Operation mode seamless switching control method of photovoltaic-energy storage system

Abstract

The invention belongs to the technical field of optical storage operation control, and particularly relates to an operation mode seamless switching control method of a photovoltaic-energy storage system, which comprises the following steps of: for an optical storage system comprising an energy storage system and a photovoltaic system, performing seamless switching on the operation modes of the photovoltaic-energy storage system; obtaining a direct current bus voltage, a photovoltaic output voltage and current, an energy storage battery voltage and current, an LC filter inductance current and a grid-connected point voltage and current in the current optical storage system; aiming at an energy storage system, constructing a switching mechanism based on photovoltaic voltage feedforward; aiming at a photovoltaic system, constructing a dynamic control strategy based on photovoltaic power feed-forward compensation; based on a switching mechanism and a dynamic control strategy, network construction control is carried out by combining an inverter virtual synchronous generator control technology, and stable operation of the optical storage system in strong power grid and weak power grid environments is realized. The system is suitable for an optical storage system, can cover various typical operation modes, and has high responsiveness and high stability.

Description

Technical Field

[0001] This invention belongs to the field of photovoltaic-energy storage operation control technology, specifically relating to a seamless switching control method for the operating modes of a photovoltaic-energy storage system. Background Technology

[0002] With the rapid development of photovoltaic (PV) power generation and energy storage technologies, PV-energy storage systems have been widely applied in distributed generation, microgrids, and islanded systems. Compared to traditional power generation equipment, PV-energy storage systems offer advantages such as cleanliness, high efficiency, rapid response, and flexible regulation, significantly improving the utilization efficiency and power quality of renewable energy. Among the various control schemes for PV-energy storage systems, grid-forming (GFM) control schemes are gradually becoming a key focus of research and engineering applications due to their ability to provide voltage source support, frequency inertia support, and rapid grid-connection / off-grid switching capabilities.

[0003] Most current mainstream GFM control strategies rely on energy storage systems (such as batteries) as the core support source for inertia and regulation capabilities. The energy storage system and photovoltaic system are connected to the DC bus via a bidirectional DC-DC converter and a boost converter, and are connected to the grid through a grid-connected inverter. During system operation, to adapt to environmental changes and load dynamics, several typical operating modes exist, including: 1) Battery charging and discharging mode switching: The battery needs to adjust its charging and discharging according to changes in photovoltaic output and load at different times. The switching process places high demands on the system's power balance. 2) Battery state of charge (SOC) exceeding limits: When the battery is close to full charge (e.g., SOC > 80%), it will no longer be able to continue charging. The system needs to quickly change the power allocation logic to prevent excess energy. 3) Severe fluctuations in photovoltaic power: Affected by climate and solar radiation changes, photovoltaic output fluctuates frequently, posing a challenge to the system's power balance. 4) Sudden load changes or switching: Rapid changes on the load side may cause power imbalance, resulting in bus voltage fluctuations or even instability. 5) Switching between control modes: Such as switching the photovoltaic operating mode from maximum power point tracking (MPPT) mode to constant voltage (CV) mode.

[0004] However, existing control strategies are typically based on proportional-integral regulators, whose response speed is limited by the control bandwidth, making it difficult to achieve timely and stable regulation under the aforementioned multimodal abrupt change scenarios. Especially in situations such as energy storage failure or offline, sudden changes in PV (photovoltaic) power, or sudden load increases, the system is prone to bus voltage overshoot or drop, leading to instability or even triggering protection mechanisms, affecting the continuous operation of the system. Summary of the Invention

[0005] In view of the shortcomings of the prior art, the purpose of this invention is to provide a seamless switching control method for the operating modes of a photovoltaic-energy storage system, which is applicable to photovoltaic-energy storage systems, can cover a variety of typical operating modes and has high responsiveness and high stability.

[0006] To achieve the above objectives, this invention provides a seamless switching control method for the operating modes of a photovoltaic-energy storage system, comprising the following steps: S1. For a photovoltaic-storage system that includes an energy storage system and a photovoltaic system, obtain the DC bus voltage, photovoltaic output voltage and current, energy storage battery voltage and current, LC filter inductor current, and grid connection point voltage and current in the current photovoltaic-storage system. S2. For energy storage systems, a switching mechanism based on photovoltaic voltage feedforward is constructed to intervene in the adjustment process in advance and respond quickly to voltage changes; S3. For photovoltaic systems, a dynamic control strategy based on photovoltaic power feedforward compensation is constructed to achieve rapid suppression of transient fluctuations in DC bus voltage. S4. Based on the switching mechanism in S2 and the dynamic control strategy in S3, grid control is carried out in combination with inverter virtual synchronous generator control technology to achieve stable operation of the photovoltaic-storage system in both strong and weak grid environments.

[0007] As a preferred embodiment of the present invention, the switching mechanism based on photovoltaic voltage feedforward in S2 is specifically as follows: S2.1 Determine whether the photovoltaic system is operating normally, and issue an alarm signal when the photovoltaic system is not operating normally; S2.2. During normal operation of the photovoltaic system, a maximum power point tracking control strategy is adopted, and the output voltage is... After being regulated by a proportional-integral controller, it maximizes its own output power and outputs electrical energy to the DC bus to supply the load of the entire photovoltaic-storage system. S2.3 When the battery system of the energy storage system is detected to be offline or the SOC exceeds the set upper limit threshold, the energy storage converter of the energy storage system will no longer perform power exchange and will only maintain a safe standby state. At the same time, the photovoltaic system will switch from the maximum power point tracking control strategy to the constant voltage control mode. During the switching process, a switching mechanism based on photovoltaic voltage feedforward is adopted, that is, a dynamic voltage feedforward term is added to the output of the voltage loop PI regulator of the photovoltaic system. The control principle is as follows: ; In the formula, Indicates the duty cycle of the boost converter in a photovoltaic system; Represents the Laplace operator; This indicates the reference value for the DC bus voltage. Indicates the current DC bus voltage; This indicates the output voltage at the photovoltaic cell terminal; , These represent the proportional and integral coefficients of the voltage loop PI controller, respectively. The current mode is represented as: ; In the formula, Indicates the adjustment coefficient; MPPT represents the maximum power point tracking control strategy; CV represents the constant voltage control mode; This indicates the current duty cycle of the boost converter.

[0008] As a preferred embodiment of the present invention, the process for determining whether the photovoltaic system is operating normally in S2.1 is as follows: S2.1.1 Collect alarm signals from the physical alarm devices of the photovoltaic system to determine if there are any abnormal alarms. If there are any abnormal alarms, the photovoltaic system is not operating normally. If there are no abnormal alarms, proceed to S2.1.2. S2.1.2 Obtain the operating status data of the photovoltaic system, and perform data fusion based on the operating status data to obtain fused feature data; S2.1.3. Based on the fused feature data, perform LSTM-classifier prediction and output the probability of the photovoltaic system operating abnormally within a set future period. If the probability of operating abnormality is less than the set threshold, the photovoltaic system operates normally; if the probability of operating abnormality is not less than the set threshold, the photovoltaic system does not operate normally.

[0009] As a preferred embodiment of the present invention, the operating status data includes the junction temperature fluctuation amplitude of the inverter IGBT module. Voc-T curve fitting deviation of photovoltaic string Combiner box communication packet loss rate Inverter DC side input capacitor ripple current Maximum power point voltage drift of photovoltaic string Series resistance of photovoltaic modules , photovoltaic array output power fluctuation and the input voltage ripple coefficient of photovoltaic inverter The integrated feature data includes the health status of the inverter power unit. String characteristic deviation Component internal resistance aging coefficient Power output stability coefficient Monitoring communication integrity rate and System Comprehensive Degradation Index ; The process of fusing data based on operational status data to obtain fused feature data includes the following steps: right and The health status of the inverter power unit is obtained by weighting the values. ; right and The maximum value is taken to obtain the string characteristic deviation. ; right The minimum value is used to obtain the component's internal resistance aging coefficient. ; right and Perform multiplication to obtain the power output stability coefficient. ; based on The monitoring communication integrity rate was obtained. ; right , , , and The system's overall degradation index is obtained by performing fusion processing. .

[0010] As a preferred embodiment of the present invention, the health status of the inverter power unit The calculation formula is: ; This represents the upper limit of the normal threshold for IGBT junction temperature fluctuation. This represents the upper limit of the normal threshold for capacitor ripple current. and All are weighting factors; String characteristic deviation The calculation formula is: ; and They are respectively and The upper limit of the normal threshold; Component internal resistance aging coefficient The calculation formula is: ; This is the aging threshold for the series resistance. Power output stability coefficient The calculation formula is: ; and They are respectively and The normal threshold upper limit, when the product is greater than 1 Take 1; Monitoring communication integrity The calculation formula is: ; This is the upper limit of the normal packet loss rate threshold. If the value is less than 0, it is set to 0, indicating that communication has completely failed; System Comprehensive Degradation Index The calculation formula is: ; , , , and They are respectively , , , and The moving average over one hour. , , , and All are weighting factors.

[0011] As a preferred embodiment of the present invention, in S2.1.3, the step of predicting the probability of an operational anomaly of the photovoltaic system within a set future period based on LSTM-classifier prediction using fused feature data includes the following steps: S2.1.3.1. The fused feature data from multiple consecutive time points are recorded as a feature sequence; S2.1.3.2 The feature sequence is input into the first layer of the bidirectional long short-term memory network BiLSTM to extract short-term features and obtain local feature vectors. The local feature vectors are then input into the second layer of the bidirectional long short-term memory network BiLSTM to extract long-term features and obtain global feature vectors. S2.1.3.3 Obtain the weight factor of each feature vector in the global feature vector, and perform a weighted summation on the global feature vector based on the obtained weight factors to obtain a weighted compressed feature vector; S2.1.3.4 The weighted compressed feature vector is input into the first fully connected layer of the classifier to obtain intermediate features. The intermediate features are then input into the second fully connected layer of the classifier, and the sigmoid function is used to output the probability of abnormal operation.

[0012] As a preferred embodiment of the present invention, in S2.3, the process of switching the photovoltaic system from the maximum power point tracking control strategy to the constant voltage control mode is as follows: S2.3.1 Real-time detection of the SOC value of the energy storage battery to determine whether it has reached the set upper limit threshold, and at the same time, detection of whether the battery system is offline. S2.3.2 When SOC ≥ upper limit threshold, or when the battery system is offline / communication interrupted, the battery management system feeds back a charging prohibition signal, the photovoltaic system no longer maintains the maximum power point tracking control strategy, and the photovoltaic system controller issues a command to switch to constant voltage control mode. S2.3.3 The photovoltaic system switches to constant voltage control mode. In constant voltage control mode, the status of the photovoltaic system is continuously monitored. When the SOC drops to the set recovery threshold or the battery system comes back online, the photovoltaic system controller issues a command to switch to the maximum power point tracking control strategy and restores the maximum power point tracking control strategy.

[0013] As a preferred embodiment of the present invention, the upper limit threshold is set to 80%, and the recovery threshold is set to 70%.

[0014] As a preferred embodiment of the present invention, the dynamic control strategy based on photovoltaic power feedforward compensation in S3 is specifically as follows: S3.1 The energy storage converter adopts a voltage-current dual closed-loop control structure, consisting of an outer loop voltage controller and an inner loop current controller. The outer loop control objective is... ,contrast This generates the battery current command value. The inner loop uses current commands as the control target, and controls the energy storage converter to output the actual current through current loop control. ; S3.2, Based on the voltage-current dual closed-loop control structure, a control based on the output power of the photovoltaic array is introduced. The feedforward compensation term calculates the compensation current by sampling the photovoltaic output power in real time. ; In the formula, This indicates the photovoltaic power feedforward compensation current; Indicates the feedforward gain coefficient; S3.3, Total Current Reference Command for Energy Storage Batteries It is composed of the output of the outer loop voltage controller and the superposition of the feedforward compensation term: ; In the inner loop controller of the input energy storage system, the duty cycle or control voltage of the energy storage converter is adjusted.

[0015] As a preferred embodiment of the present invention, the specific method for network configuration control in step S4 is as follows: S4.1 The photovoltaic system is controlled through the switching mechanism in S2, and the DC bus voltage is regulated through the dynamic control strategy in S3. S4.2 For the inverter, a virtual synchronous generator control strategy is adopted to simulate the electromagnetic inertia and damping characteristics of a synchronous generator. Combined with active and reactive power control and inner loop output voltage regulation, the autonomous control of the inverter output voltage amplitude and phase is realized.

[0016] The beneficial effects of this invention are: This invention achieves seamless switching between multiple operating modes of a photovoltaic-storage system by introducing a photovoltaic power feedforward compensation control strategy and a photovoltaic voltage feedforward mechanism. When the battery's state of charge is saturated, it can automatically and smoothly switch the photovoltaic system from maximum power point tracking (MPPT) control to constant voltage (CV) control, significantly reducing bus voltage overshoot caused by sudden changes in the control target, effectively suppressing DC bus voltage fluctuations, and enhancing the stability and dynamic response speed of the photovoltaic-storage system under complex operating conditions.

[0017] This invention combines inverter virtual synchronous generator (VSG) control technology to simulate the inertia and damping characteristics of a synchronous generator, enabling the photovoltaic-storage system (PV-SES) to provide frequency and voltage support during grid-connected operation. It allows the system to proactively adapt to load or photovoltaic power disturbances, achieving dynamic adjustment of phase angle and voltage. This innovative control method not only enhances the power balance capability and grid adaptability of the PV-SES system but also ensures its stability and reliability in various complex scenarios such as weak grids and islanded operation, providing key technical support for the construction of highly adaptive and reliable PV-SES integrated systems.

[0018] This invention establishes steps for determining whether a photovoltaic (PV) system is operating normally. It employs a three-layer logic: physical alarm initial screening, multi-dimensional data fusion, and LSTM time-series prediction. Physical alarms quickly identify explicit faults, multi-dimensional data fusion (covering IGBT junction temperature, Voc-T curve deviation, etc.) characterizes implicit states, and time-series prediction provides early warning of faults. This solves the problems of missed and false judgments inherent in traditional single-judgment methods, reducing the risk of fault escalation. The judgment covers key dimensions such as equipment health, power generation performance, and communication quality, accurately assessing the system status and avoiding voltage fluctuations and equipment damage caused by potential PV system malfunctions (such as module aging). This ensures the stability of subsequent energy storage switching mechanisms and PV-storage network control. Timely alarms in case of anomalies help maintenance personnel quickly locate and handle issues, reducing power generation losses due to abnormal PV outages or improper operation, and improving the reliability and maintenance efficiency of the PV-storage system. Attached Figure Description

[0019] Figure 1 This is a flowchart illustrating the principle of this invention; Figure 2 This is a schematic diagram of the structure of the photovoltaic-storage grid-connected system with a common DC bus of the present invention; Figure 3 This is a schematic diagram of the photovoltaic maximum power point tracking control in this invention; Figure 4 This is a schematic diagram of the photovoltaic control strategy based on the photovoltaic voltage feedforward switching mechanism in this invention; Figure 5 This is a schematic diagram of the energy storage control strategy based on the photovoltaic voltage feedforward switching mechanism in this invention; Figure 6 This is a schematic diagram of the dynamic control strategy based on photovoltaic power feedforward compensation in this invention; Figure 7 This is a schematic diagram of the active power control loop of the inverter virtual synchronous machine control in this invention; Figure 8 This is a schematic diagram of the reactive power control loop of the inverter virtual synchronous machine control in this invention; Figure 9 It is the dynamic response waveform of the DC bus voltage when using a traditional control strategy; Figure 10 It is the dynamic response waveform of the DC bus voltage when the control strategy of this invention is adopted; Figure 11 This is a schematic diagram of the DC bus voltage response without the feedforward compensation of this invention under a scenario where the illumination conditions change abruptly. Figure 12 The waveforms of the active and reactive power of the inverter connected to the grid without the feedforward compensation of this invention are the dynamic response waveforms of the inverter under the scenario of sudden change in lighting conditions. Figure 13 This is a schematic diagram of the DC bus voltage response using the feedforward compensation of this invention under a scenario where lighting conditions change abruptly. Figure 14 This is the dynamic response waveform of the grid-connected active and reactive power of the inverter using the feedforward compensation of this invention under a scenario where lighting conditions change abruptly. Detailed Implementation

[0020] The embodiments of the present invention will be further described below with reference to the accompanying drawings: Example 1: As Figure 1 As shown, a method for seamless switching control of operating modes in a photovoltaic-energy storage system includes the following steps: S1. For a photovoltaic-storage system that includes an energy storage system and a photovoltaic system, obtain the DC bus voltage, photovoltaic output voltage and current, energy storage battery voltage and current, LC filter inductor current, and grid connection point voltage and current in the current photovoltaic-storage system. S2. For energy storage systems, a switching mechanism based on photovoltaic voltage feedforward is constructed to intervene in the adjustment process in advance and respond quickly to voltage changes (which can effectively suppress bus voltage fluctuations and enhance the dynamic adaptability of the system). S3. For photovoltaic systems, a dynamic control strategy based on photovoltaic power feedforward compensation is constructed to achieve rapid suppression of transient fluctuations in DC bus voltage (improving voltage stability). S4. Based on the switching mechanism in S2 and the dynamic control strategy in S3, grid control is carried out in combination with inverter virtual synchronous generator control technology to achieve stable operation of the photovoltaic-storage system in both strong and weak grid environments.

[0021] In S2, the switching mechanism based on photovoltaic voltage feedforward is specifically as follows: S2.1 Determine whether the photovoltaic system is operating normally, and issue an alarm signal when the photovoltaic system is not operating normally; S2.2. During normal operation of the photovoltaic system, a maximum power point tracking control strategy is adopted, and the output voltage is... After being regulated by a proportional-integral controller, it maximizes its own output power and outputs electrical energy to the DC bus to supply the load of the entire photovoltaic-storage system. S2.3 When the battery system of the energy storage system is detected to be offline or the SOC exceeds the set upper limit threshold, the energy storage converter of the energy storage system will no longer perform power exchange and will only maintain a safe standby state. At the same time, the photovoltaic system will switch from the maximum power point tracking control strategy to the constant voltage control mode. During the switching process, a switching mechanism based on photovoltaic voltage feedforward is adopted (to avoid power mismatch caused by sudden changes). That is, a dynamic voltage feedforward term is added to the output of the voltage loop PI regulator of the photovoltaic system. The control principle is as follows: ; In the formula, Indicates the duty cycle of the boost converter in a photovoltaic system; Represents the Laplace operator; This indicates the reference value for the DC bus voltage. Indicates the current DC bus voltage; This indicates the output voltage at the photovoltaic cell terminal; , These represent the proportional and integral coefficients of the voltage loop PI controller, respectively. The current mode is represented as: ; In the formula, This indicates the adjustment coefficient (typically ranging from 0.5 to 1, used to adjust the degree of photovoltaic voltage feedforward influence); MPPT indicates the maximum power point tracking control strategy; CV indicates the constant voltage control mode. This indicates the current duty cycle of the boost converter, reflecting changes in power regulation capability. This regulation structure improves the response to voltage disturbances in CV control mode by introducing a proportional term related to the duty cycle; while in MPPT mode, it completely eliminates voltage coupling, maintaining the tracking accuracy to the maximum power point.

[0022] The specific operation steps of the MPPT control strategy are as follows: Collect the output voltage of the photovoltaic cell. With output current And serve as the input for MPPT control; based on , The MPPT control algorithm is executed to obtain a voltage reference value, which is then compared with the current actual output voltage. The comparison is performed to generate a voltage deviation signal, which is then input to a proportional-integral (PI) controller, which outputs a control signal for adjusting the duty cycle of the photovoltaic converter. The duty cycle signal is transmitted to the pulse width modulation (PWM) module to generate the PWM switching signal. To control the switching behavior of the photovoltaic converter.

[0023] The MPPT control algorithm can be further refined into the following steps: Calculate the increments of output power and voltage of the photovoltaic system, and analyze the relationship between the two. The direction of voltage adjustment is determined by the sign of the power increment: when the power increment is positive, the voltage is increased; when it is negative, the voltage is decreased; when it is zero, the voltage is kept constant. The operating voltage is dynamically adjusted based on the judgment results, so that the operating point of the photovoltaic system gradually approaches the maximum power point; The photovoltaic output parameters are continuously sampled, and the disturbance-observation process is executed cyclically to ensure that the photovoltaic system operates stably near the maximum power point in the long term.

[0024] The process for determining whether the photovoltaic system is operating normally in S2.1 is as follows: S2.1.1 Collect alarm signals from the physical alarm devices of the photovoltaic system to determine if there are any abnormal alarms. If there are any abnormal alarms, the photovoltaic system is not operating normally. If there are no abnormal alarms, proceed to S2.1.2. S2.1.2 Obtain the operating status data of the photovoltaic system, and perform data fusion based on the operating status data to obtain fused feature data; S2.1.3. Based on the fused feature data, perform LSTM-classifier prediction and output the probability of the photovoltaic system operating abnormally within a set future period. If the probability of operating abnormality is less than the set threshold, the photovoltaic system operates normally; if the probability of operating abnormality is not less than the set threshold, the photovoltaic system does not operate normally.

[0025] Operating status data includes the junction temperature fluctuation range of the inverter IGBT module. Voc-T curve fitting deviation of photovoltaic string Combiner box communication packet loss rate Inverter DC side input capacitor ripple current Maximum power point voltage drift of photovoltaic string Series resistance of photovoltaic modules , photovoltaic array output power fluctuation and the input voltage ripple coefficient of photovoltaic inverter The integrated feature data includes the health status of the inverter power unit. String characteristic deviation Component internal resistance aging coefficient Power output stability coefficient Monitoring communication integrity rate and System Comprehensive Degradation Index ; The process of fusing data based on operational status data to obtain fused feature data includes the following steps: right and The health status of the inverter power unit is obtained by weighting the values. ; right and The maximum value is taken to obtain the string characteristic deviation. ; right The minimum value is used to obtain the component's internal resistance aging coefficient. ; right and Perform multiplication to obtain the power output stability coefficient. ; based on The monitoring communication integrity rate was obtained. ; right , , , and The system's overall degradation index is obtained by performing fusion processing. .

[0026] Inverter power unit health The calculation formula is: ; This represents the upper limit of the normal threshold for IGBT junction temperature fluctuation. This represents the upper limit of the normal threshold for capacitor ripple current. and All are weighting factors; String characteristic deviation The calculation formula is: ; and They are respectively and The upper limit of the normal threshold; Component internal resistance aging coefficient The calculation formula is: ; This is the aging threshold for the series resistance. Power output stability coefficient The calculation formula is: ; and They are respectively and The normal threshold upper limit, when the product is greater than 1 Take 1; Monitoring communication integrity The calculation formula is: ; This is the upper limit of the normal packet loss rate threshold. If the value is less than 0, it is set to 0, indicating that communication has completely failed; System Comprehensive Degradation Index The calculation formula is: ; , , , and They are respectively , , , and The moving average over one hour. , , , and All are weighting factors.

[0027] In S2.1.3, the LSTM classifier prediction based on fused feature data outputs the probability of the photovoltaic system experiencing operational anomalies within a set future time period, including the following steps: S2.1.3.1. The fused feature data from multiple consecutive time points are recorded as a feature sequence; S2.1.3.2 The feature sequence is input into the first layer of the bidirectional long short-term memory network BiLSTM to extract short-term features and obtain local feature vectors. The local feature vectors are then input into the second layer of the bidirectional long short-term memory network BiLSTM to extract long-term features and obtain global feature vectors. S2.1.3.3 Obtain the weight factor of each feature vector in the global feature vector, and perform a weighted summation on the global feature vector based on the obtained weight factors to obtain a weighted compressed feature vector; S2.1.3.4 The weighted compressed feature vector is input into the first fully connected layer of the classifier to obtain intermediate features. The intermediate features are then input into the second fully connected layer of the classifier, and the sigmoid function is used to output the probability of abnormal operation.

[0028] In S2.3, the process of the photovoltaic system switching from the maximum power point tracking control strategy to the constant voltage control mode is as follows: S2.3.1 Real-time detection of the SOC value of the energy storage battery to determine whether it has reached the set upper limit threshold, and at the same time, detection of whether the battery system is offline. S2.3.2 When SOC ≥ upper limit threshold, or when the battery system is offline / communication interrupted, the battery management system feeds back a charging prohibition signal, the photovoltaic system no longer maintains the maximum power point tracking control strategy, and the photovoltaic system controller issues a command to switch to constant voltage control mode. S2.3.3 The photovoltaic system switches to constant voltage control mode. In constant voltage control mode, the status of the photovoltaic system is continuously monitored. When the SOC drops to the set recovery threshold (70%), or the battery system comes back online, the photovoltaic system controller issues a command to switch to the maximum power point tracking control strategy and restores the maximum power point tracking control strategy.

[0029] In S3, the dynamic control strategy based on photovoltaic power feedforward compensation is as follows: S3.1 The energy storage converter adopts a voltage-current dual closed-loop control structure, consisting of an outer loop voltage controller and an inner loop current controller. The outer loop control objective is... ,contrast This generates the battery current command value. The inner loop uses current commands as the control target, and controls the energy storage converter to output the actual current through current loop control. ; It can be represented as: ; In the formula, This is the voltage loop proportional gain; The voltage loop integral gain is used; this dual-loop structure has good steady-state performance, but its dynamic adjustment speed is limited, especially under frequent PV power disturbances or load step changes, its adjustment capability is insufficient.

[0030] S3.2 To enhance the rapid response capability of the energy storage system to photovoltaic fluctuations, based on the voltage-current dual closed-loop control structure, a control method based on the output power of the photovoltaic array is introduced. The feedforward compensation term calculates the compensation current by sampling the photovoltaic output power in real time. ; In the formula, This indicates the photovoltaic power feedforward compensation current; This represents the feedforward gain coefficient; this compensation term directly reflects the degree of influence of photovoltaic injected power on the bus voltage, and has a pre-response function. It can actively adjust the current command before a significant voltage deviation occurs, effectively suppressing voltage disturbances.

[0031] Under CV control mode, the power balance relationship of the photovoltaic-storage system is as follows: ; In the formula, This refers to the active power output from the inverter to the power grid. This refers to internal losses within the photovoltaic energy storage system. S3.3, Total Current Reference Command for Energy Storage Batteries It is composed of the output of the outer loop voltage controller and the superposition of the feedforward compensation term: ; In the inner loop controller of the input energy storage system, the duty cycle or control voltage of the energy storage converter is adjusted.

[0032] In S4, the specific method for network control is as follows: S4.1. The photovoltaic system is controlled through the switching mechanism in S2 (to obtain the maximum power to the greatest extent), and the DC bus voltage is regulated through the dynamic control strategy in S3 (to keep the DC bus voltage stable). S4.2. For the inverter, a virtual synchronous generator control strategy is adopted to simulate the electromagnetic inertia and damping characteristics of a synchronous generator. Combined with active and reactive power control and inner-loop output voltage regulation, autonomous control of the inverter's output voltage amplitude and phase is achieved. The specific operation process of this step is as follows: S4.2.1. A virtual synchronous frequency is generated using active power control. In the active power control module, the rotational kinetic energy of a traditional synchronous generator is simulated, and a virtual rotor dynamics model is established. Its control law is as follows: ; In the formula, J is the moment of inertia of the photovoltaic-storage system; The actual angular frequency of the power grid; t is time; The damping coefficient of the photovoltaic-energy storage system; This is the rated angular frequency of the power grid; It serves as a reference value for active power; the output frequency is adjusted in real time according to the power deviation to achieve active adjustment and support of the frequency of the photovoltaic-storage system. S4.2.2 In the reactive power module, the excitation system of the synchronous machine is simulated to achieve dynamic adjustment of the output voltage amplitude. The control law is as follows: ; In the formula, The inverter output port voltage is represented by K, which is the proportional coefficient. This is the reactive power droop coefficient; This refers to the rated phase voltage amplitude of the power grid; This represents the actual output phase voltage amplitude of the power grid. This is a reference value for reactive power. Reactive power; S4.2.3. Combining the above control steps, the synchronous machine inertia and damping are simulated during grid-connected operation, and the frequency and voltage support capabilities are provided; the phase angle and voltage can be actively adjusted according to the load or PV disturbance.

[0033] Figure 2 The topology of a photovoltaic-energy storage grid-connected system (PV-energy storage system) based on a common DC bus architecture is presented. In this system, the photovoltaic power generation units and energy storage devices are connected to a unified DC bus platform via independent power converters, realizing DC-side aggregation and power dispatch of multiple energy sources. The output of the PV-energy storage system completes the DC-AC power conversion through the inverter stage, and the voltage and current waveforms are smoothed with the help of LC filters, ultimately achieving grid-connected operation.

[0034] This common DC bus topology offers significant advantages in modern distributed energy systems. First, by eliminating traditional multi-stage energy conversion paths, this architecture significantly reduces losses during energy conversion, thereby improving overall system efficiency. Second, since initial power coordination is performed on the DC side by each energy unit, its control system is simpler and easier to implement for high-dynamic-performance energy management and control compared to AC-side coordination strategies. Third, this structure demonstrates high cost-effectiveness in system construction and operation, reducing redundant power converter configurations and simplifying wiring complexity and equipment installation and maintenance. Furthermore, the unified DC bus platform provides favorable conditions for subsequent capacity expansion, modular deployment, and flexible integration with other DC sources or loads, significantly enhancing system scalability and construction flexibility.

[0035] Figure 2 middle, For photovoltaic side filter capacitor, For photovoltaic-side boost inductor, This refers to the power switching transistor in the photovoltaic-side DC-DC converter, where D represents the duty cycle. This refers to the terminal voltage of the energy storage battery. For the inductor of the energy storage side DC-DC converter, , These are the power switching transistors of the bidirectional DC-DC converter on the energy storage side. DC bus represents the DC bus. For DC bus capacitors, For filtering inductors, For damping resistor, For filtering capacitors, This refers to the three-phase voltage output by the inverter. The three-phase current output by the inverter. Let be the components of the inverter output voltage in the dq rotating coordinate system. Let represent the components of the inverter output current in the dq rotating coordinate system, and PCC be the point of common coupling. The equivalent inductance on the grid side, This is the equivalent resistance on the grid side. Let be the grid voltage, and abc and dq represent the abc coordinate system and the dq coordinate system, respectively.

[0036] Figure 3 The maximum power point tracking (MPPT) control structure for photovoltaic arrays is presented. Under normal operating conditions, to fully utilize the power generation capacity of the photovoltaic array, perturbation and observation (P&O) is employed for MPPT. Specifically, the controller first collects... and ,calculate It compares the power value with the previous moment, determines whether to continue perturbing the voltage command direction based on the power change trend, and finds the maximum power point.

[0037] Photovoltaic voltage reference value generated by this algorithm It is input to a voltage PI controller, and The error is calculated after comparison, and then an adjustment signal is output. The signal is generated by the PWM modulator. This control strategy adjusts the operating point of the photovoltaic array to stabilize it at maximum power output. It features a clear structure, high adjustment precision, good steady-state performance, and provides fundamental parameter support for subsequent constant voltage switching control modes.

[0038] Figure 4This paper presents a photovoltaic (PV) control strategy based on a PV voltage feedforward switching mechanism, enabling a smooth, disturbance-free switching between maximum power point tracking (MPPT) and constant voltage (CV) control modes for the PV system. Specifically, when the battery energy storage system reaches its maximum state of charge (e.g., SOC ≥ 80%) or temporarily shuts down due to a fault, the PV control needs to be switched from MPPT to CV mode in a timely manner to prevent the bus voltage from rising too quickly due to excess PV power. To address the voltage fluctuation problem caused by sudden changes in the control target during traditional switching processes, this embodiment introduces a PV voltage feedforward channel into the PI controller structure.

[0039] In this strategy, the sampled through The adjusted value is introduced as a compensation term into the PI controller output to correct the controller's dynamic response. The setting has mode-adaptive characteristics: it is set to zero in MPPT mode to ensure that maximum power point tracking is not disturbed; while in CV mode, its value is determined according to... and Real-time adjustment. This feedforward mechanism achieves a seamless transition between MPPT and CV control, avoiding bus voltage overshoot or fluctuations caused by adjustment lag in traditional strategies. Simulation results show that the system maintains voltage stability during switching and exhibits a smooth dynamic transition process after adopting this strategy, significantly improving the stability and continuity of the photovoltaic-storage system across different operating modes.

[0040] Figure 5 A photovoltaic voltage feedforward-based switching mechanism for energy storage control is presented. This strategy enables smooth switching between the energy storage system in shutdown mode and dual-closed-loop regulation mode to adapt to dynamic operational demands caused by changes in battery state of charge. Under normal operating conditions, the energy storage system operates in a voltage-current dual-closed-loop control mode, with the outer loop using... To control the target, the output current reference value is used, while the inner loop controls the actual current. This achieves power balance.

[0041] Figure 5 In this context, CV-BES indicates that the energy storage system is operating in constant voltage control mode, and STOP indicates the stop (energy storage charging and discharging) mode. It is the actual output voltage of the energy storage converter.

[0042] However, when the battery's state of charge (SOC) reaches the set upper limit or the energy storage system fails and goes offline, the energy storage can no longer participate in regulation. At this time, the photovoltaic-energy storage system needs to quickly switch the energy storage control state to shutdown mode, that is, the energy storage converter no longer performs power exchange and only maintains a safe standby state. To achieve this smooth switching of control state, a feedforward regulation mechanism based on photovoltaic voltage is introduced into the control logic, and the control mode is automatically switched by the internal logic of the controller. Specifically, the controller determines whether the energy storage has regulation capability based on the real-time acquired SOC signal. When the SOC is less than the upper limit threshold (e.g., 80%), the energy storage operates in a dual closed-loop control mode; when the SOC exceeds the threshold, the energy storage switches to shutdown mode, and at the same time, photovoltaic CV control is activated to maintain the stability of the bus voltage. Since there is a risk of abrupt changes in the control channel and current output during energy storage switching, by coordinating the participation of the photovoltaic voltage feedforward mechanism, the system can smoothly transition between energy storage shutdown and recovery, avoiding disturbances to the bus voltage during the control window period. This control strategy ensures that the energy storage system can switch between control modes safely, continuously, and flexibly according to its own state during dynamic operation, effectively guaranteeing the overall power coordination and voltage stability of the system, and enhancing the operational robustness and adaptability of the photovoltaic-storage system.

[0043] Figure 6 A dynamic control strategy based on photovoltaic power feedforward compensation is presented. Its core lies in utilizing the real-time output power of the photovoltaic system as a feedforward quantity to actively enhance the energy storage system's rapid response capability to DC bus disturbances. In this control structure, the sampled... through After normalization, a compensation current is formed ( This compensation term is directly superimposed on the battery current reference value output by the outer loop voltage regulator. This allows the energy storage system's output current to be adjusted in advance, before sudden changes in photovoltaic power cause a significant bus voltage shift, thus rapidly adjusting the power balance of the photovoltaic-energy storage system. Unlike traditional PI regulators that rely solely on voltage error signals, this strategy introduces an active factor into the regulation path, significantly improving the photovoltaic-energy storage system's feedforward suppression capability against dynamic disturbances, reducing bus voltage overshoot, and enhancing the system's stability and control response speed.

[0044] Figure 2 , Figures 3-6 In the middle, the red part represents the core improvement of this embodiment compared to the prior art.

[0045] Figure 7 This is a virtual synchronous machine control loop for the inverter. This loop simulates the rotor inertial response mechanism of a synchronous generator and dynamically adjusts the inverter's output frequency by constructing a virtual rotation equation. Specifically, this control loop uses the active power error as the driving signal and adjusts the frequency based on the system's set moment of inertia (equivalent moment of inertia) J and damping coefficient (active power damping coefficient). Real-time adjustment of output angular frequency (i.e.) The phase angle of the output voltage is then obtained through an integrator. This mechanism enables frequency control of the output voltage waveform. When the photovoltaic output changes or the load fluctuates, the inverter can provide virtual inertia support, mitigating frequency fluctuations and improving the dynamic stability of the photovoltaic-storage system.

[0046] Figure 8 This is the reactive power control loop in the inverter's virtual synchronous generator control structure. This loop uses... and Based on this, through the proportional coefficient K and the reactive power droop coefficient Construct an excitation regulation model to achieve... The control structure, mimicking the excitation system of a synchronous generator, dynamically adjusts the output voltage according to reactive power demand, thereby regulating the inverter's reactive power support capability. The aforementioned active and reactive power control strategies operate independently yet collaboratively, forming a complete virtual synchronous control framework, significantly enhancing the inverter's dual support capability for grid voltage and frequency. Under unsteady conditions such as grid faults, disturbances, or fluctuations, this control strategy proactively provides power regulation to maintain stable operation, enhancing the adaptability, dynamic adjustment capability, and grid compatibility between the photovoltaic-storage system and the grid, ensuring the system's stability and reliability in various complex scenarios such as weak grids and islanded operation.

[0047] Figure 9 and Figure 10 This paper presents the dynamic response waveforms of the DC bus voltage under different conditions: a conventional control strategy and a photovoltaic control strategy based on a photovoltaic voltage feedforward switching mechanism proposed in this embodiment, both when the state of charge of the energy storage system exceeds its upper limit. The results show that when the energy storage system can no longer be charged, the bus voltage fluctuation under the conventional strategy reaches as high as 10V, with a settling time of approximately 380ms. However, after introducing the switching mechanism of this embodiment, the voltage fluctuation is significantly reduced to 4.9V, and the settling time is shortened to 260ms. These results verify the effectiveness of the strategy proposed in this embodiment in ensuring smooth transition of the photovoltaic control mode and stabilizing the bus voltage.

[0048] Figure 11 and Figure 13 It exhibits a sudden change in illumination conditions (from 1500 W / m at t=2s). 2 The power dropped sharply to 700 W / m 2In this scenario, the response of the DC bus voltage is compared with and without the dynamic control strategy based on photovoltaic power feedforward compensation proposed in this embodiment. It can be clearly observed that under the traditional PI control strategy, the bus voltage experiences a 70V overshoot with a settling time as long as 750ms; while after introducing the control strategy of this embodiment, the voltage fluctuation is significantly reduced to 5V, and the settling time is shortened to 250ms. These results demonstrate that this embodiment significantly enhances the rapid response capability of the photovoltaic-storage system to illumination disturbances and improves the voltage stability and control robustness of the photovoltaic-storage system.

[0049] Figure 12 and Figure 14 The dynamic response waveforms of the inverter's grid-connected active and reactive power under the aforementioned illumination disturbance conditions are presented. At t=2s, the active power fluctuation under conventional control is approximately 600W, and the reactive power fluctuation reaches 500Var. However, after adopting the control strategy proposed in this embodiment, both active and reactive power fluctuations are rapidly suppressed to 0, indicating that a seamless transition is achieved at the power level. This effectively avoids system fluctuations caused by energy imbalance, thereby ensuring the grid-connected power quality and dynamic operational stability of the photovoltaic-storage system.

[0050] In summary, this embodiment addresses key issues such as severe fluctuations in photovoltaic power, frequent limitations on battery state of charge, and response lag and voltage disturbances during traditional control switching processes. It constructs a multi-modal collaborative control framework centered on photovoltaic power and voltage feedforward. Through dynamic power compensation and flexible switching mechanisms, it effectively achieves smooth transitions between photovoltaic MPPT and CV control, as well as energy storage start-up and shutdown states, significantly improving the dynamic performance, voltage stability, and grid-connected adaptability of the photovoltaic-energy storage system. This control method possesses engineering advantages such as a simple control structure and ease of embedding into existing platforms, providing crucial support for building highly reliable and highly adaptive photovoltaic-energy storage integrated systems, and has significant theoretical value and practical application significance.

[0051] Example 2: A seamless switching control device for the operating modes of a photovoltaic-energy storage system, comprising: One or more processors; Memory, used to store one or more computer programs; When one or more programs are executed by one or more processors, the one or more processors execute the method in Example 1.

[0052] Example 3: A computer-readable storage medium having executable instructions stored thereon, which, when executed by a processor, cause the processor to perform the method in Example 1.

Claims

1. A method for seamless switching control of operating modes of a photovoltaic- energy storage system, characterized in that The method comprises the following steps: S1, for a light storage system comprising an energy storage system and a photovoltaic system, obtaining the DC bus voltage, photovoltaic output voltage and current, energy storage battery voltage and current, LC filter inductance current, and grid-connected point voltage and current in the current light storage system; S2, for the energy storage system, constructing a switching mechanism based on photovoltaic voltage feedforward, intervening in the adjustment process in advance, and quickly responding to voltage changes; S3, for the photovoltaic system, constructing a dynamic control strategy based on photovoltaic power feedforward compensation, and realizing rapid suppression of transient fluctuations of the DC bus voltage; S4, based on the switching mechanism in S2 and the dynamic control strategy in S3, combining inverter virtual synchronous generator control technology for grid control, and realizing stable operation of the light storage system in strong and weak grid environments.

2. The method of claim 1, wherein, In S2, the switching mechanism based on photovoltaic voltage feedforward is specifically: S2.1, judging whether the photovoltaic system is normally operating, and issuing an alarm signal when the photovoltaic system is not normally operating; S2.2, when the photovoltaic system is in normal operation, a maximum power point tracking control strategy is adopted, and the output voltage is , which is adjusted by a proportional integral controller to maximize its output power and then outputs electrical energy to the DC bus to supply the load of the entire photovoltaic storage system; S2.3, when it is detected that the battery system of the energy storage system is offline or the SOC exceeds the set upper threshold, the energy storage converter of the energy storage system no longer exchanges power, and only maintains a safe standby state, and at the same time, the photovoltaic system is switched from the maximum power point tracking control strategy to the constant voltage control mode; In the switching process, a switching mechanism based on photovoltaic voltage feedforward is adopted, that is, a voltage dynamic feedforward term is added to the output of the voltage loop PI regulator of the photovoltaic system, and the control principle is as follows: ； In the formula, denotes the duty cycle of the boost converter in the photovoltaic system; denotes the Laplace operator; denotes the DC bus voltage reference value; denotes the current DC bus voltage; denotes the output voltage at the end of the photovoltaic cell; , respectively denote the proportional coefficient and the integral coefficient of the voltage loop PI controller; is the current mode, denoted as: ； In the formula, denotes a regulation factor; MPPT denotes a maximum power point tracking control strategy; CV denotes a constant voltage control mode; denotes the current duty cycle of the boost converter.

3. The method of claim 2, wherein, The judgment process of whether the photovoltaic system is normally operating in S2.1 is as follows: S2.1.1, collecting alarm signals of physical alarm devices of the photovoltaic system, determining whether there is an abnormal alarm, if there is an abnormal alarm, the photovoltaic system is not normally operating, if there is no abnormal alarm, S2.1.2 is entered; S2.1.2, obtaining operating state data of the photovoltaic system, and performing data fusion based on the operating state data to obtain fusion feature data; S2.1.3, based on the fusion feature data, an LSTM-classifier prediction is performed to output an operating abnormality probability of the photovoltaic system occurring operating abnormality in a set future period of time, if the operating abnormality probability is less than a set operating abnormality probability threshold, the photovoltaic system is normally operating, if the operating abnormality probability is not less than the set operating abnormality probability threshold, the photovoltaic system is not normally operating.

4. The method of claim 3, wherein, The running state data includes an IGBT module junction temperature fluctuation amplitude of the inverter , a Voc-T curve fitting deviation of the photovoltaic string , a communication packet loss rate of the junction box , an input capacitor ripple current of the DC side of the inverter , a maximum power point voltage drift of the photovoltaic string , a series resistance of the photovoltaic module , an output power fluctuation rate of the photovoltaic array , and an input voltage ripple coefficient of the photovoltaic inverter The fusion feature data includes an inverter power unit health degree , a string characteristic deviation degree , a module internal resistance aging coefficient , a power output stability coefficient , a monitoring communication integrity rate , and a system comprehensive degradation index ; The data fusion based on the operating state data to obtain the fusion feature data comprises the following steps: right and The health status of the inverter power unit is obtained by weighting the values. ; To and The maximum value is taken to obtain the group string characteristic deviation ; To Take the minimum value processing, get the internal resistance of the component aging coefficient ; The product of and is processed to obtain a power output stability coefficient ; Based on , the communication integrity rate is monitored ; Fusion processing is performed on , , , and to obtain a system comprehensive degradation index .

5. The method of claim 4, wherein, Inverter power unit health The formula for calculating the health is: ； is an upper limit of a normal threshold for IGBT junction temperature fluctuation, is an upper limit of a normal threshold for capacitor ripple current, and are weight factors; String characteristic deviation The calculation formula is: ； and respectively and upper normal threshold limits; Packaging material The formula for calculating the aging coefficient of the internal resistance of the assembly is: ； is a series resistance aging threshold value; Power output stability coefficient The calculation formula is: ； and are respectively and upper normal thresholds, when the product is greater than 1 take 1; Monitoring communication integrity The formula for calculating the communication integrity is: ； is a normal threshold upper limit for the packet loss rate, 0 if less than 0, indicating complete failure of communication; Systematic degradation index The formula for calculating the systematic degradation index is: ； , , , and are respectively , , , and a sliding average over one hour, , , , and are weight factors.

6. The method of claim 4, wherein, In S2.1.3, the LSTM-classifier prediction based on the fusion feature data to output the operating abnormality probability of the photovoltaic system occurring operating abnormality in a set future period of time comprises the following steps: S2.1.3.1, the fusion feature data at continuous multiple time points is recorded as a feature sequence; S2.1.3.2, the feature sequence is input into a first layer bidirectional long short-term memory network BiLSTM to extract short-term features and obtain a local feature vector, and the local feature vector is input into a second layer bidirectional long short-term memory network BiLSTM to extract long-term features and obtain a global feature vector; S2.1.3.3, obtain a weight factor of each feature vector in the global feature vector, and perform weighted summation on the global feature vector based on the obtained weight factor to obtain a weighted compressed feature vector; S2.1.3.4, input the weighted compressed feature vector into the first full connection layer of the classifier to obtain an intermediate feature, and input the intermediate feature into the second full connection layer of the classifier to output an operation abnormality probability by using a sigmoid function.

7. The method of claim 2, wherein the method further comprises: In S2.3, the process of switching the photovoltaic system from the maximum power point tracking control strategy to the constant voltage control mode is: S2.3.1, real-time detection of the SOC value of the energy storage battery to determine whether it reaches the set upper threshold value, and detection of whether the battery system is in an offline state; S2.3.2, when SOC is greater than or equal to the upper threshold value, or the battery system is offline / communication interruption, the battery management system feeds back a charging prohibition signal, the photovoltaic system no longer maintains the maximum power point tracking control strategy, and the controller of the photovoltaic system issues a switching instruction to the constant voltage control mode; S2.3.3, the photovoltaic system is switched to the constant voltage control mode, and the state of the photovoltaic system is continuously monitored in the constant voltage control mode, when the SOC is detected to drop to the set recovery threshold value, or the battery system is reconnected, the controller of the photovoltaic system issues a switching instruction to the maximum power point tracking control strategy, and the photovoltaic system returns to the maximum power point tracking control strategy.

8. The method of claim 7, wherein, The upper threshold value is set to 80%, and the recovery threshold value is set to 70%.

9. The method of claim 2, wherein the method further comprises: In S3, the dynamic control strategy based on photovoltaic power feedforward compensation is specifically: S3.1, the energy storage converter adopts a voltage-current double closed loop control structure, which is composed of an outer loop voltage controller and an inner loop current controller, and the outer loop control target is , by comparison , and then generating a battery current instruction value , the inner loop takes the current instruction as the control target, and controls the actual current output by the energy storage converter through the current loop ; S3.2, On the basis of the voltage-current double closed-loop control structure, a feedforward compensation term based on the output power of the photovoltaic array is introduced, and the compensation current is calculated by real-time sampling of the photovoltaic output power: ​ ； In the formula, represents a photovoltaic power feed-forward compensation current; represents a feed-forward gain coefficient; S3.3, total current reference command of the energy storage battery The outer loop voltage controller outputs a superposition of the feedforward compensation term: ； In the current inner loop controller of the input energy storage system, the duty cycle of the energy storage converter or control voltage is adjusted.

10. The method of claim 1, wherein, In S4, the specific method of network construction control is: S4.1, the photovoltaic system is controlled by the switching mechanism in S2, and the direct current bus voltage is adjusted by the dynamic control strategy in S3; S4.2, for the inverter, a virtual synchronous generator control strategy is adopted to simulate the electromagnetic inertia and damping characteristics of the synchronous generator, and combined with active and reactive power control and inner loop output voltage stabilization, the amplitude and phase of the inverter output voltage are autonomously controlled.

Images