowed, the permitted power range could be this target range; if not, the target range could be from 0 to a small amount of grid power drawing, with no specific restrictions. If the grid connection point power is above this target range, it requires increasing power generation, i.e., adding more generating units or increasing power output; if the grid connection point power is below this target range, it requires decreasing power generation, i.e., reducing the number of generating units or decreasing power output.

[0028] In one embodiment, before determining the range of the real-time power based on the real-time power and a preset power threshold, the method further includes: constructing a unit state model based on the real-time status of the photovoltaic unit; determining whether there is a faulty unit in the unit state model; and, if there is no faulty unit in the unit state model, performing the step of determining the range of the real-time power based on the real-time power and the preset power threshold.

[0029] In this embodiment, the real-time status of each unit can be monitored. For example, a unit status model can be constructed based on the real-time status of each photovoltaic unit to determine the fault status of each unit. Only when there is no faulty unit in the unit status model will the step of determining the power range based on the real-time power and the preset power threshold be executed. Once a fault is detected in a unit, the status of the faulty unit in the unit status model is immediately marked as unavailable, and the faulty unit is disconnected. After the fault is resolved, it can be manually or automatically restored to online status.

[0030] Step 103: When the range is the target range, determine the predicted power generation based on the pre-trained machine learning model, and adjust the power increase / decrease of the photovoltaic unit according to the predicted power generation.

[0031] When the target range is within the range of the power output, a machine learning model trained with accumulated daily operating data can be used for power generation prediction feedforward control. The power output of the photovoltaic unit can be adjusted according to the predicted power output. For example, the fluctuation of sunlight can be predicted in advance, and the unit can be preventively controlled to improve the system response speed and stability.

[0032] In one embodiment, the process of determining the predicted power generation based on a pre-trained machine learning model and adjusting the power output of the photovoltaic unit based on the predicted power generation includes: acquiring real-time meteorological data and real-time operating data of the photovoltaic unit; determining the predicted power generation within a future target time period based on the real-time meteorological data, real-time operating data, and the pre-trained machine learning model; determining the predicted power trend based on the predicted power generation and real-time power; and determining whether to adjust or maintain the power generation of the photovoltaic unit based on the predicted power trend.

[0033] In this embodiment, when the power range is the target range, the predicted power generation can be determined based on a pre-trained machine learning model, and the power increase / decrease of the photovoltaic unit can be adjusted according to the predicted power generation. Specifically, real-time meteorological data and real-time operating data of the photovoltaic unit are acquired. Real-time meteorological data may include, for example, solar irradiance, ambient temperature, and season, while real-time operating data of the photovoltaic unit may include, for example, module temperature, voltage, and current data. Based on the real-time meteorological data, real-time operating data, and the pre-trained machine learning model, the predicted power generation within the target time period is determined. The predicted power generation trend is determined based on the predicted power generation and the real-time power. The power generation of the photovoltaic unit is adjusted or maintained based on the predicted power trend. Since the predicted power generation within the target time period can be predicted in advance, the corresponding photovoltaic unit can be prepared with power increase / decrease reserves in advance, avoiding delayed adjustments.

[0034] In one specific embodiment, a flowchart illustrating power regulation based on a machine learning model is shown below. Figure 2First, the model is trained and updated based on historical operating data, meteorological data, and time data to obtain a power generation prediction model. When the power generation prediction model is running, real-time data is collected and input into the power generation prediction model. The model outputs the power generation prediction value for the next T time period. Based on the power generation prediction value, the power trend can be predicted, and it can be determined whether power regulation needs to be carried out in advance. If so, instructions are issued to the regulating units; if not, the current instructions are maintained.

[0035] In this embodiment, the predicted power generation within the target time period can be predicted in advance based on the power generation prediction model, and the corresponding photovoltaic units can be prepared for power increase and decrease in advance to avoid delayed adjustment.

[0036] In one embodiment, the training process of the machine learning model includes: acquiring historical operating data, and acquiring historical meteorological data and time features corresponding to the historical operating data; constructing training samples based on the historical operating data, historical meteorological data, and time features; training the initial machine learning model based on the training samples to obtain a trained machine learning model; after determining whether to adjust or maintain the power generation of the photovoltaic unit based on the predicted power trend, the method further includes: constructing new samples based on real-time meteorological data and real-time operating data; and incrementally learning the machine learning model based on the new samples to optimize the parameters of the machine learning model.

[0037] In this embodiment, the first step is to collect and preprocess historical data. Historical operating data is extracted from the photovoltaic power generation system's operation records, including the power generation of various types of units, real-time power at the grid connection point, photovoltaic voltage and current, and module temperature. At the same time, historical meteorological data for the corresponding time period, such as solar irradiance and ambient temperature, are obtained. Time features, such as time and season identifiers, are extracted. The above multi-source data are normalized, and an input feature vector is constructed to form a training sample set and a validation sample set.

[0038] Secondly, based on the constructed training sample set, a suitable machine learning model, such as a Long Short-Term Memory (LSTM) network, is selected for training. During the training process, historical operating data, meteorological data, and time features are used as model inputs, and the predicted photovoltaic power generation within a preset future time period is used as the output. The trained machine learning model is then used as the power generation prediction model and deployed in the inverse power control device.

[0039] Finally, during system online operation, incremental updates to the power generation prediction model are implemented. As real-time operational data and meteorological data continue to accumulate, when the number of newly added samples reaches a preset threshold, the model update process is triggered. This involves using the newly added data in the background to incrementally learn the existing model and fine-tune weights or model parameters. After the update is complete, the latest power generation prediction model continuously provides accurate power trend predictions for the target range.

[0040] In this embodiment, since machine learning models can be used to predict the operating status and power generation of photovoltaic units, feedforward regulation can be performed in advance to actively suppress reverse power trends and improve the stability of photovoltaic power generation system operation. Furthermore, the machine learning model can be updated based on incremental samples to improve the accuracy of prediction.

[0041] Step 104: If the gear range is the first range, determine the model of the standby photovoltaic unit corresponding to the first range and control the standby photovoltaic unit to start generating electricity.

[0042] If the power range is the first range of increased power, the model of the standby photovoltaic unit that needs to be turned on can be determined, and the standby photovoltaic unit can be controlled to start generating electricity.

[0043] In one embodiment, the plurality of photovoltaic units include at least a first type of photovoltaic unit, a second type of photovoltaic unit, and a third type of photovoltaic unit; the rated power of the first type of photovoltaic unit is greater than the rated power of the second type of photovoltaic unit, and the rated power of the second type of photovoltaic unit is greater than the rated power of the third type of photovoltaic unit; the first interval includes at least a first rising interval, a second rising interval, and a third rising interval.

[0044] When the gear range is the first range, determine the model of the standby photovoltaic unit corresponding to the first range, and control the standby photovoltaic unit to start generating electricity, including: when the gear range is the first increase range, determine the first target standby unit from the first type of photovoltaic units, and control the first target standby unit to start generating electricity; when the gear range is the second increase range, determine the second target standby unit from the second type of photovoltaic units, and control the second target standby unit to start generating electricity; when the gear range is the third increase range, determine the third target standby unit from the third type of photovoltaic units, and control the third target standby unit to start generating electricity.

[0045] In this embodiment, the photovoltaic power generation system may include a first type of photovoltaic unit, a second type of photovoltaic unit, and a third type of photovoltaic unit. For example, the first type of photovoltaic unit is a rated large-power unit (such as a photovoltaic centrifuge), the second type of photovoltaic unit is a rated medium-power unit (such as a photovoltaic screw compressor), and the third type of photovoltaic unit is a rated small-power unit (such as a photovoltaic multi-split unit). For example, seven power threshold levels are preset: A level > B level > C level > D level > E level > F level > G level. The target range can be the range between C level and D level. The first range may include a first increase range (greater than A level), a second increase range (greater than B level and less than or equal to A level), and a third increase range (greater than C level and less than or equal to B level).

[0046] When the real-time power is greater than the power value of the A-level range, it is in the first increase range. At this time, the largest rated power unit is restored first, so that it starts generating electricity with the minimum power output. If there is no large rated power unit in the system, other available units of the same type with the largest rated power are searched in turn and restored with the minimum power output.

[0047] When the real-time power is greater than level B but less than level A, it is in the second increase range. In this case, priority is given to restoring the rated medium-sized unit with the largest rated power, so that it starts generating electricity at the minimum power output. If no rated medium-sized unit is available, then among the rated small-sized units, the available unit with the largest rated power is selected in sequence and restored at the minimum power output.

[0048] When the real-time power is greater than C level but not B level, it is in the third rising range. At this time, the rated small power unit with the largest rated power will be restored first, so that it can start generating electricity with the minimum power output.

[0049] In this embodiment, the model of the standby photovoltaic unit to be turned on can be determined according to the range of the real-time power. Then, one of the photovoltaic units of that model can be turned on. Since the rated power of the turned-on standby photovoltaic unit is adapted to the current real-time power, the photovoltaic power generation system can be quickly stabilized.

[0050] Step 105: If the range is the second range, determine the model of the working photovoltaic unit corresponding to the second range, reduce the power generation of the working photovoltaic unit or control the working photovoltaic unit to stop generating electricity.

[0051] This method determines the power level range based on the real-time power at the grid connection point and a preset power threshold. Different strategies are then selected for different power level ranges. If the range is a target range with stable power, the predicted power generation can be determined using a pre-trained machine learning model, and the power output of the photovoltaic (PV) units can be adjusted accordingly. If the range is the first range where power increases, the model of the standby PV unit to be activated can be determined, and that standby PV unit can be controlled to start generating power. If the range is the second range where power decreases, the model of the working PV unit to be reduced or shut down can be determined, and the power generation of that working PV unit can be reduced or stopped. Because the model of the PV unit to be regulated can be accurately selected in different ranges, precise reverse power regulation of multi-model PV power generation systems is achieved, balancing power generation efficiency and grid security.

[0052] In one embodiment, the second interval includes at least a first reduction interval, a second reduction interval, a third reduction interval, and a fourth reduction interval. When the range is the second interval, determining the model of the working photovoltaic unit corresponding to the second interval and reducing the power generation of the working photovoltaic unit or controlling the working photovoltaic unit to stop generating electricity includes: when the range is the first reduction interval, determining a first target working unit from the third type of photovoltaic units and reducing the power generation of the first target working unit; when the range is the second reduction interval, determining a second target working unit from the second type of photovoltaic units and reducing the power generation of the second target working unit; when the range is the third reduction interval, determining a third target working unit from the first type of photovoltaic units and reducing the power generation of the third target working unit; and when the range is the fourth reduction interval, reducing the power generation of all photovoltaic units.

[0053] In this embodiment, the photovoltaic power generation system may include a first type of photovoltaic unit, a second type of photovoltaic unit, and a third type of photovoltaic unit. For example, the first type of photovoltaic unit is a rated large-power unit (such as a photovoltaic centrifuge), the second type of photovoltaic unit is a rated medium-power unit (such as a photovoltaic screw compressor), and the third type of photovoltaic unit is a rated small-power unit (such as a photovoltaic multi-split unit). For example, seven power threshold levels are preset: A level > B level > C level > D level > E level > F level > G level. The target range can be the range between C level and D level. The second range may include a first reduction range (less than D level and greater than or equal to E level), a second reduction range (less than E level and greater than or equal to F level), a third reduction range (less than F level and greater than or equal to G level), and a fourth reduction range (less than G level).

[0054] When the real-time power is less than D level but greater than or equal to E level, it is in the first reduction range. At this time, the power generation of rated small power units is reduced first. If there are no rated small power units, the power of rated medium power units and rated large power units is reduced in sequence. When all generating units report that they can no longer reduce the power generation, the unit with the lowest current power generation is shut down (the shutdown target is selected in the order of rated small power units, rated medium power units, and rated large power units).

[0055] When the real-time power is less than level E but greater than or equal to level F, it is in the second reduction range. In this case, the power generation of the rated medium-sized power unit is reduced first. If there is no rated medium-sized power unit, the power of the rated large-sized power unit and then the rated small-sized power unit are reduced in sequence. When all units can no longer reduce their power, the unit with the lowest power generation is shut down in the following order: first the rated medium-sized power unit, then the rated small-sized power unit, and finally the rated large-sized power unit.

[0056] When the real-time power is less than F level but greater than or equal to G level, it is in the third reduction range. In this case, the power generation of the rated large power units is reduced first. If there are no rated large power units, the power of the rated medium power units and rated small power units is reduced in sequence. When all units cannot reduce their power, the unit with the lowest power generation is shut down: first the rated large power units, then the rated medium power units, and finally the rated small power units.

[0057] When the real-time power is lower than the G level, it is in the fourth reduction zone. At this time, all generator sets are reduced, regardless of model. When all units cannot reduce power, the photovoltaic unit with the highest generating capacity is disconnected. At the same time, the photovoltaic power generation cutoff timer is started. If the real-time power at the grid connection point cannot rise back to above the G level within the set time period, a full photovoltaic cutoff operation is performed, disconnecting all photovoltaic units generating power.

[0058] In this embodiment, the model of the working photovoltaic unit that needs to be reduced or turned off can be determined, and a photovoltaic unit of that model can be selected for adjustment, such as reducing the power generation of the working photovoltaic unit or controlling the working photovoltaic unit to stop generating electricity. Since the rated power of the adjusted working photovoltaic unit is adapted to the current real-time power, the photovoltaic power generation system can be quickly stabilized.

[0059] In one specific embodiment, a flowchart of a photovoltaic power generation system reverse power regulation method is shown below. Figure 3 ,include: Collect the real-time power P at the grid connection point.

[0060] Build and update the unit status model.

[0061] The system determines if the generator unit is faulty. If a fault is found, it stops sending power regulation commands to the faulty unit and triggers the unit to shut down. If no fault is found, it determines the power value P at the appropriate level. The photovoltaic power generation system can include three types of photovoltaic generator units: Type 1, Type 2, and Type 3. For example, Type 1 photovoltaic generator units are rated large-power units (such as photovoltaic centrifugal chillers), Type 2 photovoltaic generator units are rated medium-power units (such as photovoltaic screw chillers), and Type 3 photovoltaic generator units are rated small-power units (such as photovoltaic multi-split systems). For example, seven power threshold levels are preset: A > B > C > D > E > F > G. The target range can be the range between C and D gears. The first range can include the first rise range (greater than A gear), the second rise range (greater than B gear and less than or equal to A gear), and the third rise range (greater than C gear and less than or equal to B gear). The second range can include the first fall range (less than D gear and greater than or equal to E gear), the second fall range (less than E gear and greater than or equal to F gear), the third fall range (less than F gear and greater than or equal to G gear), and the fourth fall range (less than G gear).

[0062] If P > A, then priority is given to restoring the rated large-capacity units, and power regulation of these units is prioritized. For example, the largest rated large-capacity unit is restored first, allowing it to start generating power at its minimum output. If no rated large-capacity unit exists in the system, then other available units of the same type with the largest rated power are searched sequentially and restored at their minimum output.

[0063] If B < P ≤ A, then the rated medium-power units are restored first, and power regulation of medium-power units is prioritized. For example, the rated medium-power unit with the largest rated power is restored first, allowing it to start generating electricity at its minimum power output. If no rated medium-power unit exists, then among the rated small-power units, the available unit with the largest rated power is selected in sequence and restored at its minimum power output.

[0064] If C < P ≤ B, then the rated small-capacity generator units will be restored first, and power regulation of the small-capacity generator units will be prioritized. For example, the rated small-capacity generator unit with the largest rated power will be restored first, so that it starts generating electricity at its minimum power output.

[0065] If D < P ≤ C, then feedforward control is predicted based on machine learning models to anticipate and adjust light fluctuations in advance.

[0066] If E < P ≤ D, then the rated small-capacity generator units are adjusted first, and when all units cannot reduce their power, the rated small-capacity generator units are shut down first. For example, the generating power of the rated small-capacity generator units is reduced first. If there are no rated small-capacity generator units, then the rated medium-capacity generator units and rated large-capacity generator units are reduced in sequence. When all generating units report that they cannot continue to reduce their generating power, the generator unit with the lowest current generating power is shut down (the shutdown target is selected in the order of rated small-capacity generator units, rated medium-capacity generator units, and rated large-capacity generator units).

[0067] If F < P ≤ E, then the power of the rated medium-power units is adjusted first, and when all units cannot reduce their power, the rated medium-power units are shut down first. For example, the power generation of the rated medium-power units is reduced first. If there are no rated medium-power units, then the power of the rated large-power units is reduced in sequence, followed by the rated small-power units. When all units cannot further reduce their power, the units with the smallest power generation are shut down in sequence: first the rated medium-power units, then the rated small-power units, and finally the rated large-power units.

[0068] If G < P ≤ F, then the rated power of large-capacity generating units is adjusted first, and if all units cannot reduce their power, the rated large-capacity generating units are shut down first. For example, the generating power of rated large-capacity generating units is reduced first. If no rated large-capacity generating units exist, the power of rated medium-capacity generating units is reduced in sequence, followed by rated small-capacity generating units. When all units cannot reduce their power, the unit with the lowest generating power is shut down: first the rated large-capacity generating units, then the rated medium-capacity generating units, and finally the rated small-capacity generating units.

[0069] If P ≤ G, then all generator sets are adjusted, and if none of the units can reduce their power output, the unit with the highest generating capacity is prioritized for shutdown. For example, all generator sets are reduced in power output, regardless of model. When none of the units can reduce their power output, the photovoltaic unit with the highest generating capacity is shut down. Simultaneously, a photovoltaic power generation shutdown timer is started. If the real-time power output at the grid connection point fails to recover to above G level within the set time period, a full photovoltaic shutdown operation is performed, disconnecting all photovoltaic units generating power from the grid.

[0070] In this embodiment, because the model of the photovoltaic unit to be regulated can be accurately selected in different ranges, precise reverse power regulation of photovoltaic power generation systems of multiple models is achieved, taking into account both power generation efficiency and grid security.

[0071] Based on the same technical concept, the second embodiment of this application provides a reverse power regulation device for a photovoltaic power generation system, such as... Figure 4 The device is applied to a photovoltaic power generation system, wherein the photovoltaic power generation system includes multiple photovoltaic units, and the multiple photovoltaic units include at least two models; the device includes: The acquisition module 401 is used to acquire the real-time power of the grid connection point of the photovoltaic power generation system; The determining module 402 is used to determine the range of the real-time power based on the real-time power and the preset power threshold range; wherein the range of the range includes at least a target range representing power stability, a first range representing power increase, and a second range representing power decrease. The first control module 403 is used to determine the predicted power generation based on a pre-trained machine learning model when the gear range is the target range, and to adjust the power increase / decrease of the photovoltaic unit based on the predicted power generation. The second control module 404 is used to determine the model of the standby photovoltaic unit corresponding to the first interval when the gear range is the first interval, and control the standby photovoltaic unit to start generating electricity. The third control module 405 is used to determine the model of the working photovoltaic unit corresponding to the second range when the range is the second range, and reduce the power generation of the working photovoltaic unit or control the working photovoltaic unit to stop generating electricity.

[0072] This device can determine the power level range of the real-time power based on the real-time power at the grid connection point and the preset power threshold level. Then, it selects different strategies for different power level ranges. If the power level range is the target range where the power is stable, it can determine the predicted power generation based on a pre-trained machine learning model and adjust the power increase / decrease of the photovoltaic units according to the predicted power generation. If the power level range is the first range where the power increases, it can determine the model of the standby photovoltaic unit that needs to be turned on and control the standby photovoltaic unit to start generating power. If the power level range is the second range where the power decreases, it can determine the model of the working photovoltaic unit that needs to be reduced or turned off and reduce the power generation of the working photovoltaic unit or control the working photovoltaic unit to stop generating power. Because it can accurately select the model of the photovoltaic unit to be regulated in different ranges, it realizes precise reverse power regulation of multi-model photovoltaic power generation systems, taking into account both power generation efficiency and grid security.

[0073] like Figure 5 As shown in the figure, this application provides a photovoltaic power generation system, including a processor 111, a communication interface 112, a memory 113, and a communication bus 114, wherein the processor 111, the communication interface 112, and the memory 113 communicate with each other through the communication bus 114. Memory 113 is used to store computer programs; In one embodiment of this application, the processor 111, when executing the program stored in the memory 113, implements the photovoltaic power generation system reverse power regulation method provided in any of the foregoing method embodiments.

[0074] The communication bus mentioned above can be a Peripheral Component Interconnect (PCI) bus or an Extended Industry Standard Architecture (EISA) bus, etc. This communication bus can be divided into address bus, data bus, control bus, etc. For ease of illustration, only one thick line is used to represent it in the diagram, but this does not mean that there is only one bus or one type of bus.

[0075] The communication interface is used for communication between the aforementioned terminal and other devices.

[0076] The memory may include random access memory (RAM) or non-volatile memory, such as at least one disk storage device. Optionally, the memory may also be at least one storage device located remotely from the aforementioned processor.

[0077] The processors mentioned above can be general-purpose processors, including central processing units (CPUs), network processors (NPs), etc.; they can also be digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, or discrete hardware components.

[0078] This application also provides a computer-readable storage medium storing a computer program thereon, which, when executed by a processor, implements the photovoltaic power generation system reverse power regulation method as provided in any of the foregoing method embodiments.

[0079] The device embodiments described above are merely illustrative. The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the modules can be selected to achieve the purpose of this embodiment according to actual needs.

[0080] Through the above description of the embodiments, those skilled in the art can clearly understand that each embodiment can be implemented using software plus a general-purpose hardware platform, or of course, using hardware. Based on this understanding, the above technical solutions, in essence or the parts that contribute to the related technology, can be embodied in the form of a software product. This computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, magnetic disk, optical disk, etc., and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute the methods described in the various embodiments or some parts of the embodiments.

[0081] It should be understood that the terminology used herein is for the purpose of describing particular exemplary embodiments only and is not intended to be limiting. Unless the context clearly indicates otherwise, the singular forms “a,” “an,” and “described” as used herein may also include the plural forms. The terms “comprising,” “including,” “containing,” and “having” are inclusive and therefore indicate the presence of the stated features, steps, operations, elements, and / or components, but do not exclude the presence or addition of one or more other features, steps, operations, elements, components, and / or combinations thereof. The method steps, processes, and operations described herein are not construed as requiring them to be performed in a particular order described or illustrated unless the order of performance is explicitly indicated. It should also be understood that additional or alternative steps may be used.

[0082] It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the scope of this application. In the description, suffixes such as "module," "part," or "unit" used to denote elements are used solely for illustrative purposes and have no specific meaning in themselves. Therefore, "module," "part," or "unit" may be used interchangeably.

[0083] The above description is merely a specific embodiment of this application, enabling those skilled in the art to understand or implement this application. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of this application. Therefore, this application is not to be limited to the embodiments shown herein, but is to be accorded the widest scope consistent with the principles and novel features claimed herein.

Claims

1. A method for reverse power regulation in a photovoltaic power generation system, characterized in that, The method is applied to a photovoltaic power generation system, wherein the photovoltaic power generation system includes multiple photovoltaic units, and the multiple photovoltaic units include at least two models; the method includes: Collect the real-time power of the grid connection point of the photovoltaic power generation system; Based on the real-time power and the preset power threshold, the range in which the real-time power is located is determined; wherein, the range includes at least a target range representing power stability, a first range representing power increase, and a second range representing power decrease; When the gear range is the target range, the predicted power generation is determined according to a pre-trained machine learning model, and the power output of the photovoltaic unit is adjusted according to the predicted power generation. When the gear range is the first range, determine the model of the standby photovoltaic unit corresponding to the first range, and control the standby photovoltaic unit to start generating electricity; When the gear range is the second range, determine the model of the working photovoltaic unit corresponding to the second range, reduce the power generation of the working photovoltaic unit or control the working photovoltaic unit to stop generating electricity.

2. The method according to claim 1, characterized in that, The process includes determining the predicted power generation based on a pre-trained machine learning model, and adjusting the power output of the photovoltaic unit based on the predicted power generation, including: Acquire real-time meteorological data and real-time operating data of photovoltaic units; The predicted power generation within the target time period is determined based on the real-time meteorological data, the real-time operational data, and the pre-trained machine learning model. The predicted power trend is determined based on the predicted power generation and the real-time power. The power generation of the photovoltaic unit is adjusted or maintained based on the predicted power trend.

3. The method according to claim 2, characterized in that, The training process of the machine learning model includes: Acquire historical operational data, and acquire the corresponding historical meteorological data and time characteristics; Training samples are constructed based on the historical operational data, the historical meteorological data, and the time characteristics; The initial machine learning model is trained based on the training samples to obtain the trained machine learning model. After determining whether to adjust or maintain the power generation of the photovoltaic unit based on the predicted power trend, the method further includes: New samples are constructed based on the real-time meteorological data and the real-time operational data; The machine learning model is incrementally learned based on the new samples to optimize the parameters of the machine learning model.

4. The method according to claim 1, characterized in that, The plurality of photovoltaic units include at least a first type of photovoltaic unit, a second type of photovoltaic unit, and a third type of photovoltaic unit; the rated power of the first type of photovoltaic unit is greater than the rated power of the second type of photovoltaic unit, and the rated power of the second type of photovoltaic unit is greater than the rated power of the third type of photovoltaic unit; the first interval includes at least a first rise interval, a second rise interval, and a third rise interval; When the gear range is the first range, determine the model of the standby photovoltaic unit corresponding to the first range, and control the standby photovoltaic unit to start generating electricity, including: When the gear range is the first rising range, a first target standby unit is determined from the first type of photovoltaic units, and the first target standby unit is controlled to start generating electricity; When the gear range is the second increase range, a second target standby unit is determined from the second type of photovoltaic units, and the second target standby unit is controlled to start generating electricity; When the gear range is the third increase range, a third target standby unit is determined from the third type of photovoltaic units, and the third target standby unit is controlled to start generating electricity.

5. The method according to claim 4, characterized in that, The second range includes at least a first reduction range, a second reduction range, a third reduction range, and a fourth reduction range; when the range is the second range, the model of the working photovoltaic unit corresponding to the second range is determined, and the power generation of the working photovoltaic unit is reduced or the working photovoltaic unit is controlled to stop generating electricity, including: When the gear range is the first reduction range, a first target working unit is determined from the third type of photovoltaic units, and the power generation of the first target working unit is reduced; When the gear range is the second reduction range, a second target working unit is determined from the second type of photovoltaic units, and the power generation of the second target working unit is reduced; When the gear range is the third reduction range, a third target working unit is determined from the first type of photovoltaic units, and the power generation of the third target working unit is reduced; When the gear range is the fourth reduction range, the power generation of all photovoltaic units is reduced.

6. The method according to claim 1, characterized in that, Before determining the power level range of the real-time power based on the real-time power and the preset power threshold, the method further includes: A unit status model is constructed based on the real-time status of the photovoltaic unit. Determine whether there is a faulty unit in the unit status model; If there are no faulty units in the unit status model, the step of determining the range of the real-time power based on the real-time power and the preset power threshold is executed.

7. The method according to claim 6, characterized in that, If a faulty unit exists in the unit status model, the faulty unit will be marked as unavailable in the unit status model.

8. A reverse power regulation device for a photovoltaic power generation system, characterized in that, The device is applied to a photovoltaic power generation system, which includes multiple photovoltaic units, of which at least two models are included; the device includes: The data acquisition module is used to acquire the real-time power of the grid connection point of the photovoltaic power generation system; The determining module is used to determine the range of the real-time power based on the real-time power and a preset power threshold; wherein the range of the power includes at least a target range representing power stability, a first range representing power increase, and a second range representing power decrease; The first control module is used to determine the predicted power generation based on a pre-trained machine learning model when the gear range is the target range, and to adjust the power increase / decrease of the photovoltaic unit based on the predicted power generation. The second control module is used to determine the model of the standby photovoltaic unit corresponding to the first interval when the gear range is the first interval, and control the standby photovoltaic unit to start generating electricity. The third control module is used to determine the model of the working photovoltaic unit corresponding to the second range when the range is the second range, and to reduce the power generation of the working photovoltaic unit or control the working photovoltaic unit to stop generating electricity.

9. A photovoltaic power generation system, characterized in that, It includes a processor, a communication interface, a memory, and a communication bus, wherein the processor, the communication interface, and the memory communicate with each other through the communication bus; Memory, used to store computer programs; The processor, when executing a program stored in the memory, implements the reverse power regulation method of the photovoltaic power generation system according to any one of claims 1-7.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by the processor, it implements the reverse power regulation method for a photovoltaic power generation system as described in any one of claims 1-7.

Images