Control method of photovoltaic power generation system and photovoltaic power generation system

Abstract

The application discloses a control method of a photovoltaic power generation system and the photovoltaic power generation system. The control method comprises the following steps: acquiring a current state parameter of the photovoltaic power generation system in a current control period, and predicting a switching state of a switching unit of a photovoltaic power generation unit when the photovoltaic power generation unit works at maximum power according to the current state parameter and a preset prediction model to obtain a switching state prediction result; and when a next control period of the current control period arrives, the switching unit in a voltage reduction and voltage increase branch is controlled according to the switching state prediction result. In this way, compared with a traditional closed-loop control scheme, a control link is shorter, a dynamic response is faster, the photovoltaic power generation unit can work at maximum power, the utilization efficiency of solar power generation is improved, the power delivered to a vehicle battery is maximized, for an electric vehicle with a smart driving system or an auxiliary driving system and a large power consumption, the vehicle charging time can be effectively shortened, and the utilization efficiency of renewable energy by the vehicle is improved.

Description

Technical Field

[0001] This application relates to the field of photovoltaic power generation technology, and in particular to a control method and a photovoltaic power generation system. Background Technology

[0002] In related technologies, photovoltaic power generation systems typically employ buck-boost converters to regulate the output voltage of the photovoltaic cells and use maximum power point tracking (MPPT) algorithms to control the converter's switching state, thereby maximizing the power output of the photovoltaic cells. However, traditional converter topologies often use single-phase four-switch buck-boost structures, resulting in large inductor current ripple that affects power quality. Furthermore, the converter topology control strategy often relies on input-output voltage comparison for mode switching, leading to frequent switching when voltages are close, resulting in poor system stability and slow dynamic response. Summary of the Invention

[0003] This application provides a control method for a photovoltaic power generation system and a photovoltaic power generation system.

[0004] This application provides a control method for a photovoltaic power generation system. The photovoltaic power generation system includes a photovoltaic power generation unit installed on the outer side of a building roof and a buck-boost circuit. The electrical energy generated by the photovoltaic power generation unit is output to the buck-boost circuit. The buck-boost circuit includes two buck-boost branches that operate alternately and in parallel. Each buck-boost branch includes at least four switching units. The method includes: Obtain the current state parameters of the photovoltaic power generation system in the current control cycle; Based on the current state parameters and the preset prediction model, the switching state of the switching unit when the photovoltaic power generation unit operates at maximum power is predicted, and the switching state prediction result is obtained. When the next control cycle of the current control cycle arrives, a switching unit in one of the buck-boost branches is controlled according to the switching state prediction result.

[0005] In the embodiments of this application, the current state parameters of the photovoltaic power generation system in the current control cycle can be obtained, and the switching state of the switching unit when the photovoltaic power generation unit operates at maximum power can be predicted based on the current state parameters and the preset prediction model, so as to obtain the switching state prediction result. When the next control cycle of the current control cycle arrives, the switching unit in a buck-boost branch is controlled according to the switching state prediction result. Thus, the photovoltaic power generation system can solve the switching state of the switching unit in the buck-boost circuit when operating at maximum power through the operating state parameters of the current control cycle. Compared with the traditional closed-loop control scheme, the control link is shorter and the dynamic response is faster. Furthermore, by directly solving the switching states of the switching units in the buck-boost circuit through a preset predictive model, the mode detection and independent modulation stages in traditional control are eliminated. This ensures that the input and output polarities are the same and achieves seamless and smooth switching between buck and boost modes, avoiding system oscillations caused by frequent mode switching. Simultaneously, by combining the maximum power point tracking target with the underlying circuit control, the photovoltaic power generation unit can operate at maximum power, improving the efficiency of solar power generation and maximizing the power delivered to the vehicle battery. Therefore, for electric vehicles equipped with intelligent driving or assisted driving systems, which consume a lot of electricity, this effectively shortens the charging time and improves the efficiency of renewable energy utilization in intelligent driving vehicles. In addition, the interleaved parallel branch topology combined with model predictive control of the switching states effectively reduces the current ripple amplitude and increases the input current ripple frequency. This reduces the design difficulty and hardware size of the filter circuit and facilitates soft switching of the switching transistors, reducing conduction and turn-off losses. Therefore, it is suitable for a wide input voltage range and can improve the overall system conversion efficiency and power density, adapting to the wide operating conditions required for residential photovoltaic energy storage scenarios.

[0006] In some embodiments provided in this application, the buck-boost circuit includes a first branch and a second branch. The first branch includes a first switching unit, a second switching unit, a third switching unit, a fourth switching unit, and a first inductor. The second branch includes a fifth switching unit, a sixth switching unit, a seventh switching unit, an eighth switching unit, and a second inductor. The first terminal of the first switching unit is electrically connected to the first output terminal of the photovoltaic power generation unit. The second terminal of the first switching unit and the first terminal of the second switching unit are both electrically connected to the first terminal of the first inductor. The second terminal of the second switching unit is electrically connected to a preset busbar. The second terminal of the first inductor is electrically connected to the second terminal of the third switching unit and the first terminal of the fourth switching unit. The first terminal of the third switching unit is electrically connected to the first output terminal of the buck-boost circuit. The second terminal of the fourth switching unit is electrically connected to the preset busbar. The preset busbar is formed by electrically connecting the second output terminal of the photovoltaic power generation unit and the second output terminal of the buck-boost circuit. The first end of the fifth switch unit is electrically connected to the first output end of the photovoltaic power generation unit. The second end of the fifth switch unit and the first end of the sixth switch unit are both electrically connected to the first end of the second inductor. The second end of the sixth switch unit is electrically connected to the preset bus. The second end of the second inductor is electrically connected to the second end of the seventh switch unit and the first end of the eighth switch unit. The first end of the seventh switch unit is electrically connected to the first output end of the buck-boost circuit. The second end of the eighth switch unit is electrically connected to the preset bus.

[0007] Thus, in this embodiment, the buck-boost circuit can achieve dual-path power current sharing through the interleaved parallel structure of the first and second branches. Compared to a single-path buck-boost circuit, the current stress on a single device is halved, allowing for the selection of devices with lower specifications and lower conduction losses, effectively reducing hardware costs and operating losses. Furthermore, the interleaved operation of the two branches causes the ripple components of the inductor current to cancel each other out at the input and output terminals, significantly reducing the total current ripple amplitude and doubling the ripple frequency. This can greatly reduce the size and capacity of the filter components and improve the power density of the system.

[0008] In some embodiments provided in this application, the step of predicting the switching state of the switching unit when the photovoltaic power generation unit operates at maximum power based on the current state parameters and a preset prediction model, and obtaining the switching state prediction result, includes: Based on the current state parameters and the preset prediction model, the switching states of the two switching units in one of the step-down and step-up branches are predicted when the photovoltaic power generation unit operates at maximum power, and the switching state prediction results are obtained.

[0009] Thus, in this embodiment, the switching states of two switching units in a step-down / boost branch when the photovoltaic power generation unit operates at maximum power can be predicted based on the current state parameters and the preset prediction model, thereby obtaining the switching state prediction results. This changes the solution scope of the model prediction from eight switching units in a dual-branch system to two switching units in a single-branch system, reducing the solution dimension and the amount of computation required, thereby reducing the computational delay of a single prediction and effectively reducing the requirements for the computing power of the controller.

[0010] In some embodiments provided in this application, the current state parameter includes at least the output current of the first inductor or the output current of the second inductor. The step of predicting the switching state of the switching unit when the photovoltaic power generation unit operates at maximum power based on the current state parameter and a preset prediction model, to obtain the switching state prediction result, includes: Based on the current state parameters and the preset prediction model, a cost function containing target constraints is constructed, wherein the target constraints are used to constrain the ripple in the output current and / or the duty cycle of the switching unit; The cost function is solved to predict the switching state of the switching unit when the photovoltaic power generation unit operates at maximum power, and the switching state prediction result is obtained.

[0011] Thus, in this embodiment of the application, a cost function containing target constraints can be constructed and solved based on the current state parameters and the preset prediction model to predict the switching state of the switching unit when the photovoltaic power generation unit operates at maximum power, thereby obtaining the switching state prediction result. In this way, the current ripple and / or the duty cycle of the switching unit can be used as constraints of the cost function, thereby reducing the ripple amplitude of the branch output current, improving the output power quality, and / or optimizing the working state of the switching devices, reducing switching losses, and improving the overall energy conversion efficiency of the photovoltaic power generation system.

[0012] In some embodiments provided in this application, each of the buck-boost branches further includes an inductor. The current state parameters include the first input voltage, first input current, and first output voltage of the buck-boost circuit in the current control cycle, and the second output current of the inductor in the current control cycle. The step of predicting the switching state of the switching unit when the photovoltaic power generation unit operates at maximum power based on the current state parameters and a preset prediction model, to obtain the switching state prediction result, includes: Based on the first input voltage, the first input current, and the second output current, determine the reference operating voltage of the buck-boost circuit and the reference operating current of the inductor when the photovoltaic power generation unit operates at maximum power. The first input voltage, the first output voltage, the reference operating voltage, and the reference operating current are input into the preset prediction model to obtain the switch state prediction result.

[0013] Thus, in this embodiment, the reference operating voltage of the buck-boost circuit and the reference operating current of the inductor when the photovoltaic power generation unit operates at maximum power can be determined based on the first input voltage, the first input current, and the second output current. The first input voltage, the first output voltage, the reference operating voltage, and the reference operating current are then input to a preset prediction model to obtain the switching state prediction result. This achieves the combination of maximum power point tracking and model predictive control, avoids the control deviation and delay caused by multi-layer control loop switching, and effectively improves the energy utilization efficiency of the photovoltaic power generation unit.

[0014] In some embodiments provided in this application, determining the reference operating voltage of the buck-boost circuit and the reference operating current of the inductor when the photovoltaic power generation unit operates at maximum power, based on the first input voltage, the first input current, and the second output current, includes: Based on the first input voltage and the first input current, determine the reference operating voltage of the photovoltaic power generation unit when it operates at maximum power; The reference operating current is determined based on the voltage difference between the first input voltage and the reference operating voltage.

[0015] Thus, in this embodiment, the reference operating voltage of the photovoltaic power generation unit when it operates at maximum power can be determined based on the first input voltage and the first input current, and the reference operating current can be determined based on the voltage difference between the first input voltage and the reference operating voltage. This can, to a certain extent, ensure the validity of the reference operating voltage and reference operating current, and avoid the cumulative error caused by intermediate power conversion, thereby improving the tracking accuracy of the maximum power point. Furthermore, the method of deriving the reference operating current based on the voltage difference has a simple and clear computational logic, eliminating the need for complex multi-parameter iterations and multi-loop control parameter tuning, reducing the computational load, and enabling adaptation to higher frequency control cycles.

[0016] In some embodiments provided in this application, determining the reference operating voltage of the photovoltaic power generation unit when it operates at maximum power based on the first input voltage and the first input current includes: Based on the first input voltage and the preset power-voltage curve of the photovoltaic power generation unit, determine the target power peak point on the preset power-voltage curve that is closest to the first input voltage; Gradually increase the first input voltage; Based on the power and voltage corresponding to the target power peak point, and the first input voltage and first input current before and after each increase of the first input voltage, determine the target value of the first input voltage when the photovoltaic power generation unit operates at maximum power; The target value is determined as the reference operating voltage.

[0017] Thus, in this embodiment, based on the first input voltage and the preset power-voltage curve of the photovoltaic power generation unit, the target power peak point closest to the first input voltage on the preset power-voltage curve can be determined, and the first input voltage can be gradually increased. Based on the power and voltage corresponding to the target power peak point, as well as the first input voltage and first input current before and after each increase in the first input voltage, the target value of the first input voltage when the photovoltaic power generation unit operates at maximum power can be determined, and the target value can be determined as the reference operating voltage. This can adapt to complex lighting scenarios with partial shading, and accurately break free from the limitation of local power peak points when there are multiple peaks in the power-voltage curve, locate the operating voltage corresponding to the global maximum power point, avoid the situation where the traditional single-peak algorithm gets stuck in local optima, and improve the overall power generation efficiency of the system.

[0018] In some embodiments provided in this application, determining the target power peak point on the preset power-voltage curve that is closest to the first input voltage based on the first input voltage and the preset power-voltage curve of the photovoltaic power generation unit includes: The target starting voltage value is determined based on the illumination status of the photovoltaic power generation unit; Control the opening and closing state of the switching unit in the buck-boost circuit so that the value of the first input voltage changes to the target starting voltage value; Based on the first input voltage and the preset power-voltage curve, determine the target power peak point on the preset power-voltage curve that is closest to the first input voltage.

[0019] Thus, in this embodiment, the target starting voltage value can be determined according to the illumination state of the photovoltaic power generation unit, and the opening and closing state of the switching unit in the buck-boost circuit can be controlled so that the value of the first input voltage changes to the target starting voltage value. According to the first input voltage and the preset power-voltage curve, the target power peak point on the preset power-voltage curve that is closest to the first input voltage can be determined. This enables adaptive switching of the power peak point search starting point, matching the corresponding starting voltage for different illumination states, and improving the peak point search efficiency under different operating conditions to a certain extent.

[0020] In some embodiments provided in this application, the photovoltaic power generation system further includes a bidirectional conversion circuit, and the buck-boost circuit and the power supply network are electrically connected to the target load through the bidirectional conversion circuit. The method further includes: The operating state of the bidirectional conversion circuit is controlled so that the electrical energy output by the photovoltaic power generation unit and the electrical energy output by the power supply network flows to the target load, or the electrical energy output by the target load flows to the power supply network.

[0021] Thus, in this embodiment, the operating state of the bidirectional conversion circuit can be controlled to allow the electrical energy output from the photovoltaic power generation unit and the power supply network to flow to the target load, or to allow the electrical energy output from the target load to flow to the power supply network. This realizes the transmission of photovoltaic and grid electrical energy to the target load, as well as the feedback of target load electrical energy to the power supply network, thereby improving the operational economy of the photovoltaic power generation system to a certain extent. This application also provides a photovoltaic power generation system, including a memory and a processor. The memory stores a computer program, which, when executed by the processor, implements the steps in the control method of the photovoltaic power generation system described in some of the above embodiments.

[0022] This application also provides an electronic device, including a memory and a processor, wherein the memory stores a computer program, and when the computer program is executed by the processor, it implements the steps in the control method of the photovoltaic power generation system described in some of the above embodiments.

[0023] This application also provides a computer-readable storage medium storing a computer program that, when executed by one or more processors, implements the steps in the control method of the photovoltaic power generation system described in some of the above embodiments.

[0024] This application also provides a computer program product, including a computer program / instructions that, when executed by a processor, implement the steps in the control method of the photovoltaic power generation system described in some of the above embodiments.

[0025] The photovoltaic power generation system, electronic equipment, computer-readable storage medium, and computer program product provided in this application can obtain the current state parameters of the photovoltaic power generation system in the current control cycle, and predict the switching state of the switching unit when the photovoltaic power generation unit operates at maximum power based on the current state parameters and a preset prediction model, thereby obtaining the switching state prediction result. When the next control cycle of the current control cycle arrives, the switching unit in a buck-boost branch is controlled according to the switching state prediction result. Thus, the photovoltaic power generation system can solve the switching state of the switching unit in the buck-boost circuit when operating at maximum power by using the operating state parameters of the current control cycle. Compared with the traditional closed-loop control scheme, the control link is shorter and the dynamic response is faster. Furthermore, by directly solving the switching states of the switching units in the buck-boost circuit through a preset predictive model, the mode detection and independent modulation stages in traditional control are eliminated. This ensures that the input and output polarities are the same and achieves seamless and smooth switching between buck and boost modes, avoiding system oscillations caused by frequent mode switching. Simultaneously, by combining the maximum power point tracking target with the underlying circuit control, the photovoltaic power generation unit can operate at maximum power, improving the efficiency of solar power generation and maximizing the power delivered to the vehicle battery. Therefore, for electric vehicles equipped with intelligent driving or assisted driving systems, which consume a lot of electricity, this effectively shortens the charging time and improves the efficiency of renewable energy utilization in intelligent driving vehicles. In addition, the interleaved parallel branch topology combined with model predictive control of the switching states effectively reduces the current ripple amplitude and increases the input current ripple frequency. This reduces the design difficulty and hardware size of the filter circuit and facilitates soft switching of the switching transistors, reducing conduction and turn-off losses. Therefore, it is suitable for a wide input voltage range and can improve the overall system conversion efficiency and power density, adapting to the wide operating conditions required for residential photovoltaic energy storage scenarios.

[0026] Additional aspects and advantages of embodiments of this application will be set forth in part in the description which follows, and in part will be obvious from the description, or may be learned by practice of embodiments of this application. Attached Figure Description

[0027] The above and / or additional aspects and advantages of this application will become apparent and readily understood from the description of the embodiments taken in conjunction with the following drawings, wherein: Figure 1 This is one of the flowcharts illustrating the control method of a photovoltaic power generation system in certain embodiments of this application; Figure 2 This is a schematic diagram illustrating one of the application scenarios of the control method for a photovoltaic power generation system in certain embodiments of this application; Figure 3 This is a schematic diagram of a buck-boost circuit in some embodiments of this application; Figure 4a This is a schematic diagram of the current-voltage output characteristic curves of a photovoltaic power generation unit under uniform illumination in certain embodiments of this application; Figure 4b This is a schematic diagram of the power-voltage output characteristic curves of a photovoltaic power generation unit under uniform illumination in certain embodiments of this application; Figure 5a This is a schematic diagram of the current-voltage output characteristic curves of a photovoltaic power generation unit under non-uniform illumination in certain embodiments of this application; Figure 5b This is a schematic diagram of the power-voltage output characteristic curves of a photovoltaic power generation unit under non-uniform illumination in certain embodiments of this application; Figure 6 This is a second schematic flowchart of the control method for a photovoltaic power generation system in certain embodiments of this application; Figure 7 This is a second schematic diagram illustrating an application scenario of the control method for a photovoltaic power generation system in certain embodiments of this application. Figure 8 This is the third schematic diagram illustrating an application scenario of the control method for a photovoltaic power generation system in certain embodiments of this application. Detailed Implementation

[0028] The embodiments of this application are described in detail below. Examples of the embodiments are shown in the accompanying drawings, wherein the same or similar reference numerals denote the same or similar elements or elements having the same or similar functions throughout. The embodiments described below with reference to the accompanying drawings are exemplary and are only used to explain the embodiments of this application, and should not be construed as limiting the embodiments of this application.

[0029] Among related technologies, with the continuous decline in the cost of photovoltaic modules and electrochemical energy storage systems, residential photovoltaic-energy storage systems are increasingly favored, especially in overseas markets. High electricity costs, coupled with government subsidies and incentives, have greatly promoted the popularization and development of these systems. Meanwhile, electric vehicles are gradually replacing traditional gasoline vehicles, becoming a new choice for people's travel. However, electric vehicles mainly rely on low-power AC charging stations, resulting in a relatively slow charging process. Furthermore, with technological advancements, the V2G (Vehicle-to-Grid) model has begun to show certain commercial potential. Some electric vehicle battery management systems already support V2G functionality, allowing vehicles to supply power to the grid in reverse. It is foreseeable that households owning electric vehicles will increasingly tend to utilize the V2G model to achieve efficient energy management and utilization. The principle of V2G is not complicated: electric vehicles are connected to the grid through charging and swapping facilities, charging during off-peak hours and supplying power back to the grid during peak hours or when needed. Like distributed batteries, they can supply power to the grid during peak demand periods and charge and store energy during off-peak hours, thereby achieving peak shaving and valley filling of the grid and improving grid stability and efficiency.

[0030] In photovoltaic (PV) power generation systems, solar cells are crucial components. Their primary function is to convert solar energy into electrical energy for use by downstream batteries or loads. Solar cells exhibit a non-linear power-voltage characteristic, and under certain environmental conditions, there exists a point of maximum power output, known as the maximum power point (MPPT). The purpose of Maximum Power Point Tracking (MPPT) technology is to find this MPPT point and enable the solar cell to operate at that point, thereby improving its conversion efficiency. Traditional MPPT algorithms primarily use incremental conductance methods and perturbation-observation methods. However, these methods suffer from drawbacks such as the need for empirical step size settings, poor stability, low output efficiency, and limited applicability. They are also prone to losing the MPPT point under shading conditions. To address these issues, researchers have proposed several improved schemes, such as a range-optimized PV MPPT method. This method divides the duty cycle range equally, assigns a set number of particles, analyzes the output power characteristics of the PV module during each iteration to obtain the duty cycle corresponding to the possible maximum power point, then narrows the duty cycle range, reallocates the duty cycle, and continues executing MPPT. In addition, some studies have proposed a hybrid algorithm that combines particle swarm optimization and perturbation observation, which can find the maximum power point more accurately and quickly under partial shading conditions.

[0031] In scenarios combining photovoltaic power generation systems with electric vehicle charging and discharging, the system architecture design is particularly critical. A proposed residential photovoltaic-storage-charging-discharging integrated system includes a photovoltaic array, a first bidirectional DC-DC converter, a first power switching module, an energy storage battery system, a second bidirectional DC-DC converter, a second power switching module, a DC bus, a bidirectional AC-DC converter, an Energy Management System (EMS), a cloud management platform, and a terminal application (APP). This system can rapidly charge and discharge electric vehicle batteries, while also functioning as a DC charging station and V2G, thereby improving the utilization rate of the power electronic equipment in the photovoltaic-storage system and the overall economic efficiency of the system. However, this system does not integrate a bidirectional capacitor-inductor-inductor-capacitor (CLLC) circuit. Therefore, it cannot convert AC power from the grid to DC power to charge the power battery, nor can it convert DC power from the power battery to AC power for output to the grid or other devices. Consequently, it is difficult to achieve functions such as vehicle-to-vehicle charging, vehicle power supply to external devices, and participation in grid peak shaving and valley filling. It is worth noting that the bidirectional CLLC resonant converter, as an isolated bidirectional DC-DC converter, is highly favored in DC distribution networks, electric vehicles, and energy storage systems due to its advantages such as symmetrical structure, relatively simple parameter design, ease of high-frequency scaling, and magnetic integration. Furthermore, the bidirectional CLLC resonant converter offers advantages such as electrical isolation, high efficiency, and high power density in battery charging and discharging systems. Therefore, integrating bidirectional CLLC circuits into photovoltaic-energy storage-charging integrated systems is of great significance for achieving efficient and flexible bidirectional energy exchange between vehicles and the power grid.

[0032] Another photovoltaic-storage-charging system and its control method are proposed in related technologies. This system includes a charging device, an energy storage module, a photovoltaic module, and a power converter. The charging device includes a first bidirectional DC-DC converter module, used to connect between the energy storage module and the power converter, or between the photovoltaic module and the power converter. The power converter is used to connect to the energy storage module, the photovoltaic module, and the AC network respectively, and to control the energy dispatch among the charging device, the energy storage module, the photovoltaic module, and the AC network. The innovation of this scheme lies in the separate installation of the charging device, the energy storage module, and the power converter, breaking through the grid-side power limit for household installation, and offering less restriction on installation location and high installation flexibility. However, the photovoltaic DC-DC (Direct Current to Direct Current) module in this system uses a traditional Boost circuit, which can only boost voltage and cannot buck voltage, resulting in high voltage stress on the devices. Furthermore, this system also does not integrate a CLLC circuit, therefore requiring an additional DC charging pile to achieve complete charging and discharging functionality.

[0033] A related technology proposes a model predictive control adaptive adjustment method for a four-switch buck-boost converter. This method involves calculating disturbance observations, real-time correction of the current prediction model based on the estimated disturbance observations, and treating the output voltage, inductance, and inductor internal resistance as system disturbances for overall observation. The observer's estimation results are then used to compensate the prediction model. This method establishes random disturbance observations based on a Luenberger disturbance observer model to account for the uncertainty of the total equivalent resistance of the inductor's series circuit. While ensuring smooth switching of the converter across multiple input voltage modes, it effectively reduces the impact of inaccurate model parameters on the converter's dynamic and steady-state performance. However, the addition of the Luenberger disturbance observer for disturbance compensation increases the algorithm's complexity. Furthermore, using a single-phase four-switch buck-boost topology results in significant inductor current ripple.

[0034] In a model predictive control system for a four-switch Buck-Boost converter, designed for smooth switching between Buck and Boost modes, the inner-loop control module is a predictive current controller that rapidly adjusts the inductor current using model predictive control. The outer-loop control module is a voltage loop based on a proportional-integral regulator, which uses the error between a preset reference voltage and the converter output voltage to calculate the required current reference value for the inductor, thus performing inner-loop inductor current predictive control. However, due to the use of a single-phase four-switch buck-boost topology, this scheme also suffers from significant inductor current ripple.

[0035] In summary, with the widespread adoption of residential photovoltaic-energy storage systems and the gradual maturation of V2G technology, the deep integration of photovoltaic power generation and electric vehicle charging / discharging has become a significant development trend in the energy sector. In this integration process, maximum power point tracking (MPPT) technology determines the photovoltaic energy harvesting efficiency, while the bidirectional DC-DC converter topology and its control strategy determine the flexibility and efficiency of energy flow between the photovoltaic array, energy storage battery, electric vehicle battery, and the power grid. While traditional BOOST circuits and single-phase four-switch buck-boost topologies offer advantages in terms of cost and complexity, they suffer from significant shortcomings in voltage regulation range, current ripple, and system integration.

[0036] Based on the above issues, please refer to the following: Figure 1 , Figure 1 This is a flowchart illustrating a control method for a photovoltaic power generation system in certain embodiments of this application. Specifically, this application provides a control method for a photovoltaic power generation system, which includes a photovoltaic power generation unit installed on the outer side of a building roof and a buck-boost circuit. The electrical energy generated by the photovoltaic power generation unit is output to the buck-boost circuit. The buck-boost circuit includes two alternating and parallel buck-boost branches, each of which includes at least four switching units. The method includes: 01: Obtain the current state parameters of the photovoltaic power generation system in the current control cycle; 02: Based on the current state parameters and the preset prediction model, predict the switching state of the switching unit when the photovoltaic power generation unit operates at maximum power, and obtain the switching state prediction results; 03: When the next control cycle arrives at the current control cycle, control the switching unit in a buck-boost branch according to the switch state prediction result.

[0037] This application also provides a control device for a photovoltaic power generation system. The control method for the photovoltaic power generation system provided in this application can be implemented by the control device provided in this application. Specifically, the control device for the photovoltaic power generation system provided in this application includes a parameter acquisition module, a model prediction module, and a control module. The parameter acquisition module is used to acquire the current state parameters of the photovoltaic power generation system in the current control cycle. The model prediction module is used to predict the switching state of the switching unit when the photovoltaic power generation unit operates at maximum power based on the current state parameters and a preset prediction model, obtaining a switching state prediction result. The control module is used to control the switching unit in a buck-boost branch according to the switching state prediction result when the next control cycle arrives.

[0038] This application also provides a photovoltaic power generation system. The control method of the photovoltaic power generation system provided in this application can be implemented through the photovoltaic power generation system provided in this application. Specifically, the photovoltaic power generation system provided in this application includes a memory and a processor. The memory stores a computer program, and the processor is used to acquire the current state parameters of the photovoltaic power generation system in the current control cycle, and predict the switching state of the switching unit when the photovoltaic power generation unit operates at maximum power according to the current state parameters and a preset prediction model, to obtain the switching state prediction result, and to control the switching unit in a buck-boost branch according to the switching state prediction result when the next control cycle of the current control cycle arrives.

[0039] Specifically, traditional residential photovoltaic energy storage systems often employ a single-branch Buck-Boost topology for DC-DC step-up / step-down conversion. This topology suffers from drawbacks in high-power charging scenarios, including high current stress on switching devices, high output current ripple amplitude, large filter circuit size, and low overall power density. It fails to meet the high-power, miniaturized, and high-efficiency design requirements of residential energy storage systems. Conversely, using a conventional multi-branch parallel scheme results in poor current sharing between branches, unreasonable topology connections leading to unsmooth step-up / step-down mode switching, cumulative switching losses, and insufficient system stability. Furthermore, the complex topology with multiple switches further increases control difficulty, making it difficult to simultaneously achieve wide-range step-up / step-down capability with low ripple and high efficiency.

[0040] Based on this, this application provides a buck-boost circuit topology employing an interleaved parallel dual-branch architecture. Two symmetrical four-switch buck-boost branches are connected in parallel. Each buck-boost branch consists of four switching units and one inductor forming a complete buck-boost conversion path. The two buck-boost branches work collaboratively, in conjunction with a model predictive control mechanism, to transform the traditional single-branch centralized power transmission process into a dual-branch interleaved collaborative power transmission process. Specifically, in this application embodiment, the buck-boost circuit can distribute the total transmission power through the dual-branch topology, reducing the current stress on individual devices. Furthermore, interleaved operation cancels branch current ripple, and a unified model predictive control logic enables seamless switching between buck and boost modes. Thus, the dual-branch interleaved topology design optimizes the power conversion process, ensuring a wide range of buck-boost capabilities while also guaranteeing system operating efficiency and power density.

[0041] In some embodiments, a photovoltaic power generation system can be understood as a complete power conversion and dispatching system consisting of photovoltaic power generation units, power conversion circuits, control units, and supporting loads, energy storage, grid interfaces, and other modules. It can convert solar energy into DC power and complete voltage regulation, transmission, and distribution, and is the core carrier for realizing the local consumption and storage of photovoltaic energy.

[0042] In one example, the photovoltaic power generation system is applied to a residential energy storage scenario, which includes a photovoltaic module array installed on the building roof, an interleaved parallel four-switch buck-boost circuit, an energy storage battery pack, a bidirectional inverter, an AC load and a grid access port, enabling multiple operating modes such as self-consumption, surplus power storage, and grid-connected power feeding.

[0043] In one example, see Figure 2 , Figure 2 This application scenario diagram illustrates the implementation of the present application, namely, in situations such as... Figure 2 In one example shown, the photovoltaic power generation system includes a photovoltaic power generation unit 210 and a buck-boost circuit 220 disposed on the outer side of the roof of building 301, a DC-DC storage circuit 240 and a storage battery 250 electrically connected to the buck-boost circuit 220, and an inverter 260 and a bidirectional conversion circuit 270 electrically connected to the DC-DC storage circuit 240. Additionally, the inverter 260 is electrically connected to the power supply network 302 and the AC load 303, the bidirectional conversion circuit 270 is electrically connected to the vehicle 304, the power supply network 302 is also electrically connected to the vehicle 304, and the vehicle 304 is communicatively connected to a cloud application 305.

[0044] Furthermore, such as Figure 2 As shown, in the photovoltaic power generation system provided in this application embodiment, the electrical energy generated by the photovoltaic power generation unit 210 can be stored in the energy storage battery 250 through the buck-boost circuit 220. Furthermore, the electrical energy generated by the photovoltaic power generation unit 210 and / or the electrical energy stored in the energy storage battery 250 can flow into the power grid through the energy storage DC-DC circuit 240 and the inverter 260, or flow into the vehicle 304 to charge the vehicle through the energy storage DC-DC circuit 240 and the bidirectional conversion circuit 270.

[0045] In some embodiments, a photovoltaic power generation unit can be understood as a power generation device that converts solar energy into direct current based on the photovoltaic effect. It is the source of power input for the entire photovoltaic power generation system and is usually composed of multiple photovoltaic modules connected in series and parallel according to rated voltage and power requirements. It is installed on the outer side of the building roof to obtain sufficient sunlight.

[0046] In one example, a photovoltaic power generation unit can be understood as a photovoltaic string consisting of 6 photovoltaic modules connected in series, with a rated operating voltage of about 180V. Under standard illumination conditions, it can output stable DC power. The output voltage and current of the photovoltaic string change in real time with the light intensity and ambient temperature.

[0047] In some embodiments, the buck-boost circuit can be understood as a DC-DC converter circuit that combines buck and boost functions, also known as a Buck-Boost circuit. It can achieve stable voltage output when the input voltage is higher or lower than the output voltage, adapting to the wide range of fluctuations in photovoltaic output voltage, and ensuring stable power transmission and the realization of maximum power point tracking.

[0048] As an example, in a residential photovoltaic energy storage scenario, the photovoltaic output voltage fluctuates between 100V and 250V, while the subsequent energy storage bus voltage is stable at 200V. The buck-boost circuit can operate in buck mode when the photovoltaic voltage is higher than the bus voltage and in boost mode when it is lower than the bus voltage, thus always maintaining the bus voltage stability.

[0049] In some embodiments, the buck-boost branch can be understood as an independent single-channel conversion unit constituting the buck-boost circuit. Each branch has complete buck-boost conversion capability. Two branches connected in parallel and operating alternately can jointly undertake power transmission, reducing the current stress on single-channel devices.

[0050] In one example, the buck-boost circuit includes two identical buck-boost branches, each capable of independent buck-boost conversion. When the two branches are connected in parallel, the total output power can be doubled, while the current ripple is offset by interleaving the conduction.

[0051] In some embodiments, the switching unit can be understood as a fully controllable power electronic device whose conduction and cutoff states can be controlled by level signals. The conduction loop of the circuit can be changed by the combination of different switching states, thereby realizing the regulation of the output voltage.

[0052] In one example, the switching unit uses an N-channel metal-oxide-semiconductor field-effect transistor (MOSFET) device. The gate is connected to the pulse width modulation (PWM) signal output by the drive circuit. When the drive signal is high, the switch is turned on, and when it is low, the switch is turned off. Each buck-boost branch is configured with four switching units, which can form various working circuits such as buck and boost through different on / off combinations.

[0053] In some embodiments, the control cycle can be understood as a fixed time interval for a complete state sampling, calculation, and control output update. It is the basic time unit of digital control, and the cycle length determines the control bandwidth and dynamic response capability of the photovoltaic power generation system.

[0054] In one example, the control cycle of the buck-boost circuit is set to 20 microseconds, which means that 50,000 state samples and control updates are completed per second, ensuring that the system can quickly track changes in photovoltaic output characteristics.

[0055] In some embodiments, the current state parameter can be understood as the starting time of the current control cycle. The operating state quantity of important nodes in the photovoltaic system, which is collected by the sampling circuit, can reflect the current operating point state of the photovoltaic system.

[0056] For example, the current state parameters include the input voltage and input current of the photovoltaic power generation unit, the output voltage of the buck-boost circuit, and the inductor current of each branch. All parameters are converted into digital signals by the sampling circuit and then input to the controller for calculation.

[0057] In some embodiments, the preset prediction model can be understood as a discretized mathematical model established based on the volt-ampere characteristics of the circuit topology and the volt-second balance principle of the inductor. It can accurately predict the system operating state of the next control cycle based on the current state variables and the on / off combination of the switching units, and is the core computational basis of model predictive control.

[0058] In one example, the preset prediction model is a discrete state-space equation. The input is the current inductor current, input voltage, output voltage and the combination of switching states at the current moment. The output is the predicted value of the inductor current in the next control cycle. It can traverse all feasible switching states and calculate the corresponding system response.

[0059] In some embodiments, maximum power operation can be understood as the photovoltaic power generation unit operating at the peak point of its power-voltage characteristic curve, at which point the output power is the maximum under the current illumination and temperature conditions, which is the core optimization objective of photovoltaic power generation control.

[0060] As an example, under uniform illumination, the power-voltage curve of a photovoltaic power generation unit has a single peak point. By controlling the equivalent load impedance of the buck-boost circuit, the operating voltage of the photovoltaic power generation unit can be maintained at the voltage value corresponding to the peak point, thereby achieving maximum power output.

[0061] In some embodiments, the switch state prediction result can be understood as the optimal switching combination scheme for the next control cycle obtained by traversing all feasible switch states through a preset prediction model and optimizing the solution. This can be converted into a drive signal to control the switching device, so that the photovoltaic power generation system operates close to the expected target.

[0062] In one example, a single buck-boost branch contains four switching units, which can result in a variety of feasible switching state combinations. The optimal on / off combination obtained after optimization by the cost function is the switching state prediction result, which can be directly output to the switching drive circuit.

[0063] To more clearly illustrate the control method of the photovoltaic power generation system provided in the embodiments of this application, please refer to the following exemplary description: At the beginning of the current control cycle, the controller first collects the current state parameters of the photovoltaic power generation system through the sampling circuit, and fully obtains the input voltage and input current of the photovoltaic power generation unit, the output voltage of the buck-boost circuit, and the inductor current of each buck-boost branch, etc. The analog quantities are converted into digital quantities (i.e., current state parameters) and then sent to the controller.

[0064] Then, the controller inputs the collected current state parameters into a preset prediction model, using the goal of ensuring the photovoltaic power generation unit continues to operate at its maximum power point, and performs a step-by-step deduction of all feasible switching state combinations. The preset prediction model can be based on the discretized state equations of the circuit, used to calculate the state quantities such as inductor current, input and output voltage for the next cycle corresponding to each switching combination. It can also filter out the switching state combinations that make the photovoltaic system closest to its maximum power operating point through preset optimization rules, and finally output the switching state prediction results. This process integrates multiple links in traditional control, such as mode judgment, closed-loop regulation, and pulse modulation, into a unified optimization solution process, without the need to separately distinguish between buck and boost operating modes, nor the need for an additional modulator to generate drive pulses.

[0065] Finally, at the end of the current control cycle and the arrival of the next control cycle, the controller converts the switch state prediction results obtained from the aforementioned solution into corresponding level drive signals and outputs them to all switching units in one of the buck-boost branches, controlling each switch to perform the corresponding turn-on or turn-off action. The two buck-boost branches operate alternately, with their respective state sampling, prediction calculations, and control outputs executed alternately with a fixed phase difference. The inductor current ripples of the two branches cancel each other out, jointly maintaining the stable power transmission of the buck-boost circuit.

[0066] Thus, in this embodiment, the current state parameters of the photovoltaic power generation system in the current control cycle can be obtained, and the switching state of the switching unit when the photovoltaic power generation unit operates at maximum power can be predicted based on the current state parameters and the preset prediction model, thereby obtaining the switching state prediction result. When the next control cycle of the current control cycle arrives, the switching unit in a buck-boost branch is controlled based on the switching state prediction result. This allows the photovoltaic power generation system to solve the switching state of the switching unit in the buck-boost circuit when operating at maximum power using the operating state parameters of the current control cycle. Compared with the traditional closed-loop control scheme, the control link is shorter and the dynamic response is faster. Furthermore, by directly solving the switching states of the switching units in the buck-boost circuit through a preset predictive model, the mode detection and independent modulation stages in traditional control are eliminated. This ensures that the input and output polarities are the same and achieves seamless and smooth switching between buck and boost modes, avoiding system oscillations caused by frequent mode switching. Simultaneously, by combining the maximum power point tracking target with the underlying circuit control, the photovoltaic power generation unit can operate at maximum power, improving the efficiency of solar power generation and maximizing the power delivered to the vehicle battery. Therefore, for electric vehicles equipped with intelligent driving or assisted driving systems, which consume a lot of electricity, this effectively shortens the charging time and improves the efficiency of renewable energy utilization in intelligent driving vehicles. In addition, the interleaved parallel branch topology combined with model predictive control of the switching states effectively reduces the current ripple amplitude and increases the input current ripple frequency. This reduces the design difficulty and hardware size of the filter circuit and facilitates soft switching of the switching transistors, reducing conduction and turn-off losses. Therefore, it is suitable for a wide input voltage range and can improve the overall system conversion efficiency and power density, adapting to the wide operating conditions required for residential photovoltaic energy storage scenarios.

[0067] In some embodiments provided in this application, the buck-boost circuit includes a first branch and a second branch. The first branch includes a first switching unit, a second switching unit, a third switching unit, a fourth switching unit, and a first inductor. The second branch includes a fifth switching unit, a sixth switching unit, a seventh switching unit, an eighth switching unit, and a second inductor. The first terminal of the first switching unit is electrically connected to the first output terminal of the photovoltaic power generation unit. The second terminals of the first and second switching units are both electrically connected to the first terminal of the first inductor. The second terminal of the second switching unit is electrically connected to a preset busbar. The second terminal of the first inductor is electrically connected to the second terminal of the third switching unit and the first terminal of the fourth switching unit. The first terminal of the third switching unit is electrically connected to the first output terminal of the buck-boost circuit. The second terminal of the fourth switching unit is electrically connected to the preset busbar, which is formed by electrically connecting the second output terminal of the photovoltaic power generation unit and the second output terminal of the buck-boost circuit. The first end of the fifth switch unit is electrically connected to the first output end of the photovoltaic power generation unit. The second end of the fifth switch unit and the first end of the sixth switch unit are both electrically connected to the first end of the second inductor. The second end of the sixth switch unit is electrically connected to the preset bus. The second end of the second inductor is electrically connected to the second end of the seventh switch unit and the first end of the eighth switch unit. The first end of the seventh switch unit is electrically connected to the first output end of the buck-boost circuit. The second end of the eighth switch unit is electrically connected to the preset bus.

[0068] Specifically, in residential photovoltaic energy storage scenarios, when a single-channel four-switch buck-boost circuit handles high-power charging demands, the switching devices and inductors must bear the full current stress, resulting in severe device overheating, high selection costs, and large inductor current ripple amplitude in a single circuit, requiring large-capacity filter components, thus limiting the improvement of system power density. Furthermore, if a simple two-circuit direct parallel scheme is adopted, the synchronous conduction of the two switches will cause the current ripples to overlap, failing to improve ripple characteristics, and is also prone to uneven current distribution due to differences in device parameters, causing overcurrent damage to some devices and reducing system operational reliability.

[0069] Based on this, the embodiments of this application provide a topology structure of two symmetrical four-switch buck-boost branches interleaved in parallel. The buck-boost circuit is divided into a first branch and a second branch with completely identical structures. Each branch is independently configured with four switching units and one energy storage inductor, forming a symmetrical parallel architecture with common input, common output, and common bus. The two branches are interleaved with a certain phase difference, such as 180°. The total ripple amplitude is reduced by the mutual cancellation of inductor current ripple, and the power current is distributed to reduce the current stress of individual devices. This provides a stable and reliable hardware topology carrier for the model predictive control of photovoltaic power generation systems, effectively improving the system's operating efficiency and reliability.

[0070] In some implementations, the first branch can be understood as an independent conversion unit in one of two interleaved parallel buck-boost branches, possessing complete buck-boost power conversion capabilities, and is structurally symmetrical with the second branch, jointly undertaking the power transmission task of the system.

[0071] In one example, the first branch is configured with four switching units and one energy storage inductor, which can independently perform DC voltage buck-boost conversion and carry half of the system's transmission power.

[0072] In some implementations, the second branch can be understood as another independent conversion unit in the two interleaved parallel buck-boost branches, which has the same topology and device parameters as the first branch, and is connected in parallel to the input and output terminals of the circuit. During operation, it maintains a fixed phase difference with the first branch and conducts alternately.

[0073] As an example, the component selection parameters for the second branch are exactly the same as those for the first branch. When the two branches are connected in parallel, the total transmission power of the circuit can be doubled, while achieving ripple cancellation.

[0074] In some implementations, the first switching unit can be understood as a switching device disposed on the upper bridge arm of the first branch input side, connected in series between the first output terminal of the photovoltaic power generation unit and the first inductor, used to control the conduction circuit on the input side under the step-down condition, and cooperates with the second switching unit to form an input side half-bridge structure.

[0075] In one example, the first switching unit uses an N-channel MOSFET device. When turned on, the power of the photovoltaic power generation unit can be transferred to the first inductor, and when turned off, the upper bridge path on the input side is cut off.

[0076] In some implementations, the second switching unit can be understood as a switching device disposed on the lower bridge arm of the first branch input side, connected in series between the first inductor and the preset bus, and complementary to the first switching unit to jointly complete the circuit switching on the input side.

[0077] Exemplarily, the drive signals of the second switching unit are complementary to those of the first switching unit. When the first switching unit is turned off, the second switching unit is turned on to provide a freewheeling circuit for the first inductor.

[0078] In some implementations, the third switching unit can be understood as a switching device disposed on the upper bridge arm of the first branch output side, connected in series between the first output terminal of the buck-boost circuit and the first inductor, used to control the conduction circuit on the output side under boost conditions, and cooperates with the fourth switching unit to form an output side half-bridge structure.

[0079] In one example, in boost mode, the third switching unit periodically turns on and off, releasing the electrical energy stored in the first inductor to the circuit output, thereby boosting the output voltage.

[0080] In some implementations, the fourth switching unit can be understood as a switching device located on the lower bridge arm of the first branch output side, connected in series between the first inductor and the preset bus, and complementary to the third switching unit to jointly complete the circuit switching on the output side.

[0081] As an example, when the circuit operates in buck mode, the fourth switching unit remains off, and electrical energy is directly transmitted to the output after being regulated by the input half-bridge.

[0082] In some implementations, the first inductor can be understood as an energy storage inductor element located in the middle of the first branch. It is the core energy storage device in the power conversion process, and completes the charging and discharging process by switching the switching state to realize the voltage rise and fall conversion.

[0083] In one example, the first inductor is an iron-silicon-aluminum magnetic core inductor with an inductance value of 100μH. It alternately completes charging and discharging energy storage and release during the switching cycle, and works with the switching unit to realize voltage transformation.

[0084] In some implementations, the fifth switching unit can be understood as a switching device disposed on the upper bridge arm of the second branch input side, with the same function as the first switching unit, and cooperates with the sixth switching unit to form the input side half-bridge structure of the second branch.

[0085] As an example, the drive signal of the fifth switching unit maintains a 180° phase difference with that of the first switching unit, thereby enabling the interleaved operation of the two branches.

[0086] In some implementations, the sixth switching unit can be understood as a switching device located on the lower bridge arm of the second branch input side, with the same function as the second switching unit and complementary conduction with the fifth switching unit.

[0087] In one example, the sixth switching unit operates synchronously and complementaryly with the fifth switching unit to provide a freewheeling path for the second inductor, ensuring stable operation of the branch.

[0088] In some implementations, the seventh switching unit can be understood as a switching device located on the upper bridge arm of the output side of the second branch, with the same function as the third switching unit, and together with the eighth switching unit, it forms the output side half-bridge structure of the second branch.

[0089] As an example, the switching action of the seventh switching unit maintains a 180° phase difference with that of the third switching unit, so that the ripples of the two output currents cancel each other out.

[0090] In some implementations, the eighth switching unit can be understood as a switching device located on the lower bridge arm of the second branch output side, with the same function as the fourth switching unit and complementary conduction with the seventh switching unit.

[0091] In one example, the eighth switch unit works in conjunction with the seventh switch unit to complete the loop control on the output side of the second branch, thereby realizing the step-up and step-down voltage function.

[0092] In some implementations, the second inductor can be understood as an energy storage inductor element located in the middle of the second branch, with parameters completely identical to the first inductor, working alternately with the first inductor to jointly complete the storage and transmission of electrical energy.

[0093] As an example, the inductance value and rated current parameters of the second inductor are exactly the same as those of the first inductor, ensuring that the impedances of the two branches are consistent and achieving uniform current distribution.

[0094] In some implementations, the first output terminal of the photovoltaic power generation unit can be understood as the positive terminal of the DC output of the photovoltaic power generation unit, providing positive voltage input to the two buck and boost branches, serving as the positive terminal of the power input for the entire circuit.

[0095] In one example, the first output terminal of the photovoltaic power generation unit has an output voltage range of 100V~250V, and is connected to the input terminals of the first and fifth switching units to provide the same input voltage for the two branches.

[0096] In some implementations, the second output terminal of the photovoltaic power generation unit can be understood as the negative terminal of the DC output of the photovoltaic power generation unit, which is directly connected to a preset bus and serves as a reference potential for the entire circuit.

[0097] As an example, the second output terminal of the photovoltaic power generation unit is directly connected to the preset busbar to provide a common current loop for all lower arm switching units.

[0098] In some implementations, the preset bus can be understood as a common bus formed by electrically connecting the second output terminal of the photovoltaic power generation unit and the second output terminal of the buck-boost circuit, serving as the negative reference bus of the entire buck-boost circuit and providing a common circuit for all lower bridge arm switching units.

[0099] In one example, the preset bus is the zero potential reference point of the system, and the lower ends of the second, fourth, sixth, and eighth switching units are all connected to this bus, which simplifies the wiring structure of the circuit.

[0100] In some implementations, the first output terminal of the buck-boost circuit can be understood as the positive terminal of the DC output of the buck-boost circuit. The upper bridge arm of the output side of the two branches is connected to this terminal to provide a regulated positive voltage output to the subsequent energy storage unit or load.

[0101] As an example, the first output terminal of the buck-boost circuit is connected to the positive terminal of the energy storage battery, outputting a stable charging voltage and current.

[0102] In some implementations, the second output terminal of the buck-boost circuit can be understood as the negative terminal of the DC output of the buck-boost circuit, which is connected to a preset bus and serves as the negative terminal of the reference potential on the output side.

[0103] In one example, the second output terminal of the buck-boost circuit is connected to the negative terminal of the energy storage battery, and is at the same potential as the preset bus, forming a complete current loop.

[0104] For a clearer illustration of the buck-boost circuit in the embodiments of this application, please refer to [link to relevant documentation]. Figure 3 , Figure 3 This is a schematic diagram of a buck-boost circuit provided in an embodiment of this application. Specifically, in... Figure 3 In the buck-boost circuit shown, V i The input DC voltage of the step-down / boost circuit is also called the output voltage of the photovoltaic power generation unit, which is the potential difference between the first and second output terminals of the photovoltaic power generation unit.

[0105] S1 to S8 are eight fully controlled power electronic switching devices, such as MOSFETs, with anti-parallel freewheeling diodes, which are the first to eighth switching units, belonging to two independent buck-boost branches. Specifically, the switching units belonging to the first buck-boost branch include S1 to S4, where S1 is the first switching unit, S2 is the second switching unit, S3 is the third switching unit, and S4 is the fourth switching unit. S1 is the upper bridge arm switch on the input side, with one end connected to the input voltage V. i Positive terminal; S2 is the lower bridge arm switch on the input side, one end of which is connected to the common negative bus; S3 is the upper bridge arm switch on the output side, one end of which is connected to the output voltage V. out Positive terminal; S4 is the lower bridge arm switch on the output side, with one end connected to the common negative bus. The switching units belonging to the second buck-boost branch include S5 to S8, where S5 is the fifth switching unit, S6 is the sixth, S7 is the seventh, and S8 is the eighth. It is worth noting that the topological positions of the switching units in the first buck-boost branch are completely symmetrical to those in the second buck-boost branch, and the two branches are connected in parallel to the input and output terminals.

[0106] In some implementations, S1 to S8 switch between on / off states via level control signals, and change the circuit conduction loop through different on / off combinations to achieve DC voltage step-up / step-down conversion; during operation, the drive signals of the two branches maintain a 180° phase difference and operate alternately.

[0107] L1 is the energy storage inductor belonging to the first buck-boost branch, i.e., the first inductor, connected in series at the midpoint between the input and output half-bridges of the first buck-boost branch. L2 is the energy storage inductor belonging to the second buck-boost branch, i.e., the second inductor, with parameters and specifications identical to L1, connected in series at the midpoint of the second branch. It can be understood that, through switching states, L1 and L2 alternately complete the charging and discharging energy processes, working in conjunction with the switching unit to achieve voltage level conversion from input to output.

[0108] C is the output filter capacitor, connected in parallel across the output of the buck-boost circuit. It smooths the ripple component of the output voltage and stabilizes the output DC voltage. Combined with the ripple cancellation characteristics of the interleaved parallel topology, it can further reduce output voltage fluctuations and provide stable DC power to the downstream load.

[0109] RL is the output equivalent load, which is the power supply object of the subsequent stage of the buck-boost circuit, such as... Figure 2 The energy storage battery and subsequent DC-DC energy storage circuit in the circuit are the final carriers of electrical energy after voltage regulation.

[0110] V out The output voltage of the buck-boost circuit is the potential difference between the first and second output terminals of the buck-boost circuit. It can achieve stable voltage regulation over a wide range according to the needs of the downstream load / energy storage battery.

[0111] also, Figure 3 China V i The second output terminal (negative) and V out The connection between the second output terminals is the preset bus, which serves as the common reference potential and current return path for the entire circuit.

[0112] In one example, such as Figure 3 When the buck-boost circuit shown is working, the two buck-boost branches operate alternately with a 180° phase difference. When S1 of the first buck-boost branch is turned on, S5 of the second buck-boost branch is turned off, and the charging and discharging processes of L1 and L2 are staggered. During the power transmission process, the electrical energy S output by the photovoltaic power generation unit... i The current is divided into two paths, flowing into the first buck-boost branch and the second buck-boost branch respectively. After voltage regulation by their respective switching units and inductors, the voltages converge at the output terminal and are output to the subsequent circuit. The controller can control the switching state of the two branches separately. By adjusting the on and off timing of the switching units, buck-boost conversion is achieved. At the same time, by utilizing the opposite phase characteristics of the current ripples in the two paths, ripples are canceled out at the input and output terminals, reducing the amplitude of the total current ripple.

[0113] Thus, in this embodiment, the buck-boost circuit can achieve dual-path power current sharing through the interleaved parallel structure of the first and second branches. Compared to a single-path buck-boost circuit, the current stress on a single device is halved, allowing for the selection of devices with lower specifications and lower conduction losses, effectively reducing hardware costs and operating losses. Furthermore, the interleaved operation of the two branches causes the ripple components of the inductor current to cancel each other out at the input and output terminals, significantly reducing the total current ripple amplitude and doubling the ripple frequency. This can greatly reduce the size and capacity of the filter components and improve the power density of the system.

[0114] Furthermore, the symmetrical topology makes it easier to achieve two-way current sharing control, reducing the overcurrent risk caused by differences in device parameters. When a single branch fails, the other branch can continue to operate at a reduced rate, improving the system's operational reliability. At the same time, the four-switch topology combined with the interleaved parallel design can adapt to the wide range of photovoltaic input voltage fluctuations. With the help of model predictive control, it can achieve a seamless and smooth transition between buck and boost modes, ensuring that the photovoltaic power generation unit always operates at the maximum power point, meeting the complex and ever-changing operational needs of residential photovoltaic energy storage systems.

[0115] In some embodiments provided in this application, step 02 above includes: Based on the current state parameters and the preset prediction model, the switching states of two switching units in a step-down and step-up branch of a photovoltaic power generation unit are predicted when the unit operates at maximum power, and the switching state prediction results are obtained.

[0116] The model prediction module in this application embodiment is also used to predict the switching state of two switching units in a step-down and step-up branch when the photovoltaic power generation unit is operating at maximum power, based on the current state parameters and the preset prediction model, and to obtain the switching state prediction result.

[0117] The processor in this embodiment is also used to predict the switching state of two switching units in a buck-boost branch when the photovoltaic power generation unit is operating at maximum power, based on the current state parameters and a preset prediction model, and to obtain the switching state prediction result.

[0118] Specifically, the interleaved parallel four-switch buck-boost circuit has a large number of switching units. If the switching states of all switching units are included in the optimization range of the model prediction, the number of switching state combinations will increase exponentially with the number of switches. This leads to a significant increase in the online computation load of the controller and an increase in the operation delay, making it difficult to match the real-time requirements of high-frequency control cycles. This not only restricts the space for increasing the switching frequency, but also weakens the dynamic response speed of the system to light fluctuations and load changes. At the same time, if the prediction solution is performed simultaneously on all switching units of the two parallel branches, the computing power will increase exponentially, and the interleaved operation characteristics of the two branches cannot be effectively utilized, resulting in control logic redundancy and waste of computing resources. In addition, the excessively large optimization space will increase the risk of the solution deviating from the optimal operating point, making it difficult to ensure the current ripple suppression effect while maintaining computational efficiency.

[0119] Based on this, in some embodiments provided in this application, when performing switch state prediction based on a preset prediction model, only one buck-boost branch is selected as the prediction object, and the switch state of the two switch units in the branch is predicted and solved only. Relying on the inherent characteristic of complementary conduction of the upper and lower switch units of the same bridge arm in the four-switch topology, the switch state of the other two switch units in the same branch is derived from the prediction results of the two switch units. At the same time, the switching timing of the other branch is derived by combining the staggered phase shift operation rules of the two branches. Under the premise of not weakening the buck-boost control and maximum power point tracking effect, the solution dimension and computational amount of the prediction model are compressed.

[0120] In some implementations, two switching units in a buck-boost branch can be understood as two switching units selected from the four switching units of a single four-switch buck-boost branch, belonging to two different bridge arms and without complementary conduction constraints. These two switching units are used as the main control variables for model predictive control. The optimal switching state can be found through finite state enumeration, locking the minimum effective variable set for prediction. Since the upper and lower switching units of the same bridge arm in a four-switch buck-boost topology have a fixed complementary conduction relationship, determining the state of one switching unit directly yields the state of the other. Therefore, it is only necessary to select one switching unit from the input side bridge arm and one from the output side bridge arm to uniquely determine all the switching states of the entire branch.

[0121] In some examples, the two switching units can be selected from the upper bridge arm devices of the input-side bridge arm and the output-side bridge arm, respectively, for example, selecting... Figure 3 In the first step-up and step-down branch, S1 on the input side upper bridge arm and S3 on the output side upper bridge arm are used as prediction objects. When S1 is predicted to be in the on state, S2 on the same bridge arm can be directly deduced to be in the off state. When S3 is predicted to be in the on state, S4 on the same bridge arm can be directly deduced to be in the off state. It is not necessary to include all four switching units in the prediction optimization range.

[0122] In some examples, the switching state combinations of two switching units contain a total of 4 possible permutations, far fewer than the 16 combinations of full prediction for four switches. This can effectively reduce the optimization calculation amount for a single prediction and reduce the computational load on the controller.

[0123] Thus, in this embodiment, the switching states of two switching units in a step-down / boost branch when the photovoltaic power generation unit operates at maximum power can be predicted based on the current state parameters and the preset prediction model, thereby obtaining the switching state prediction results. This changes the solution scope of the model prediction from eight switching units in a dual-branch system to two switching units in a single-branch system, reducing the solution dimension and the amount of computation required, thereby reducing the computational delay of a single prediction and effectively reducing the requirements for the computing power of the controller.

[0124] Furthermore, reducing the solution dimension improves the system's dynamic response speed to fluctuations in illumination and changes in load, ensuring that the photovoltaic power generation unit always operates stably near its maximum power point, thus improving overall power generation efficiency. Moreover, the dimension reduction prediction achieved based on the inherent characteristics of the topology does not change the seamless buck-boost switching capability of the four-switch buck-boost circuit, nor does it affect the ripple cancellation effect of the interleaved parallel structure. It simplifies the implementation of the control algorithm while retaining all the advantages of the original topology, thereby improving the stability and reliability of the system operation.

[0125] In some embodiments provided in this application, the current state parameters include at least the output current of the first inductor or the output current of the second inductor. Therefore, step 02 above includes: Based on the current state parameters and the preset prediction model, a cost function containing target constraint terms is constructed, whereby the target constraint terms are used to constrain the ripple in the output current and / or the duty cycle of the switching unit. The cost function is solved to predict the switching state of the switching unit when the photovoltaic power generation unit operates at maximum power, and the switching state prediction results are obtained.

[0126] The model prediction module in this embodiment is further used to construct a cost function containing target constraints based on the current state parameters and a preset prediction model, and to solve the cost function, so as to predict the switching state of the switching unit when the photovoltaic power generation unit operates at maximum power, and obtain the switching state prediction result. The target constraints are used to constrain the ripple in the output current and / or the duty cycle of the switching unit.

[0127] The processor in this embodiment is further configured to construct a cost function containing target constraints based on current state parameters and a preset prediction model, and to solve the cost function, so as to predict the switching state of the switching unit when the photovoltaic power generation unit operates at maximum power, and obtain the switching state prediction result. The target constraints are used to constrain the ripple in the output current and / or the duty cycle of the switching unit.

[0128] Specifically, traditional control schemes for interleaved parallel four-switch buck-boost circuits often employ a proportional-integral (PI) current loop combined with a dedicated modulator architecture. This requires designing independent control loops and modulation logic for different operating modes such as Buck and Boost. When the input voltage and output voltage are close, the system needs to frequently switch operating modes, which can easily cause control fluctuations and reduce the stability of system operation.

[0129] Based on this, in some embodiments provided in this application, the output current of the inductor in the buck-boost branch can be used as one of the model input parameters, and control objectives such as current ripple suppression and duty cycle optimization can be transformed into target constraint terms in the cost function. By solving the optimal solution of the cost function, the switching state prediction result can be obtained directly. There is no need to design control logic for different modes. Multi-dimensional control objectives can be achieved simultaneously through a single optimization. While ensuring that the photovoltaic power generation unit operates stably at the maximum power point, it also takes into account the performance requirements of low current ripple, low switching loss and fast dynamic response.

[0130] In some implementations, the output current of the first inductor and the output current of the second inductor can be understood as the real-time current values ​​of the branches where the first inductor and the second inductor are located in the two interleaved parallel buck and boost branches, respectively, which can reflect the power transmission status and current ripple level of the branches.

[0131] In some examples, the output current of the first inductor and the output current of the second inductor can be acquired in real time by connecting a current sampling circuit in series in each inductor branch. The sampling frequency is consistent with the control cycle of the system, ensuring that the latest current operation data can be obtained in each control cycle.

[0132] In some implementations, the target constraint can be understood as a quantitative constraint condition embedded in the cost function, which is used to transform the system's control performance index into a computable optimization objective, guide the optimization process to converge toward the desired operating state, and the constraint dimension and weight can be flexibly configured according to the system design requirements.

[0133] In some examples, the target constraint can be set as an output current ripple constraint alone, or as a switching unit duty cycle constraint alone, or both constraints can be set simultaneously and assigned different weight coefficients to adapt to different system performance requirements.

[0134] In some implementations, the cost function can also be called the value function, which can be understood as a quantitative calculation function used in the model predictive control system to evaluate the merits of different switching states. Generally, the smaller the function value, the better the system performance under the corresponding switching state matches the preset control target.

[0135] In some examples, the cost function can be constructed based on the sum of squared errors, which weights and sums the deviations of each constraint term, and the numerical value can intuitively reflect the performance of different switching states.

[0136] In some implementations, the ripple in the output current can be understood as the periodic fluctuation component of the inductor output current based on the DC average value. Excessive ripple amplitude will increase the conduction loss of the line and devices, reduce the power quality, and under long-term effects, it will also affect the service life of downstream devices such as energy storage batteries.

[0137] In some examples, the fluctuation range of the output current can be limited to a preset threshold by constraining the peak amplitude or effective value of the ripple, thereby ensuring the stability of the output power.

[0138] In some implementations, the duty cycle of a switching unit can be understood as the ratio of the on-time of the switching unit to the total duration of the entire cycle within a complete switching cycle. The magnitude of the duty cycle can determine the voltage transformation ratio of the circuit, and is also directly related to the switching losses of the switching device and the branch current ripple level.

[0139] In some examples, by constraining the range of duty cycle values, the switching device can be prevented from operating in the extreme on or extreme off states, thereby reducing stress loss and extending the device's lifespan.

[0140] In some implementations, solving the cost function can be understood as a calculation process that, within all feasible combinations of switching states, calculates the cost function value for each switching state one by one through traversal calculations or numerical optimization algorithms, and finally selects the switching state combination that makes the cost function value optimal.

[0141] In some examples, an enumeration method can be used to calculate all feasible switch state combinations for a single branch one by one, and the optimal solution can be selected within a single control cycle, ensuring the real-time performance of the control.

[0142] To more clearly illustrate the implementation process of switch state prediction in the embodiments of this application, please refer to the following exemplary description: At the beginning of each control cycle, the photovoltaic power generation system first collects the state parameters of the current cycle through the sampling circuit, and at least collects the output current of the first inductor or the output current of the second inductor as the basis input for subsequent prediction calculations.

[0143] The photovoltaic power generation system then invokes a pre-set prediction model and constructs a corresponding cost function based on the collected current state parameters. During the construction process, corresponding target constraints can be added according to the pre-set control objectives of the photovoltaic power generation system. For example, if the photovoltaic power generation system prioritizes ensuring the quality of output power, then the output current ripple is used as the main constraint to limit the maximum amplitude of current fluctuations and ensure the stability of the output current. Alternatively, if the photovoltaic power generation system prioritizes reducing switching losses and optimizing device lifespan, then the duty cycle of the switching unit is used as the main constraint to limit the extreme values ​​of the duty cycle and reduce the working stress on the switching devices. Furthermore, output current ripple and the switching unit duty cycle can also be used as two separate constraints with different weighting coefficients to achieve balanced optimization of multiple control objectives.

[0144] After the cost function is constructed, the photovoltaic power generation system performs a solution operation. For all feasible switching state combinations of the target buck-boost branch, the cost function value under each state is calculated one by one, and the switching state combination that makes the cost function value optimal is selected. This combination is the switching state prediction result of the corresponding buck-boost branch switching unit when the photovoltaic power generation unit is working at the maximum power point.

[0145] Finally, the switch state prediction result is output to the drive module of the photovoltaic power generation system. When the next control cycle arrives, the corresponding buck-boost branch switch unit is driven to perform on / off actions according to the predicted state, completing the adjustment process of a complete control cycle.

[0146] Thus, in this embodiment of the application, a cost function containing target constraints can be constructed and solved based on the current state parameters and the preset prediction model to predict the switching state of the switching unit when the photovoltaic power generation unit operates at maximum power, thereby obtaining the switching state prediction result. In this way, the current ripple and / or the duty cycle of the switching unit can be used as constraints of the cost function, thereby reducing the ripple amplitude of the branch output current, improving the output power quality, and / or optimizing the working state of the switching devices, reducing switching losses, and improving the overall energy conversion efficiency of the photovoltaic power generation system.

[0147] Furthermore, since the cost function is solved based on the current state, there is no need for complex parameter tuning and multi-controller coordination. It can respond quickly to transient conditions such as light fluctuations and load changes. It also has high current tracking accuracy, does not have surge current, and can stably cooperate with the front-end maximum power point tracking algorithm to ensure that the photovoltaic power generation unit always works in the optimal output state, which is fully adaptable to the complex and ever-changing operating scenarios of residential energy storage systems.

[0148] In some embodiments provided in this application, each buck-boost branch further includes an inductor, and the current state parameters include the first input voltage, first input current, and first output voltage of the buck-boost circuit in the current control cycle, and the second output current including the inductor in the current control cycle. Therefore, step 02 above includes: Based on the first input voltage, the first input current, and the second output current, determine the reference operating voltage of the buck-boost circuit and the reference operating current of the inductor when the photovoltaic power generation unit operates at maximum power. The first input voltage, the first output voltage, the reference operating voltage, and the reference operating current are input into the preset prediction model to obtain the switch state prediction result.

[0149] The model prediction module in this application embodiment is also used to determine the reference operating voltage of the buck-boost circuit and the reference operating current of the inductor when the photovoltaic power generation unit operates at maximum power, based on the first input voltage, the first input current and the second output current, and to input the first input voltage, the first output voltage, the reference operating voltage and the reference operating current into the preset prediction model to obtain the switching state prediction result.

[0150] The processor in this embodiment is further configured to determine the reference operating voltage of the buck-boost circuit and the reference operating current of the inductor when the photovoltaic power generation unit operates at maximum power, based on the first input voltage, the first input current and the second output current, and to input the first input voltage, the first output voltage, the reference operating voltage and the reference operating current to a preset prediction model to obtain the switching state prediction result.

[0151] Specifically, considering that in the control scheme of traditional photovoltaic power generation systems, the maximum power point tracking algorithm only outputs the voltage target on the photovoltaic side, which cannot directly match the current control requirements of the buck-boost circuit, it needs to go through multiple control loops in the middle. This not only increases the control delay, but also easily generates control deviations during the conversion process, causing the actual operating state of the circuit to deviate from the photovoltaic maximum power point, thus reducing the utilization efficiency of photovoltaic energy.

[0152] Based on this, in some embodiments provided in this application, the input voltage, input current, output voltage, and inductor current can be collected as basic operating parameters within the current control cycle. The reference operating voltage corresponding to the maximum power point is calculated based on the electrical parameters of the photovoltaic side. Then, the reference operating current of the inductor is derived by combining the circuit operating characteristics. Finally, the actual operating parameters and the two types of reference targets are input into the preset prediction model to achieve a close connection between maximum power point tracking and model predictive control. This provides an accurate target benchmark and operating boundary for switch state prediction, thereby ensuring that the photovoltaic power generation unit always works at the maximum power point while improving the accuracy of predictive control and dynamic response speed.

[0153] In some implementations, the first input voltage can be understood as the DC voltage value at the input terminal of the buck-boost circuit, i.e., the output side of the photovoltaic power generation unit, which can reflect the current operating point position of the photovoltaic power generation unit.

[0154] In some examples, the first input voltage can be acquired by connecting a voltage sampling circuit in parallel across the output of the photovoltaic power generation unit. The sampling frequency is synchronized with the system control cycle to ensure that accurate photovoltaic side voltage data can be obtained in each control cycle.

[0155] In some implementations, the first input current can be understood as the DC current value at the input terminal of the buck-boost circuit and the output side of the photovoltaic power generation unit, and its product with the first input voltage is the real-time output power of the photovoltaic power generation unit.

[0156] In some examples, the first input current can be acquired by connecting a current sensor or sampling resistor in series in the main input circuit. Combined with the sampled first input voltage, the current output power of the photovoltaic power generation unit can be calculated in real time, thereby determining the power change trend.

[0157] In some implementations, the first output voltage can be understood as the DC voltage value at the output of the buck-boost circuit, reflecting the voltage condition of the downstream energy storage unit or load.

[0158] In some examples, the first output voltage may correspond to the real-time terminal voltage of the back-end energy storage battery, and the value of the first output voltage may fluctuate with the battery's charging and discharging state.

[0159] In some implementations, the second output current can be understood as the real-time current value of the inductor in the buck-boost branch.

[0160] In some examples, the second output current can be obtained by connecting independent current sampling units in series in each inductor branch, and the sampling value is updated once in each control cycle, thereby determining the real-time operating status of the branch current.

[0161] In some implementations, the reference operating voltage can be understood as the target voltage value that the input terminal of the buck-boost circuit should maintain when the photovoltaic power generation unit is operating at its maximum power point.

[0162] In some implementations, the reference operating current can be understood as the target current value that the buck-boost branch inductor should reach under the corresponding reference operating voltage.

[0163] In some examples, the reference operating current can be derived from the reference operating voltage, the current output voltage, and the current inductor current based on the power balance of the circuit, thus converting the voltage on the photovoltaic side into a current target that can be directly controlled on the circuit side.

[0164] To more clearly illustrate the process of generating and calculating reference values ​​for switch state prediction in the embodiments of this application, please refer to the following exemplary description: At the beginning of each control cycle, the first input voltage, first input current, first output voltage of the buck-boost circuit, and the second output current of the branch inductor are collected by the sampling circuit.

[0165] After sampling is completed, based on the first input voltage and the first input current, and combined with the power-voltage output characteristics of the photovoltaic power generation unit, the reference operating voltage corresponding to the photovoltaic power generation unit operating at its maximum power point is calculated, thus completing the location of the optimal operating point on the photovoltaic side. After obtaining the reference operating voltage, the system combines the power balance relationship of the circuit with the current second output current to convert the voltage target into a current control target, and derives the reference operating current of the inductor under the corresponding operating condition.

[0166] Finally, the current first input voltage and first output voltage are used as the actual operating condition boundaries, and the reference operating voltage and reference operating current are used as the control tracking targets. All four parameters are input into the preset prediction model. Based on the circuit's operating characteristics and optimization objectives, the model evaluates and optimizes all feasible switching states, and finally outputs the final switching state prediction results, which are used to drive the switching unit of the corresponding branch in the next control cycle.

[0167] Thus, in this embodiment, the reference operating voltage of the buck-boost circuit and the reference operating current of the inductor when the photovoltaic power generation unit operates at maximum power can be determined based on the first input voltage, the first input current, and the second output current. The first input voltage, the first output voltage, the reference operating voltage, and the reference operating current are then input to a preset prediction model to obtain the switching state prediction result. This achieves the combination of maximum power point tracking and model predictive control, avoids the control deviation and delay caused by multi-layer control loop switching, and effectively improves the energy utilization efficiency of the photovoltaic power generation unit.

[0168] Meanwhile, the reference operating voltage and reference operating current are dynamically generated based on the real-time electrical parameters of the current cycle. This enables rapid response to fluctuations in photovoltaic characteristics caused by changes in the environment such as light and temperature. The reference values ​​are updated in a timely manner, and the tracking speed is faster under dynamic operating conditions. It can always lock the maximum power point and reduce energy loss during changes in operating conditions.

[0169] Furthermore, by inputting the actual input and output voltages along with reference values ​​into the prediction model, the prediction calculations can closely match the real-time operating boundaries of the circuit, which improves the accuracy of switching state prediction to a certain extent, ensuring the stability of circuit operation and the precision of current control. Combined with the interleaved parallel circuit topology, the output current ripple can be further reduced, switching losses can be reduced, and the overall conversion efficiency of the system can be improved.

[0170] In some embodiments provided in this application, the steps of determining the reference operating voltage of the buck-boost circuit and the reference operating current of the inductor when the photovoltaic power generation unit operates at maximum power, based on the first input voltage, the first input current, and the second output current, include: Based on the first input voltage and the first input current, determine the reference operating voltage of the photovoltaic power generation unit when it operates at maximum power; The reference operating current is determined based on the voltage difference between the first input voltage and the reference operating voltage.

[0171] The model prediction module in this application embodiment is also used to determine the reference operating voltage of the photovoltaic power generation unit when the photovoltaic power generation unit operates at maximum power based on the first input voltage and the first input current, and to determine the reference operating current based on the voltage difference between the first input voltage and the reference operating voltage.

[0172] The processor in this application embodiment is further configured to determine, based on the first input voltage and the first input current, a reference operating voltage of the photovoltaic power generation unit when the photovoltaic power generation unit operates at maximum power, and to determine a reference operating current based on the voltage difference between the first input voltage and the reference operating voltage.

[0173] Specifically, to ensure the reliability of the reference operating voltage and reference operating current, in some embodiments provided in this application, the photovoltaic power generation system can first locate the reference operating voltage corresponding to the global maximum power point of the photovoltaic power generation unit based on the real-time collected input voltage and input current, and then derive the reference operating current of the inductor branch through the difference between the current input voltage and the reference operating voltage.

[0174] To more clearly illustrate the generation process of the reference operating voltage and reference operating current in the embodiments of this application, please refer to the following exemplary description: Within each control cycle, the photovoltaic power generation unit first acquires the first input voltage and the first input current in real time. Combining the inherent power-voltage output characteristics of the photovoltaic unit, the power peak point under the current operating condition is located through the maximum power point tracking algorithm. The voltage value corresponding to the global maximum power point is then selected and determined as the reference operating voltage of the photovoltaic power generation unit under the current operating condition.

[0175] After obtaining the reference operating voltage, the photovoltaic (PV) power generation system calculates the voltage difference between the current first input voltage and the reference operating voltage. The sign of the difference determines the direction of the current operating point's deviation, and the magnitude of the difference determines the degree of deviation. Subsequently, the PV power generation system performs closed-loop regulation calculations based on this voltage difference to obtain the reference operating current for the inductor branch. Specifically, when the actual input voltage is higher than the reference operating voltage, a reference current of corresponding magnitude is generated to increase the inductor current, thereby lowering the input voltage; when the actual input voltage is lower than the reference operating voltage, a reference current of corresponding magnitude is generated to decrease the inductor current, thereby raising the input voltage, ultimately driving the PV input voltage to gradually converge to the reference operating voltage.

[0176] The final generated reference operating voltage and reference operating current will be input into the preset prediction model as the target benchmark for switching state prediction, ensuring that the subsequent optimization process of switching state always revolves around the core objective of maximum power output, and completing the complete transmission process from power optimization to current benchmark.

[0177] Thus, in this embodiment, the reference operating voltage of the photovoltaic power generation unit when it operates at maximum power can be determined based on the first input voltage and the first input current, and the reference operating current can be determined based on the voltage difference between the first input voltage and the reference operating voltage. This can, to a certain extent, ensure the validity of the reference operating voltage and reference operating current, and avoid the cumulative error caused by intermediate power conversion, thereby improving the tracking accuracy of the maximum power point. Furthermore, the method of deriving the reference operating current based on the voltage difference has a simple and clear computational logic, eliminating the need for complex multi-parameter iterations and multi-loop control parameter tuning, reducing the computational load, and enabling adaptation to higher frequency control cycles.

[0178] In some embodiments provided in this application, the step of determining the reference operating voltage of the photovoltaic power generation unit when it operates at maximum power based on the first input voltage and the first input current includes: Based on the first input voltage and the preset power-voltage curve of the photovoltaic power generation unit, determine the target power peak point on the preset power-voltage curve that is closest to the first input voltage; Gradually increase the first input voltage; Based on the power and voltage corresponding to the target power peak point, and the first input voltage and first input current before and after each increase in the first input voltage, determine the target value of the first input voltage when the photovoltaic power generation unit operates at maximum power; The target value is set as the reference operating voltage.

[0179] The model prediction module in this application embodiment is further configured to determine the target power peak point on the preset power-voltage curve that is closest to the first input voltage based on the first input voltage and the preset power-voltage curve of the photovoltaic power generation unit, and gradually increase the first input voltage, and determine the target value of the first input voltage when the photovoltaic power generation unit operates at maximum power based on the power and voltage corresponding to the target power peak point, as well as the first input voltage and first input current before and after each increase of the first input voltage, and determine the target value as the reference operating voltage.

[0180] The processor in this embodiment is further configured to determine, based on the first input voltage and the preset power-voltage curve of the photovoltaic power generation unit, the target power peak point closest to the first input voltage on the preset power-voltage curve, and gradually increase the first input voltage, and based on the power and voltage corresponding to the target power peak point, as well as the first input voltage and first input current before and after each increase of the first input voltage, determine the target value of the first input voltage when the photovoltaic power generation unit operates at maximum power, and determine the target value as the reference operating voltage.

[0181] Specifically, during the actual operation of a photovoltaic power generation system, the photovoltaic power generation unit is often affected by factors such as cloud cover, building projection, and surrounding shadows, resulting in uneven local illumination. At this time, the power-voltage characteristic curve of the photovoltaic power generation unit will form multiple local power peaks, with only one global maximum power point.

[0182] For a clearer explanation of this situation, please refer to [link / reference]. Figure 4a , Figure 4b , Figure 5a and Figure 5b , Figure 4a This is a schematic diagram of the current-voltage output characteristic curves of a photovoltaic power generation unit under uniform illumination in certain embodiments of this application. Figure 4a This is a schematic diagram of the power-voltage output characteristic curves of a photovoltaic power generation unit under uniform illumination in certain embodiments of this application. Figure 5a This is a schematic diagram of the current-voltage output characteristic curves of a photovoltaic power generation unit under non-uniform illumination in certain embodiments of this application. Figure 5b This is a schematic diagram of the power-voltage output characteristic curves of a photovoltaic power generation unit under non-uniform illumination in certain embodiments of this application.

[0183] It should be noted that, Figure 4a and Figure 4b The diagram shows the output characteristic curves of a photovoltaic power generation system comprising six photovoltaic modules (i.e., photovoltaic power generation units), with all six modules receiving the same irradiation conditions. The irradiation conditions include: all six photovoltaic modules receiving 500 W / m². 2 The irradiance was such that all six photovoltaic modules received 1000W / m².2 The irradiance was 2000 W / m² for all six photovoltaic modules. 2 The intensity of radiation.

[0184] It should also be noted that, Figure 5a and Figure 5b The diagram shows the output characteristic curves of a photovoltaic power generation system comprising six photovoltaic modules (i.e., photovoltaic power generation units), with the six photovoltaic modules not having identical irradiation conditions. The irradiation conditions include: each of the six photovoltaic modules receiving an irradiance of 800 W / m². 2 800 W / m 2 800W / m 2 400 W / m 2 400 W / m 2 400 W / m 2 The irradiance received by each of the six photovoltaic modules was 1200 W / m². 2 1200W / m 2 800 W / m 2 800 W / m 2 500 W / m 2 500 W / m 2 The irradiance received by each of the six photovoltaic modules was 1200W / m². 2 1200 W / m 2 400 W / m 2 400 W / m 2 200 W / m 2 200 W / m 2 Furthermore, based on Figure 4a and Figure 4b It can be seen that the power-voltage characteristic curve of a photovoltaic power generation unit under uniform irradiation conditions has a single peak, that is... Figure 4b The location marked by the dashed rectangle is thus easily tracked by traditional MPPT algorithms such as the perturbation-observation method and the incremental conductance method. However, based on Figure 5a and Figure 5b It can be seen that under non-uniform lighting scenarios such as partial shading, the PV curve of a photovoltaic power generation unit has a global maximum power point (GMPP), that is... Figure 5b The area marked by the dashed circular box also includes multiple local maximum power points (LMPPs). The presence of multiple LMPPs makes the traditional MPPT algorithm prone to getting stuck in LMPPs, thus reducing the power generation efficiency of the photovoltaic power generation unit.

[0185] Based on this, in some embodiments provided in this application, based on the pre-calibrated power-voltage curve of the photovoltaic power generation unit, the power peak point closest to the current operating voltage is taken as the benchmark, and then the operating voltage is gradually increased. The power trend is judged by combining the parameters of the benchmark peak point with the real-time sampled voltage and current data. Finally, the voltage target value corresponding to the global maximum power point is locked as the reference operating voltage. In this way, it is possible to accurately jump out of the local peak and find the global optimal operating point in the multi-peak scenario of partial shading. It is also possible to compress the search interval by using the prior peak point information and shorten the tracking time of the maximum power point.

[0186] In some implementations, the preset power-voltage curve can also be called the photovoltaic power generation curve. It can be understood as a pre-calibrated characteristic curve that reflects the relationship between the output power and output voltage of the photovoltaic power generation unit. The horizontal axis of the curve is the output voltage of the photovoltaic unit, and the vertical axis is the output power under the corresponding voltage. The shape of the curve will change with the working conditions such as light intensity and ambient temperature. The peaks on the curve correspond to the power extreme points under different working conditions.

[0187] In some examples, power-voltage curve sets under different light levels and ambient temperatures can be obtained through preliminary calibration tests. When the system is running, the power-voltage curves under the corresponding conditions can be called according to the real-time collected operating parameters to ensure the accuracy of peak point location.

[0188] In some examples, the preset power-voltage curve may be as described above. Figure 5b and Figure 4b .

[0189] In some implementations, the target power peak can be understood as the power extreme point on a preset power-voltage curve, i.e., the power value at this point is higher than the power values ​​of its adjacent voltage ranges. In partial shading scenarios, there will be multiple power peaks on the curve, among which the one with the largest power value is the global maximum power point, and the rest are local power peaks, i.e. Figure 4b As shown.

[0190] In some implementations, the target value can be understood as the first input voltage value corresponding to the photovoltaic power generation unit operating at the global maximum power point, which is finally determined after search calculation.

[0191] In some examples, after completing the search and judgment across the entire range, the voltage value corresponding to the operating point with the highest output power in the global range can be determined as the target value, providing a precise voltage reference for the control of subsequent buck and boost circuits.

[0192] To more clearly illustrate the process of determining the reference operating voltage in the embodiments of this application, please refer to the following exemplary description: When determining the reference operating voltage of a photovoltaic power generation unit, the photovoltaic power generation system first obtains the first input voltage sampled in the current control cycle, retrieves the preset power-voltage curve corresponding to the photovoltaic power generation unit, matches and finds the power peak point closest to the current first input voltage value on the curve, determines this point as the target power peak point, and records the power value and voltage value corresponding to the target power peak point.

[0193] After the benchmark peak point is located, the photovoltaic power generation system gradually increases the value of the first input voltage in a preset fixed step size. After each voltage adjustment, the adjusted first input voltage and first input current are collected synchronously, the output power at the corresponding operating point is calculated, and the trend of power change with voltage is identified.

[0194] As the voltage gradually increases, the photovoltaic power generation system combines the power and voltage parameters of the target power peak point, as well as the first input voltage and first input current data collected before and after each voltage increase, to segment and judge the changing trend of the power curve, and identify whether there is a higher power peak point in the subsequent voltage range. If it is determined that there is a higher peak point, the search continues in the direction of higher voltage. If it is determined that there is no higher peak point, the search is terminated, the position of the global maximum power point is locked, and finally the target value of the first input voltage when the photovoltaic power generation unit operates at maximum power is calculated.

[0195] After the target value is calculated, the photovoltaic power generation system determines the target voltage value as the reference operating voltage of the photovoltaic power generation unit and outputs it to the subsequent model predictive control link as a reference quantity for controlling the buck-boost circuit and achieving maximum power point tracking.

[0196] Thus, in this embodiment, based on the first input voltage and the preset power-voltage curve of the photovoltaic power generation unit, the target power peak point closest to the first input voltage on the preset power-voltage curve can be determined, and the first input voltage can be gradually increased. Based on the power and voltage corresponding to the target power peak point, as well as the first input voltage and first input current before and after each increase in the first input voltage, the target value of the first input voltage when the photovoltaic power generation unit operates at maximum power can be determined, and the target value can be determined as the reference operating voltage. This can adapt to complex lighting scenarios with partial shading, and accurately break free from the limitation of local power peak points when there are multiple peaks in the power-voltage curve, locate the operating voltage corresponding to the global maximum power point, avoid the situation where the traditional single-peak algorithm gets stuck in local optima, and improve the overall power generation efficiency of the system.

[0197] Meanwhile, by first locating the nearest power peak as a benchmark before conducting a targeted search, it is not necessary to traverse the entire range starting from the lowest voltage, effectively compressing the search range. Combined with the power trend judgment during the gradual voltage increase process, it can also directly skip voltage ranges without higher peaks, significantly shortening the tracking time of the global maximum power point, improving the response speed of maximum power point tracking control, and better adapting to operating scenarios with rapidly changing light intensity.

[0198] In some embodiments provided in this application, the step of determining the target power peak point on the preset power-voltage curve that is closest to the first input voltage based on the first input voltage and the preset power-voltage curve of the photovoltaic power generation unit includes: The target starting voltage value is determined based on the illumination status of the photovoltaic power generation unit; Control the opening and closing state of the switching unit in the buck-boost circuit so that the value of the first input voltage changes to the target starting voltage value; Based on the first input voltage and the preset power-voltage curve, determine the target power peak point on the preset power-voltage curve that is closest to the first input voltage.

[0199] The model prediction module in this application embodiment is also used to determine the target starting voltage value based on the illumination state of the photovoltaic power generation unit, and to control the opening and closing state of the switching unit in the buck-boost circuit so that the value of the first input voltage changes to the target starting voltage value, and to determine the target power peak point on the preset power-voltage curve that is closest to the first input voltage based on the first input voltage and the preset power-voltage curve.

[0200] The processor in this application embodiment is further configured to determine the target starting voltage value based on the illumination state of the photovoltaic power generation unit, and control the opening and closing state of the switching unit in the buck-boost circuit so that the value of the first input voltage changes to the target starting voltage value, and determine the target power peak point on the preset power-voltage curve that is closest to the first input voltage based on the first input voltage and the preset power-voltage curve.

[0201] Specifically, in the maximum power point tracking (MPPT) process of a photovoltaic (PV) power generation system, if the current real-time input voltage is used directly as the search starting point when locating the power peak, the MPPT tracking efficiency will be affected to some extent. Specifically, in a uniform illumination scenario, if the current input voltage deviates from the optimal search range, it will lead to redundancy in the single-peak search path, lengthening the MPPT tracking time. In a multi-peak scenario with partial shading, an unreasonable starting voltage can easily cause some power peaks to be missed during the search process, or a local peak to be mistakenly identified as the target reference peak, directly affecting the accuracy of the subsequent global MPPT search and ultimately reducing the power generation efficiency of the PV power generation unit.

[0202] Based on this, in some embodiments provided in this application, when locating the target power peak on the preset power-voltage curve, a light state recognition and starting voltage calibration mechanism is introduced. First, the light state of the photovoltaic power generation unit is detected, and the target starting voltage value under the corresponding scenario is matched. The first input voltage is adjusted to the starting value by controlling the switching state of the buck-boost circuit. Then, based on the calibrated input voltage, the nearest target power peak on the power-voltage curve is located. This adapts to different operating scenarios such as uniform illumination and partial shading, optimizes the search starting position of the power peak, and improves search efficiency and positioning accuracy.

[0203] In some implementations, the illumination state can be understood as the uniformity of the photovoltaic power generation unit being irradiated by light. The number of peaks and voltage distribution range of the curves differ significantly under different states.

[0204] In some examples, the illumination conditions can generally be divided into uniform illumination conditions and non-uniform illumination conditions. Non-uniform illumination conditions can also be called local shading conditions, which can be determined by the output current and voltage characteristics of the photovoltaic array, or by the detection data of the matching illumination sensor.

[0205] In some implementations, the target starting voltage value can be understood as the initial adjustment target value of the first input voltage preset when starting the power peak point search process, which serves as the starting reference for subsequent power peak point searches. The value corresponds to the illumination state and is used to ensure that the search starting point is within a reasonable voltage range.

[0206] In some examples, corresponding voltage values ​​can be preset for different lighting conditions. For example, in a uniform lighting scenario, 60% of the open-circuit voltage of the entire photovoltaic array can be set as the target starting voltage value; in a partial shading scenario, 60% of the open-circuit voltage of a single photovoltaic module can be set as the target starting voltage value to adapt to the peak distribution characteristics under different scenarios.

[0207] To more clearly illustrate the positioning process of the maximum power point tracking process in the embodiments of this application, please refer to [link to relevant documentation]. Figure 6 , Figure 6 This is a flowchart illustrating a control method for a photovoltaic power generation system in certain embodiments of this application. Specifically, as shown... Figure 6 As shown, in the process of locating maximum power point tracking in the photovoltaic system according to the embodiments of this application, an initialization operation is first performed to complete the initial configuration of system parameters and operating variables, such as setting the reference voltage V. ref Set as the open-circuit voltage V of a single photovoltaic module oc-mod 60%, that is, V ref =60%V oc-mod At the same time, the segmented counting variable j is set to 1, i.e., j=1.

[0208] Then the incremental conductance method was used to update P.m V m That is, by using the incremental conductance method, the local power peak point is searched within the current voltage range, and after locating the peak point, the maximum power P corresponding to that point is recorded. m With corresponding operating voltage V m .

[0209] After completing the first peak record, the algorithm gradually increases the reference voltage in fixed steps, executing V... ref =V ref +ΔV, continuously scanning towards higher voltage ranges. Then V ref Is it greater than V? end That is, to determine "V" ref >V end Is the statement true or false? Where V is... end This is the preset upper limit of the voltage search; if the judgment result is yes (Y), it means that the full voltage range scan has been completed, and we can directly enter the final output stage, that is, "when V_pv=V m At that time, "GMPP" is obtained, which is the input voltage V of the photovoltaic array. pv Operating at the recorded peak voltage V m This allows us to pinpoint the global maximum power point.

[0210] If "V" ref >V end If the result of the judgment is negative (N), then continue to judge the power change trend, that is, judge whether the rate of change of power with voltage dP / dV is less than 0, that is, judge whether "dP / dV<0?" is true. If the result of the judgment is positive (Y), it means that we are still in the power decrease range of the current peak and have not yet reached the dividing point between the two peaks, so return to "V ref =V ref The "+ΔV" step means continuing to increase the reference voltage in fixed steps and continuously scanning backward.

[0211] If the result of the judgment "dP / dV<0?" is negative (N), it means that dP / dV has changed from negative to positive, and the current position is the segmentation point between the two peak intervals. The algorithm then enters the jump phase. Specifically, first, the segmentation point current I_sc(j+1) is obtained, that is, the current value corresponding to the (j+1)th segmentation point is recorded; then, the formula V is used... ref =P m / I_sc(j+1) calculates the new reference voltage and jumps directly to that voltage position, skipping the voltage interval between the segment point and the new operating point that does not need to be searched, thereby shortening the overall search time; after completing the voltage jump, execute j=j+1 to increment the segment count variable by 1 and enter the judgment process of the next segment.

[0212] After updating the count j, a second voltage boundary check is performed, that is, the "V" condition is checked again.ref >V end Is the condition true? If the result is yes (Y), proceed to the final step, i.e., "when V_pv=V m When GMPP is obtained, the global maximum power point GMPP is output.

[0213] If we judge "V" again ref >V end If the result of the judgment for "?" is negative (N), then the process proceeds to the third stage of the judgment process to determine whether the next peak value has search value. Specifically, the judgment for "I" is... pv Is the condition >0.85I_sc(j) and dp / dv>0 true? That is, simultaneously verify the current photovoltaic output current I. pv Is it greater than 0.85 times the current at the current segment point I_sc(j), and is the power change rate dp / dv greater than 0?

[0214] If "I pv The judgment result for ">0.85I_sc(j) and dp / dv>0?" is negative (N), indicating that the power of the next peak is less than the currently recorded P. m Since it is only a local maximum power point, no additional search is needed, so we return to "V". ref =V ref The "+ΔV" step continues scanning backward.

[0215] If "I pv The judgment result for ">0.85I_sc(j) and dp / dv>0?" is yes (Y), indicating that the power of the next peak is greater than the currently recorded P. m This is a better power peak point, therefore we return to "starting with the incremental conductance method and updating P". m V m "This step involves performing an incremental conductance search on the new peak range to update the recorded maximum power P." m With the corresponding voltage V m This will initiate a new round of scanning and assessment.

[0216] After the algorithm completes the full-range scan and obtains the Global Maximum Power Point (GMPP), it will continuously monitor the system's operating status to determine whether the restart conditions are met. Figure 6 The question asks for "Restart conditions?". Specifically, if the restart conditions are met, the system returns to "V". ref =60%V oc-mod The step of "j=1" restarts global maximum power point tracking to adapt to dynamic changes in operating conditions such as illumination and temperature; if the restart condition is not met, the current global maximum power point operation state is maintained and the maximum power is continuously output.

[0217] Thus, in this embodiment, the target starting voltage value can be determined according to the illumination state of the photovoltaic power generation unit, and the opening and closing state of the switching unit in the buck-boost circuit can be controlled so that the value of the first input voltage changes to the target starting voltage value. According to the first input voltage and the preset power-voltage curve, the target power peak point on the preset power-voltage curve that is closest to the first input voltage can be determined. This enables adaptive switching of the power peak point search starting point, matching the corresponding starting voltage for different illumination states, and improving the peak point search efficiency under different operating conditions to a certain extent.

[0218] In some embodiments provided in this application, the photovoltaic power generation system further includes a bidirectional conversion circuit, and the buck-boost circuit and power supply network are electrically connected to the target load through the bidirectional conversion circuit. Furthermore, the control method of the photovoltaic power generation system in the embodiments of this application also includes: Control the operating state of the bidirectional conversion circuit so that the electrical energy output by the photovoltaic power generation unit and the electrical energy output by the power supply network flows to the target load, or so that the electrical energy output by the target load flows to the power supply network.

[0219] The control module in the photovoltaic power generation system provided in this application embodiment is also used to control the working state of the bidirectional conversion circuit so that the electrical energy output by the photovoltaic power generation unit and the electrical energy output by the power supply network flows to the target load, or so that the electrical energy output by the target load flows to the power supply network.

[0220] The processor in this embodiment is also used to control the operating state of the bidirectional conversion circuit so that the electrical energy output by the photovoltaic power generation unit and the electrical energy output by the power supply network flows to the target load, or so that the electrical energy output by the target load flows to the power supply network.

[0221] Specifically, traditional residential photovoltaic (PV) power generation systems have a relatively singular energy flow direction, typically supporting only one-way power supply from PV units to the load, or only one-way grid connection of PV power to the grid. They lack flexible two-way energy dispatch capabilities and cannot adapt to the grid's peak-valley pricing mechanism for energy off-peak allocation. During the day, when PV output is excessive, it cannot efficiently feed back to the grid to generate revenue. At night, when electricity prices are high during peak hours, it cannot utilize energy storage resources on the load side for power supply, resulting in insufficient system operating economy. Furthermore, the singular energy flow direction makes it difficult to cope with emergency scenarios such as grid outages, limiting power supply reliability. Voltage surges are easily generated when switching between multiple power sources and loads, affecting the stable operation of the load.

[0222] Based on this, in some embodiments provided in this application, a bidirectional conversion circuit can be added to the photovoltaic power generation system to electrically connect the buck-boost circuit, the power supply network, and the target load. The energy transmission direction can be switched by adjusting the operating state of the bidirectional conversion circuit. This allows the photovoltaic power generation unit and the power supply network to jointly supply power to the target load, ensuring a stable power supply to the load; it also supports the target load to output power back to the power supply network, realizing functions such as energy storage peak-shaving and grid-connected power sales, thereby improving the system's operational economy and energy utilization rate.

[0223] In some implementations, a bidirectional conversion circuit can be understood as a power conversion circuit with bidirectional power transmission capability. It can flexibly switch the direction of energy transmission according to control commands, and can also realize functions such as voltage level conversion and electrical isolation. It is the core power unit for flexible energy scheduling in multi-power supply and multi-load systems.

[0224] In some examples, the bidirectional converter circuit can adopt a bidirectional CLLC resonant topology, which can realize soft switching of switching devices under all operating conditions, with high conversion efficiency, low electromagnetic interference, and adaptability to the operating conditions of bidirectional charging and discharging. It can efficiently complete the bidirectional buck-boost conversion of DC power.

[0225] In some implementations, the power supply network can be understood as a public mains power supply network that can provide stable power frequency energy and support the purchase and sale of electricity.

[0226] In some examples, the power supply network is a residential-side industrial frequency AC grid network with a peak-valley pricing mechanism, where electricity prices are higher during peak hours and lower during off-peak hours, providing an economical operating basis for the system's peak-shifting energy dispatch.

[0227] In some implementations, the target load can be understood as the electrical energy interaction object of the photovoltaic power generation system, including both power-consuming loads and energy storage loads with charging and discharging capabilities. It can receive electrical energy to consume electricity or charge energy storage, and can also output electrical energy under specific operating conditions.

[0228] In some examples, the target load may include household daily low-voltage electrical loads, residential energy storage batteries, electric vehicle power batteries, etc., where energy storage loads can participate in the system's energy dispatch as mobile energy storage units.

[0229] To more clearly illustrate the implementation process of bidirectional energy dispatch in the embodiments of this application, please refer to... Figure 2 And the following exemplary description, namely: When vehicle 304 (i.e., the target load) has a power demand, such as when a user sends a charging command to vehicle 304 via cloud application 305, the photovoltaic power generation system controls the bidirectional conversion circuit 270 to operate in the forward power supply state. For example, if the output of photovoltaic power generation unit 210 is sufficient, and the electrical energy output by photovoltaic power generation unit 210 undergoes voltage conversion and maximum power point adjustment via buck-boost circuit 220, it is then delivered to vehicle 304 through energy storage DC-DC circuit 240 and bidirectional conversion circuit 270 to meet the power demand of vehicle 304. Alternatively, if the output of photovoltaic power generation unit 210 is insufficient to cover the load demand, the electrical energy from power supply network 302 is synchronously connected to bidirectional conversion circuit 270 via inverter 260, working together with photovoltaic power generation unit 210 to supply power to vehicle 304, ensuring the continuity and stability of power supply to vehicle 304.

[0230] When the vehicle 304 has sufficient electrical energy and there is a need to feed back to the power supply network 302, such as during peak nighttime electricity consumption when electricity prices are high and energy storage loads have surplus power, the photovoltaic power generation system controls the bidirectional conversion circuit 270 to switch to reverse feedback mode. At the same time, the user sends corresponding control commands to the vehicle 304 through the cloud application 305, so that the electrical energy output by the vehicle 304 is converted into voltage and system by the bidirectional conversion circuit 270 and the inverter 260 before flowing to the power supply network 302, realizing grid-connected electricity sales and converting stored electrical energy into economic benefits. In emergency scenarios where the power supply network 302 is shut down, the bidirectional conversion circuit 270 can also be used to dispatch the energy stored in the vehicle 304 to continuously supply power to critical loads and improve the power supply reliability of the photovoltaic power generation system.

[0231] Thus, in this embodiment, the operating state of the bidirectional conversion circuit can be controlled to allow the electrical energy output from the photovoltaic power generation unit and the power supply network to flow to the target load, or to allow the electrical energy output from the target load to flow to the power supply network. This realizes the transmission of photovoltaic and grid power to the target load, as well as the feedback of target load power to the power supply network, thereby improving the operational economy of the photovoltaic power generation system to a certain extent. Furthermore, it can improve the system's energy utilization rate and scenario adaptability. It can fully absorb the output of photovoltaic power generation units and avoid wasting solar energy. It can also support bidirectional charging and discharging of energy storage loads, adapting to various scenarios such as daily household electricity use, electric vehicle charging and discharging, and emergency backup power supply, thus expanding the functional boundaries of residential photovoltaic power generation systems.

[0232] Furthermore, for a clearer explanation of the control method for the photovoltaic power generation system provided in this application, please refer to [link to relevant documentation]. Figure 6 and Figure 7 , Figure 7 This is a schematic diagram illustrating an application scenario of the control method for a photovoltaic power generation system in certain embodiments of this application. Specifically, as shown... Figure 7As shown, the input current I is first sampled from the output side of the photovoltaic power generation unit. pv With input voltage V pv The two sets of operating parameters are sent to the maximum power point tracking module. The maximum power point tracking module is based on... Figure 6 The maximum power point tracking method shown is used to calculate the reference operating voltage V corresponding to the photovoltaic power generation unit operating at the maximum power point. Ref .

[0233] Subsequently, the reference operating voltage V Ref With the feedback input voltage V pv Both are connected to the voltage error comparison node, and their deviation is input to the voltage loop proportional-integral (PI) module. The PI module performs closed-loop regulation calculations on the voltage deviation and outputs a reference operating current I that matches the target transmission power. ref .

[0234] Because the buck-boost circuit uses a two-phase interleaved parallel topology, the total reference operating current I... ref After multiplying by 1 / 2 and dividing equally, we obtain the inductor current setpoint I corresponding to the two parallel branches. L1 *、I L2 *

[0235] Next, the photovoltaic power generation system will assign a setpoint I to the two inductor currents. L1 *、I L2 *, and the actual operating inductor current I of the two branches obtained through real-time sampling. L1 Inductor current I L2 and input voltage V pv Output voltage V out These parameters are used together as input parameters and sent to the model prediction and control module.

[0236] The model predictive control module is based on a preset predictive model. It combines all input reference values ​​and real-time sampling quantities, and selects the prediction results of the switching state combination for the next control cycle by constructing and solving the cost function. That is, the switching drive signals corresponding to S1 to S8 are applied to each switching unit of the buck-boost circuit, thereby completing the closed-loop control of the entire photovoltaic power generation system.

[0237] For a clearer explanation of the operation process of the photovoltaic power generation system in the embodiments of this application, please refer to [link / reference needed]. Figure 6 and Figure 7 Based on the above, refer to Figure 8 , Figure 8 This is a schematic diagram illustrating an application scenario of the control method for a photovoltaic power generation system in certain embodiments of this application. Specifically, as shown... Figure 8 As shown, in the real-time operation of the buck-boost circuit, the photovoltaic power generation system can collect the inductor current I during the current control cycle k.L1 (k) Inductor current I L2 (k) Input voltage V pv (k) Output voltage V out (k). These sampled data are transmitted in two paths: one path is input to the current delay compensation module, and the other path is input to the inductor current slope calculation module.

[0238] The current delay compensation module receives the sampled data of the current cycle and combines it with the predicted inductor current I for the next control cycle k+1. L1 (k+1), Inductor current I L2 (k+1) compensates for and corrects the current phase deviation caused by sampling operation delay and drive execution delay, eliminates the impact of control lag on control accuracy, and provides a more accurate state reference for the front-end prediction stage.

[0239] The inductor current slope calculation module receives the input voltage V for the current cycle. pv (k) and output voltage V out (k) Combining the topology characteristics of the buck-boost circuit with the inductor volt-second characteristics, the slope of the inductor current change in the two branches under different switching combinations is calculated, and two sets of slope parameters, M1~M3 and M4~M6, are output. Among them, M1~M3 correspond to the current slope of the first branch under different switching states, and M4~M6 correspond to the current slope of the second branch under different switching states, providing core calculation parameters for subsequent state prediction.

[0240] The current and duty cycle prediction module receives reference values ​​for the two branch currents from the system, namely the inductor current I. L1 *、I L2 * Simultaneously, the slope parameters M1~M3 and M4~M6 output from the inductor current slope calculation module are used to predict and calculate the branch current and duty cycle corresponding to all feasible switching states based on a preset prediction model, and the prediction parameter group D for the first branch is output respectively. i 、i、I i2 With the prediction parameter group D of the second branch j j, I j2 Where Di and Dj are the predicted duty cycle values ​​for the corresponding branches, i and j are the indicator values ​​for the switch status, and I... i2 I j2 This represents the predicted branch current under the corresponding switching state.

[0241] The value function optimization module receives two sets of prediction parameters D. i 、i、I i2 With D j j, I j2A value function is constructed with the control objectives of minimizing current tracking error, suppressing current ripple, and optimizing switching losses. All feasible switching state combinations are evaluated and sorted, and the switching state combination that optimizes the value function is selected. The corresponding optimal control parameter D is then output. i i and D j 、j.

[0242] The pulse width modulation signal generation module receives the optimal control parameters Di, i and Dj, j output by the value function optimization, converts the discrete switching states and duty cycle parameters into pulse drive signals that can directly drive power devices, and generates eight switching control signals from S1 to S8.

[0243] Finally, the S1~S8 drive signals are output to the buck-boost circuit to control the conduction and cutoff of the eight switching units in the buck-boost circuit, completing the power regulation of a single cycle; at the same time, the operating status of the buck-boost circuit is continuously sampled and enters the closed-loop regulation of the next control cycle, thereby realizing the stable and efficient operation of the entire photovoltaic charging system.

[0244] This application also provides a computer-readable storage medium storing a computer program that, when executed by one or more processors, implements the steps in the control method of the photovoltaic power generation system described in some of the above embodiments.

[0245] This application also provides a computer program product, including a computer program / instructions, which, when executed by a processor, implement the steps in the control method of the photovoltaic power generation system of some of the above embodiments.

[0246] It is understood that a computer program includes computer program code. Computer program code can be in the form of source code, object code, executable files, or some intermediate form. Computer-readable storage media can include: any entity or device capable of carrying computer program code, recording media, USB flash drives, portable hard drives, magnetic disks, optical disks, computer memory, read-only memory (ROM), random access memory (RAM), and software distribution media, etc.

[0247] In this specification, the terms "specifically," "furthermore," "particularly," "understandably," etc., refer to specific features, structures, materials, or characteristics described in connection with embodiments or examples that are included in at least one embodiment or example of this application. In this specification, the illustrative expressions of the above terms do not necessarily refer to the same embodiment or example. Furthermore, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in one or more embodiments or examples. Moreover, without contradiction, those skilled in the art can combine and integrate the different embodiments or examples described in this specification, as well as the features of different embodiments or examples.

[0248] Any process or method description in the flowchart or otherwise herein can be understood as representing a module, segment, or portion of executable request code comprising one or more steps for implementing a particular logical function or process, and the scope of the preferred embodiments of this application includes additional implementations in which functions may be performed not in the order shown or discussed, including substantially simultaneously or in reverse order according to the functions involved, as should be understood by those skilled in the art to which embodiments of this application pertain.

[0249] Although embodiments of this application have been shown and described above, it is understood that the above embodiments are exemplary and should not be construed as limiting this application. Those skilled in the art can make changes, modifications, substitutions and variations to the above embodiments within the scope of this application.

Claims

1. A control method for a photovoltaic power generation system, characterized in that, The photovoltaic power generation system includes photovoltaic power generation units installed on the outer side of the building roof and a buck-boost circuit. The electrical energy generated by the photovoltaic power generation units is output to the buck-boost circuit. The buck-boost circuit includes two buck-boost branches that operate alternately and in parallel. Each buck-boost branch includes at least four switching units. The method includes: Obtain the current state parameters of the photovoltaic power generation system in the current control cycle; Based on the current state parameters and the preset prediction model, the switching state of the switching unit when the photovoltaic power generation unit operates at maximum power is predicted, and the switching state prediction result is obtained. When the next control cycle of the current control cycle arrives, a switching unit in one of the buck-boost branches is controlled according to the switching state prediction result.

2. The method according to claim 1, characterized in that, The buck-boost circuit includes a first branch and a second branch. The first branch includes a first switching unit, a second switching unit, a third switching unit, a fourth switching unit, and a first inductor. The second branch includes a fifth switching unit, a sixth switching unit, a seventh switching unit, an eighth switching unit, and a second inductor. The first terminal of the first switching unit is electrically connected to the first output terminal of the photovoltaic power generation unit. The second terminal of the first switching unit and the first terminal of the second switching unit are both electrically connected to the first terminal of the first inductor. The second terminal of the second switching unit is electrically connected to a preset busbar. The second terminal of the first inductor is electrically connected to the second terminal of the third switching unit and the first terminal of the fourth switching unit. The first terminal of the third switching unit is electrically connected to the first output terminal of the buck-boost circuit. The second terminal of the fourth switching unit is electrically connected to the preset busbar. The preset busbar is formed by electrically connecting the second output terminal of the photovoltaic power generation unit and the second output terminal of the buck-boost circuit. The first end of the fifth switch unit is electrically connected to the first output end of the photovoltaic power generation unit. The second end of the fifth switch unit and the first end of the sixth switch unit are both electrically connected to the first end of the second inductor. The second end of the sixth switch unit is electrically connected to the preset bus. The second end of the second inductor is electrically connected to the second end of the seventh switch unit and the first end of the eighth switch unit. The first end of the seventh switch unit is electrically connected to the first output end of the buck-boost circuit. The second end of the eighth switch unit is electrically connected to the preset bus.

3. The method according to claim 2, characterized in that, The step of predicting the switching state of the switching unit when the photovoltaic power generation unit operates at maximum power based on the current state parameters and a preset prediction model, and obtaining the switching state prediction result, includes: Based on the current state parameters and the preset prediction model, the switching states of the two switching units in one of the step-down and step-up branches are predicted when the photovoltaic power generation unit operates at maximum power, and the switching state prediction results are obtained.

4. The method according to claim 3, characterized in that, The current state parameters include at least the output current of the first inductor or the output current of the second inductor. The step of predicting the switching state of the switching unit when the photovoltaic power generation unit operates at maximum power, based on the current state parameters and a preset prediction model, to obtain the switching state prediction result includes: Based on the current state parameters and the preset prediction model, a cost function containing target constraints is constructed, wherein the target constraints are used to constrain the ripple in the output current and / or the duty cycle of the switching unit; The cost function is solved to predict the switching state of the switching unit when the photovoltaic power generation unit operates at maximum power, and the switching state prediction result is obtained.

5. The method according to claim 1, characterized in that, Each of the buck-boost branches also includes an inductor. The current state parameters include the first input voltage, first input current, and first output voltage of the buck-boost circuit in the current control cycle, and the second output current of the inductor in the current control cycle. Based on the current state parameters and a preset prediction model, the switching state of the switching unit when the photovoltaic power generation unit operates at maximum power is predicted to obtain the switching state prediction result, including: Based on the first input voltage, the first input current, and the second output current, determine the reference operating voltage of the buck-boost circuit and the reference operating current of the inductor when the photovoltaic power generation unit operates at maximum power. The first input voltage, the first output voltage, the reference operating voltage, and the reference operating current are input into the preset prediction model to obtain the switch state prediction result.

6. The method according to claim 5, characterized in that, The step of determining the reference operating voltage of the buck-boost circuit and the reference operating current of the inductor when the photovoltaic power generation unit operates at maximum power, based on the first input voltage, the first input current, and the second output current, includes: Based on the first input voltage and the first input current, determine the reference operating voltage of the photovoltaic power generation unit when it operates at maximum power; The reference operating current is determined based on the voltage difference between the first input voltage and the reference operating voltage.

7. The method according to claim 6, characterized in that, Determining the reference operating voltage of the photovoltaic power generation unit when it operates at maximum power based on the first input voltage and the first input current includes: Based on the first input voltage and the preset power-voltage curve of the photovoltaic power generation unit, determine the target power peak point on the preset power-voltage curve that is closest to the first input voltage; Gradually increase the first input voltage; Based on the power and voltage corresponding to the target power peak point, and the first input voltage and first input current before and after each increase of the first input voltage, determine the target value of the first input voltage when the photovoltaic power generation unit operates at maximum power; The target value is determined as the reference operating voltage.

8. The method according to claim 7, characterized in that, The step of determining the target power peak point on the preset power-voltage curve that is closest to the first input voltage based on the first input voltage and the preset power-voltage curve of the photovoltaic power generation unit includes: The target starting voltage value is determined based on the illumination status of the photovoltaic power generation unit; Control the opening and closing state of the switching unit in the buck-boost circuit so that the value of the first input voltage changes to the target starting voltage value; Based on the first input voltage and the preset power-voltage curve, determine the target power peak point on the preset power-voltage curve that is closest to the first input voltage.

9. The method according to claim 1, characterized in that, The photovoltaic power generation system further includes a bidirectional conversion circuit, and the buck-boost circuit and the power supply network are electrically connected to the target load through the bidirectional conversion circuit. The method further includes: The operating state of the bidirectional conversion circuit is controlled so that the electrical energy output by the photovoltaic power generation unit and the electrical energy output by the power supply network flows to the target load, or the electrical energy output by the target load flows to the power supply network.

10. A photovoltaic power generation system, characterized in that, The method includes a memory and a processor, wherein the memory stores a computer program, which, when executed by the processor, implements the method according to any one of claims 1-9.

Images